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,context_mla,vllm_triton_mla,float16,float16,128,1,1,1,0,0.49452801545461017
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,1,1,0,0.3080853422482808
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,1,8,0,0.3141760031382243
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,1,2,0,0.32154132922490436
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,1,8,0,0.3203413287798564
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,1,2,0,0.317248006661733
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,1,4,0,0.34937067826588947
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,1,32,0,0.3224533398946126
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,1,16,0,0.4873973528544108
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,1,1,0,0.4160906473795573
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,1,2,0,0.41368532180786133
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,1,4,0,0.33852799733479816
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,1,8,0,0.3038240075111389
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,1,4,0,0.31866133213043213
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,1,16,0,0.33557331562042236
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,1,32,0,0.3146773378054301
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,1,64,0,0.32733333110809326
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,1,1,0,0.3145013252894084
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,1,64,0,0.310805340607961
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,1,2,0,0.312991996606191
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,1,4,0,0.30267200867335003
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,1,8,0,0.4459253152211507
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,1,16,0,0.3347413142522176
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,1,64,0,0.32657599449157715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,1,32,0,0.4553866783777873
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,1,1,0,0.4875946839650472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,1,2,0,0.3328746755917867
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,1,4,0,0.3150879939397176
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,1,8,0,0.3405653238296509
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,1,16,0,0.3071626623471578
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,1,32,0,0.33779199918111164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,1,16,0,0.43689600626627606
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,1,64,0,0.3128906687100728
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,1,32,0,0.4907146692276001
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,1,64,0,0.4933653275171916
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,1,1,0,0.3437439997990926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,1,2,0,0.45073068141937256
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,1,4,0,0.4675146738688151
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,1,8,0,0.31146132946014404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,1,16,0,0.33585067590077716
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,1,32,0,0.31626667579015094
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,1,64,0,0.3343520164489746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,1,1,0,0.33261332909266156
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,1,2,0,0.31043734153111774
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,1,4,0,0.3208906650543213
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,1,8,0,0.3641066551208496
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,1,16,0,0.32997334003448486
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,1,32,0,0.3239840070406596
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,1,64,0,0.31009066104888916
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,1,1,0,0.31626667579015094
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,1,2,0,0.33079467217127484
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,1,4,0,0.37467201550801593
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,1,8,0,0.33420801162719727
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,1,16,0,0.3472586472829183
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,1,32,0,0.3104106585184733
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,1,64,0,0.3093760013580322
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,1,128,0,0.3120159904162089
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,256,1,1,0,0.5709120035171509
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,256,1,2,0,0.31940267483393353
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,1,128,0,0.32130134105682373
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,256,1,4,0,0.3077546755472819
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,256,1,8,0,0.3452906608581543
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,256,1,16,0,0.3137813409169515
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,256,1,32,0,0.3086453278859456
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,256,1,64,0,0.3213919997215271
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,16,1,0,0.2648906707763672
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,16,2,0,0.28640000025431317
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,256,1,128,0,0.29545066754023236
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,1,128,0,0.3190079927444458
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,16,4,0,0.29688533147176105
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,16,8,0,0.2926186720530192
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,16,16,0,0.2945280075073242
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,16,32,0,0.2850400010744731
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,16,64,0,0.2895573377609253
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,16,128,0,0.2807146708170573
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,16,1,0,0.2599839965502421
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,16,2,0,0.2860693335533142
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,16,4,0,0.26765332619349164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,16,8,0,0.2595733404159546
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,16,64,0,0.2897760073343913
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,16,16,0,0.28818132479985553
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,16,32,0,0.2932426730791728
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,16,128,0,0.2988319993019104
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,16,1,0,0.2829973300298055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,16,2,0,0.290613333384196
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,16,4,0,0.2895680069923401
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,16,8,0,0.28964799642562866
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,16,16,0,0.287882665793101
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,16,32,0,0.2809866666793823
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,16,64,0,0.291866660118103
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,16,128,0,0.2704053322474162
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,16,4,0,0.287663996219635
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,16,1,0,0.28934399286905926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,16,2,0,0.28728532791137695
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,16,16,0,0.3011893431345622
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,16,32,0,0.29632532596588135
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,16,64,0,0.2627519965171814
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,16,8,0,0.28731199105580646
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,16,128,0,0.24075732628504434
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,1,128,0,0.34703465302785236
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,16,4,0,0.287498672803243
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,16,1,0,0.3055359919865926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,16,2,0,0.29280000925064087
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,16,8,0,0.29843733708063763
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,16,16,0,0.2897973259290059
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,16,32,0,0.2880000074704488
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,16,64,0,0.31701334317525226
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,16,128,0,0.26289065678914386
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,16,2,0,0.2996533314387004
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,16,1,0,0.5155839920043945
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,16,4,0,0.2847200036048889
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,16,8,0,0.2872213323911031
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,16,32,0,0.28710933526357013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,16,16,0,0.2938986619313558
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,16,64,0,0.2860906720161438
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,16,128,0,0.24304000536600748
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,16,1,0,0.9720106919606527
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,16,2,0,0.5167626539866129
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,16,4,0,0.29132266839345294
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,1,128,0,0.31273066997528076
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,16,16,0,0.2889973322550456
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,16,32,0,0.2643839915593465
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,16,8,0,0.2970026731491089
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,16,128,0,0.2722453276316325
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,16,64,0,0.2698720097541809
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,16,2,0,0.9775359630584717
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,1,128,0,0.3360319932301839
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,16,4,0,0.5241546630859375
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,16,8,0,0.30052800973256427
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,16,1,0,1.881696065266927
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,16,32,0,0.2940799991289775
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,16,16,0,0.2945493261019389
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,16,64,0,0.2707413236300151
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,16,128,0,0.26766933997472125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,256,16,2,0,1.8942987124125164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,256,16,4,0,0.9873066743214926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,256,16,16,0,0.3073599934577942
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,256,16,32,0,0.2939573327700297
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,256,16,8,0,0.527728001276652
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,32,2,0,0.2918933431307475
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,32,1,0,0.2878719965616862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,256,16,1,0,3.697733243306478
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,32,4,0,0.28463999430338544
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,32,8,0,0.2881866693496704
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,1,128,0,0.33992000420888263
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,256,16,64,0,0.2977760036786397
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,32,16,0,1.30459729830424
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,32,128,0,1.7904853820800781
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,256,16,128,0,0.27089067300160724
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,32,32,0,0.28180267413457233
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,32,64,0,1.0113653341929119
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,32,4,0,0.28431999683380127
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,32,8,0,0.5432906548182169
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,32,2,0,0.285098671913147
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,32,1,0,0.2906293272972107
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,32,32,0,1.3201759656270344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,32,16,0,0.2857813239097595
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,32,64,0,0.946773370107015
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,32,128,0,0.27931733926137287
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,32,2,0,1.4066826502482097
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,32,4,0,0.2926773428916931
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,32,8,0,0.29249600569407147
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,32,1,0,0.29502399762471515
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,32,16,0,2.1124800046284995
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,32,64,0,0.28733332951863605
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,32,32,0,0.2910613417625427
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,32,1,0,1.210378646850586
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,32,128,0,0.2676266630490621
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,32,4,0,0.29363733530044556
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,32,2,0,0.2680746714274089
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,32,8,0,2.1370293299357095
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,32,32,0,1.3114346663157146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,32,16,0,0.29474133253097534
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,32,128,0,0.27640533447265625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,32,64,0,1.2190773487091064
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,32,1,0,0.9026986757914225
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,32,4,0,0.2938186724980672
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,32,2,0,0.2959413329760234
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,32,8,0,0.2893066604932149
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,32,32,0,0.29151467482248944
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,32,16,0,0.2887146671613057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,32,128,0,0.2694293260574341
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,32,64,0,0.29344000418980914
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,32,2,0,0.40826666355133057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,32,4,0,0.27743999163309735
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,32,1,0,0.750495990117391
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,32,8,0,0.2728746732076009
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,32,32,0,0.3614879846572876
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,32,128,0,0.27138133843739826
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,32,64,0,0.2884426712989807
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,32,16,0,0.28273600339889526
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,32,4,0,0.7761440277099609
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,32,1,0,1.4301493962605794
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,32,8,0,0.2951359947522481
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,32,2,0,0.7544906934102377
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,32,64,0,0.29042667150497437
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,32,32,0,0.2824266751607259
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,32,16,0,0.29097600777943927
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,32,128,0,0.26129599412282306
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,32,8,0,0.41599468390146893
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,32,4,0,0.7633066972096761
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,32,16,0,0.3208373387654622
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,32,64,0,0.31835200389226276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,32,32,0,0.8795626958211263
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,32,1,0,2.820741335550944
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,32,2,0,1.4739947319030762
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,1,128,0,0.3118879993756612
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,32,128,0,0.26921600103378296
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,256,32,8,0,0.7787573337554932
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,256,32,16,0,0.44021332263946533
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,256,32,4,0,1.4571679433186848
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,256,32,2,0,2.813098589579264
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,256,32,32,0,0.29632532596588135
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,256,32,64,0,0.2666719953219096
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,64,2,0,0.2893600066502889
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,64,1,0,0.26659200588862103
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,256,32,1,0,5.52677853902181
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,64,8,0,0.30249067147572833
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,64,16,0,0.29180266459782916
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,64,32,0,0.29014400641123456
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,64,4,0,0.2788800001144409
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,64,64,0,0.29228800535202026
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,64,128,0,0.27771733204523724
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,256,32,128,0,0.2675466736157735
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,64,2,0,0.2946186661720276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,64,16,0,0.32506134112675983
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,64,4,0,0.2940479914347331
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,64,32,0,0.2925119996070862
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,64,8,0,0.28646934032440186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,64,64,0,0.3097599943478902
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,64,1,0,0.2961600025494893
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,64,128,0,0.24183466037114462
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,64,4,0,1.1988266309102376
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,64,32,0,0.31836267312367755
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,64,8,0,0.9469172954559326
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,64,2,0,0.29444799820582074
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,64,64,0,0.29550399382909137
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,64,1,0,0.290175994237264
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,64,16,0,1.3347147305806477
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,64,128,0,0.27080533901850384
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,64,8,0,0.2971946597099304
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,64,1,0,0.368341326713562
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,64,4,0,0.37043734391530353
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,64,64,0,0.320848007996877
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,64,16,0,0.2909066677093506
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,64,32,0,0.2944693366686503
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,64,2,0,0.2922666668891907
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,64,128,0,0.2675146659215291
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,64,2,0,0.40261868635813397
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,64,4,0,0.837562640508016
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,64,8,0,0.29078932603200275
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,64,1,0,0.6706240177154541
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,64,128,0,0.2606400052706401
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,64,32,0,1.0543200174967449
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,64,16,0,0.2891146739323934
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,64,64,0,0.30186132589975995
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,64,2,0,0.6604479948679606
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,64,32,0,0.2895946701367696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,64,1,0,1.2353066603342693
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,64,16,0,0.5278453429539999
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,64,128,0,0.2783626715342204
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,64,4,0,0.3733493487040202
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,64,64,0,0.2898293336232503
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,64,8,0,0.26917866865793866
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,64,16,0,0.29997867345809937
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,64,2,0,1.2353599866231282
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,64,4,0,0.6650079886118571
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,64,8,0,0.37196266651153564
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,64,1,0,2.3802293141682944
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,64,32,0,0.2922453284263611
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,64,128,0,0.26368000109990436
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,64,64,0,0.28994133075078327
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,64,16,0,0.39638932545979816
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,64,8,0,0.6815893650054932
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,64,32,0,0.28884265820185345
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,64,64,0,0.7966986497243246
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,64,128,0,0.328821341196696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,64,2,0,2.438549359639486
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,64,4,0,1.2518133322397869
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,64,1,0,4.692565282185872
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,256,64,32,0,1.008682648340861
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,256,64,16,0,1.0068000157674153
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,256,64,8,0,1.2829066912333171
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,256,64,64,0,0.28702400128046673
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,256,64,4,0,2.4286932945251465
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,256,64,128,0,0.2675466736157735
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,256,64,2,0,5.444592157999675
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,128,8,0,0.2930240035057068
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,128,1,0,0.3262453277905782
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,128,4,0,0.29874666531880695
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,128,2,0,0.2850826581319173
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,128,16,0,0.29254400730133057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,256,64,1,0,9.290277481079102
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,128,128,0,0.26600533723831177
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,128,64,0,0.28836800654729206
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,128,32,0,0.2961813410123189
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,128,1,0,0.2967733343442281
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,128,4,0,0.292303999265035
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,128,2,0,0.2937973340352376
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,128,8,0,0.2882399956385295
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,128,64,0,0.2861599922180176
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,128,16,0,0.296122670173645
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,128,32,0,1.3382879892985027
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,128,128,0,0.27195199330647785
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,128,2,0,1.024341344833374
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,128,4,0,0.29474133253097534
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,128,1,0,0.3556853135426839
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,128,32,0,0.28150399525960285
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,128,8,0,0.2893600066502889
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,128,16,0,0.36615999539693195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,128,64,0,0.26576000452041626
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,128,128,0,0.27242134014765423
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,128,2,0,0.3561280171076457
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,128,1,0,0.6538026730219523
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,128,16,0,0.29148799180984497
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,128,8,0,0.29390933116277057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,128,4,0,0.2882026632626851
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,128,32,0,0.2988319993019104
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,128,64,0,0.32500267028808594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,128,128,0,0.24885332584381104
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,128,16,0,0.28678399324417114
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,128,32,0,0.2890346646308899
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,128,1,0,1.2350880304972331
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,128,8,0,0.4314986864725749
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,128,4,0,0.38417065143585205
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,128,2,0,0.6542880137761434
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,128,128,0,0.26453866561253864
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,128,64,0,0.27262399593989056
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,128,2,0,1.2368586858113606
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,128,4,0,0.6613706747690836
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,128,8,0,0.36500799655914307
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,128,32,0,0.28915733098983765
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,128,16,0,0.29875733455022174
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,128,64,0,0.2701386610666911
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,128,1,0,2.3933706283569336
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,128,128,0,0.24796799818674722
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,128,16,0,0.38787734508514404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,128,32,0,0.2933173378308614
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,128,64,0,0.293178657690684
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,128,8,0,0.6761066913604736
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,128,4,0,1.248032013575236
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,128,2,0,2.455338637034098
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,128,128,0,0.2712639967600505
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,128,1,0,4.664437294006348
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,128,16,0,0.71833602587382
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,128,8,0,1.2872640291849773
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,128,4,0,2.4151466687520347
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,128,32,0,0.4673333168029785
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,128,64,0,0.2839573423067729
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,128,2,0,4.6982879638671875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,128,128,0,0.24660799900690714
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,128,1,0,9.334778467814127
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,256,128,64,0,0.5168639818827311
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,256,128,32,0,0.8000319798787435
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,256,128,4,0,4.767162640889485
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,256,128,16,0,1.3607786496480305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,256,2,0,0.2939786712328593
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,256,4,0,0.2925493319829305
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,256,128,128,0,0.6145439942677816
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,256,128,8,0,2.4825493494669595
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,256,1,0,0.37918933232625324
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,256,16,0,0.2871786753336589
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,256,8,0,0.29305599133173627
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,256,32,0,1.0781866709391277
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,256,128,2,0,9.326218922932943
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,256,128,0,0.27461334069569904
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,256,64,0,1.2875626881917317
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,256,4,0,0.29466134309768677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,256,1,0,0.36829864978790283
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,256,2,0,0.299344003200531
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,256,8,0,0.2680533329645793
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,256,32,0,0.5919146537780762
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,256,16,0,0.29473066329956055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,256,64,0,0.28506133953730267
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,256,128,0,0.26258132855097455
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,256,4,0,0.29444799820582074
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,256,2,0,0.36937065919240314
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,256,16,0,0.2926186720530192
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,256,1,0,0.6595946550369263
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,256,8,0,0.29789867003758747
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,256,64,0,0.29331199328104657
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,256,32,0,0.29443200429280597
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,256,128,0,0.26902933915456134
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,256,4,0,0.37090667088826496
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,256,2,0,0.6629279851913452
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,256,8,0,0.295418659845988
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,256,128,1,0,18.548453013102215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,256,16,0,0.32356266180674237
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,256,32,0,0.29309332370758057
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,256,1,0,1.2588213284810383
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,256,64,0,0.29068267345428467
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,256,128,0,0.2744266589482625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,256,4,0,0.6689759890238444
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,256,8,0,0.3749599854151408
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,256,16,0,1.4052480061848958
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,256,2,0,1.2439839839935303
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,256,32,0,0.2984586755434672
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,256,1,0,2.3942294120788574
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,256,128,0,0.24876266717910767
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,256,64,0,0.2860640088717143
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,256,2,0,2.422976016998291
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,256,8,0,0.9418453375498453
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,256,32,0,0.28936533133188885
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,256,64,0,0.26970134178797406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,256,16,0,0.4174933433532715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,256,4,0,1.6558027267456055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,256,1,0,4.7158559163411455
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,256,128,0,0.2680000066757202
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,256,32,0,1.3078293005625408
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,256,64,0,0.2990933259328206
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,256,8,0,1.2952746550242107
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,256,16,0,0.7267359892527262
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,256,4,0,2.4362239837646484
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,256,128,0,0.2481600046157837
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,256,2,0,4.717354774475098
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,256,1,0,9.311637242635092
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,256,8,0,2.517850716908773
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,256,16,0,1.3830506006876628
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,256,32,0,0.8398239612579346
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,256,4,0,4.850111961364746
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,256,64,0,0.5250933170318604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,256,128,0,0.3704746564229329
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,256,2,0,9.370922724405924
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,256,256,8,0,4.94210147857666
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,256,1,0,18.607167561848957
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,256,256,16,0,2.683274586995443
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,256,256,32,0,1.549850622812907
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,512,4,0,0.2983146707216899
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,512,1,0,0.4024213155110677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,256,256,4,0,9.510165532430014
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,512,2,0,0.26869332790374756
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,512,8,0,0.29337600866953534
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,512,16,0,0.29229867458343506
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,512,32,0,0.294320007165273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,512,64,0,0.28616533676783246
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,512,128,0,0.2717706759770711
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,256,256,64,0,0.9875946839650472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,512,2,0,0.4028000036875407
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,512,16,0,0.28170132637023926
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,512,4,0,0.29839466015497845
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,256,256,2,0,18.81329600016276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,512,8,0,0.4676320155461629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,512,1,0,1.9012799263000488
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,512,32,0,0.28945066531499225
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,256,256,128,0,0.6812960306803385
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,512,64,0,0.3058346708615621
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,512,128,0,0.24352532625198364
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,512,16,0,0.2846933404604594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,512,4,0,0.4086079994837443
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,512,8,0,0.3006933331489563
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,512,2,0,0.7258666356404623
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,512,32,0,0.28774933020273846
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,512,64,0,0.29492799441019696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,512,128,0,0.26851733525594074
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,512,1,0,1.36245330174764
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,512,64,0,0.4596800009409587
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,512,16,0,0.29426666100819904
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,512,8,0,0.4091413418451945
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,512,32,0,0.2870453397432963
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,512,2,0,1.3671733538309734
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,512,4,0,0.7323520183563232
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,512,128,0,0.24592000246047974
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,512,1,0,2.6274399757385254
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,512,16,0,0.43425599733988446
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,512,8,0,0.7491146723429362
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,512,32,0,0.2739893396695455
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,256,256,1,0,37.25992075602213
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,512,4,0,1.3796639442443848
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,512,2,0,2.6588053703308105
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,512,64,0,0.9437493483225504
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,512,128,0,0.27140265703201294
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,512,1,0,5.167274792989095
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,512,16,0,0.7926987012227377
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,512,8,0,1.4186933835347493
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,512,32,0,0.476032018661499
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,512,64,0,0.3115413387616475
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,512,4,0,2.7184425989786782
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,512,128,0,0.24422399202982584
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,512,2,0,5.211093266805013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,512,64,0,1.5340426762898762
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,512,1,0,10.309247970581055
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,512,16,0,1.5113706588745117
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,512,32,0,0.8767039775848389
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,512,8,0,2.7465813954671225
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,512,128,0,0.39049601554870605
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,512,4,0,5.2698774337768555
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,512,2,0,10.297599792480469
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,128,512,32,0,1.6766239802042644
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,128,512,16,0,2.931957244873047
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,128,512,64,0,1.0503893693288167
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,1024,2,0,0.47861333688100177
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,128,512,8,0,5.445770899454753
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,1024,1,0,0.8623627026875814
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,128,512,128,0,0.7187946637471517
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,1024,8,0,0.295360008875529
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,128,512,4,0,10.452085494995117
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,1024,16,0,0.291157325108846
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,1024,4,0,0.3055306673049927
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,1024,32,0,0.2865440050760905
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,1024,128,0,0.26711465915044147
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,1024,64,0,0.28911999861399335
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,512,1,0,20.510021209716797
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,1024,2,0,0.8650720119476318
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,1024,4,0,0.4840746720631917
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,1024,8,0,0.2955999970436096
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,1024,16,0,0.2900533278783162
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,1024,32,0,0.2888373335202535
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,1024,1,0,1.6234985987345378
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,1024,128,0,0.30001066128412884
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,1024,64,0,0.30744000275929767
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,128,512,2,0,20.662490844726562
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,1024,32,0,0.304202675819397
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,1024,16,0,0.30688534180323285
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,1024,8,0,0.48766398429870605
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,1024,64,0,0.29073067506154376
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,1024,2,0,1.962959925333659
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,1024,128,0,0.27034133672714233
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,1024,4,0,0.8767680327097574
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,1024,1,0,3.2000481287638345
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,1024,8,0,0.8870399792989095
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,1024,4,0,1.654597282409668
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,1024,32,0,0.3198026617368062
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,1024,16,0,0.5118773380915324
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,1024,2,0,3.154746691385905
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,1024,64,0,0.30425065755844116
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,1024,128,0,0.2532479961713155
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,1024,1,0,6.1762135823567705
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,1024,16,0,0.9310666720072428
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,1024,8,0,1.6785173416137695
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,1024,4,0,3.189530690511068
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,1024,32,0,0.5524799823760986
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,1024,64,0,0.35771199067433673
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,1024,128,0,0.2699413299560547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,128,512,1,0,40.97907257080078
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,1024,2,0,6.194298426310222
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,1024,16,0,1.7652106285095215
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,1024,8,0,3.266624132792155
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,1024,64,0,0.6401493151982626
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,1024,1,0,12.24014409383138
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,1024,128,0,0.4383680025736491
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,1024,32,0,1.0393653710683186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,1024,4,0,6.253450393676758
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,64,1024,32,0,1.933311939239502
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,1024,2,0,12.256020863850912
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,64,1024,8,0,6.405445098876953
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,64,1024,16,0,3.4504213333129883
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,64,1024,64,0,1.1897599697113037
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,64,1024,128,0,0.795583963394165
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,1536,8,0,0.31376532713572186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,1536,4,0,0.5082613229751587
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,1536,2,0,0.8153813680013021
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,1536,1,0,1.4407092730204265
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,64,1024,4,0,12.445157368977865
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,1536,64,0,0.28858667612075806
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,1024,1,0,24.4343744913737
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,1536,16,0,0.29756800333658856
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,1536,32,0,0.29625600576400757
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,1536,128,0,0.271999994913737
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,1536,2,0,1.4446239471435547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,1536,8,0,0.4532639980316162
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,1536,32,0,0.28942400217056274
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,1536,4,0,0.7906453609466553
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,1536,1,0,2.762666702270508
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,1536,16,0,0.3165760040283203
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,1536,64,0,0.3189653356870015
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,1536,128,0,0.26985599597295123
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,64,1024,2,0,24.64307149251302
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,1536,4,0,1.4777065912882488
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,1536,8,0,0.801093339920044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,1536,64,0,0.2958933313687642
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,1536,128,0,0.27301865816116333
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,1536,2,0,2.769402821858724
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,1536,16,0,0.4660586516062419
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,1536,32,0,0.31041600306828815
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,1536,1,0,5.417317072550456
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,1536,16,0,1.2474613189697266
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,1536,64,0,0.33100799719492596
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,1536,8,0,1.4817813237508137
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,1536,32,0,0.49560534954071045
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,1536,4,0,2.79368527730306
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,1536,128,0,0.27508799235026044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,1536,2,0,5.409055709838867
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,1536,8,0,2.849578539530436
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,1536,32,0,0.8951466878255209
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,1536,16,0,1.5520426432291667
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,1536,4,0,5.465973536173503
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,1536,64,0,0.5621866782506307
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,1536,128,0,0.3847786585489909
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,1536,1,0,10.698944091796875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,1536,2,0,10.745620727539062
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,64,1024,1,0,48.85289510091146
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,1536,32,0,1.6757386525472004
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,1536,8,0,5.576480229695638
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,1536,16,0,2.9789759318033853
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,1536,64,0,1.0277067025502522
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,2048,4,0,0.6413866678873698
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,1536,128,0,0.6799093087514242
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,2048,2,0,1.1416107018788655
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,2048,1,0,2.1403892834981284
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,1536,4,0,10.909236907958984
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,2048,8,0,0.4220000108083089
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,2048,16,0,0.304202675819397
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,2048,64,0,0.28310932715733844
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,2048,128,0,0.2802613377571106
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,2048,32,0,0.3067093292872111
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,1536,1,0,21.288511912027996
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,2048,16,0,0.3955306609471639
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,2048,4,0,1.1507893403371174
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,2048,2,0,2.1444266637166343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,2048,8,0,0.6423093477884928
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,2048,32,0,0.29287999868392944
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,2048,128,0,0.27801066637039185
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,2048,64,0,0.29729066292444867
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,2048,1,0,4.139626820882161
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,2048,16,0,0.6667839686075846
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,1536,2,0,21.455599466959637
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,2048,4,0,2.1592639287312827
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,2048,8,0,1.1654773553212483
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,2048,32,0,0.7896373271942139
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,2048,128,0,0.3083840012550354
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,2048,64,0,0.3009866674741109
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,2048,2,0,4.151007970174153
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,2048,16,0,1.2082666556040447
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,2048,1,0,8.158485412597656
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,2048,8,0,2.1959147453308105
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,2048,32,0,0.707360029220581
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,2048,4,0,4.197946548461914
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,2048,128,0,0.32124799489974976
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,2048,64,0,0.45427199204762775
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,2048,2,0,8.187248229980469
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,2048,8,0,4.255434672037761
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,2048,16,0,2.284133275349935
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,1536,1,0,42.44561004638672
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,2048,1,0,16.173236846923828
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,2048,32,0,1.2916746934254963
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,2048,64,0,0.7949439684549967
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,2048,128,0,0.5320586760838827
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,2048,4,0,8.235408147176107
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,2048,2,0,16.2520268758138
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,32,2048,32,0,2.454064051310221
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,32,2048,16,0,4.434442520141602
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,32,2048,128,0,0.9518933296203613
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,32,2048,8,0,8.424794514973959
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,32,2048,64,0,1.4664799372355144
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,3072,4,0,1.1111626625061035
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,3072,8,0,0.6415466864903768
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,3072,2,0,2.042506694793701
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,3072,16,0,0.4140640099843343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,3072,1,0,3.900639851888021
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,3072,32,0,0.2993013262748718
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,3072,64,0,0.3031839927037557
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,3072,128,0,0.3145866592725118
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,3072,8,0,1.148197333017985
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,3072,4,0,2.0529173215230307
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,32,2048,4,0,16.49211247762044
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,3072,16,0,0.651962677637736
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,3072,2,0,3.9071947733561196
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,3072,32,0,0.4262346824010213
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,3072,64,0,0.30857600768407184
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,2048,1,0,32.25973765055338
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,3072,128,0,0.27448532978693646
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,3072,1,0,7.651871999104817
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,3072,16,0,1.1564586957295735
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,3072,32,0,0.6804107030232748
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,3072,8,0,2.0775465965270996
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,3072,4,0,3.95249080657959
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,3072,64,0,0.45347734292348224
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,3072,128,0,0.32522666454315186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,3072,2,0,7.7048797607421875
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,3072,8,0,4.002821286519368
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,32,2048,2,0,32.40848032633463
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,3072,16,0,2.146224021911621
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,3072,32,0,1.2171839872996013
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,3072,64,0,0.7476906776428223
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,3072,1,0,15.100672403971354
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,3072,128,0,0.5054346720377604
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,3072,4,0,7.701242446899414
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,3072,16,0,4.1212107340494795
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,3072,2,0,15.251115163167318
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,3072,8,0,7.819285074869792
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,3072,32,0,2.2717013359069824
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,3072,64,0,1.3573439915974934
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,3072,128,0,0.8674346605936686
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,4096,4,0,1.699295997619629
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,3072,4,0,15.320714314778646
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,4096,2,0,3.196117401123047
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,4096,8,0,0.9508693218231201
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,4096,16,0,0.585258682568868
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,4096,32,0,0.3936320145924886
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,4096,1,0,6.14674695332845
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,4096,64,0,0.32131733496983844
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,4096,128,0,0.29710932572682697
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,3072,1,0,30.07354736328125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,4096,8,0,1.741663932800293
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,4096,16,0,0.9732320308685303
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,4096,4,0,3.2076905568440757
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,4096,32,0,0.6006826559702555
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,4096,64,0,0.41208000977834064
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,4096,128,0,0.331386665503184
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,32,2048,1,0,64.53254699707031
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,4096,2,0,6.176597595214844
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,4096,8,0,3.23201592763265
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,3072,2,0,30.197952270507812
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,4096,4,0,6.195914586385091
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,4096,32,0,1.0152053038279216
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,4096,16,0,1.7576692899068196
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,4096,64,0,0.6415253480275472
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,4096,128,0,0.4377973477045695
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,4096,1,0,12.151920318603516
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,4096,2,0,12.1419308980306
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,4096,16,0,3.3225599924723306
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,4096,8,0,6.284287770589192
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,4096,32,0,1.8419253031412761
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,4096,64,0,1.1044853528340657
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,4096,128,0,0.726581335067749
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,4096,4,0,12.226431528727213
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,4096,1,0,24.008138020833332
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,3072,1,0,60.023101806640625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,4096,2,0,24.047882080078125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,16,4096,64,0,2.017354647318522
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,16,4096,8,0,12.335866292317709
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,16,4096,32,0,3.490053176879883
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,16,4096,16,0,6.480010350545247
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,16,4096,128,0,1.2597706317901611
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,6144,8,0,1.7671680450439453
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,6144,4,0,3.2604214350382485
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,6144,16,0,1.02237335840861
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,6144,32,0,0.6583946545918783
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,6144,2,0,6.194608052571614
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,6144,64,0,0.443994681040446
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,16,4096,4,0,24.17736053466797
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,6144,128,0,0.41793068250020343
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,6144,1,0,12.104698181152344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,6144,8,0,3.2827253341674805
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,6144,16,0,1.7958079973856609
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,4096,1,0,47.953470865885414
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,6144,4,0,6.2144317626953125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,6144,32,0,1.0570399761199951
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,6144,64,0,0.6863679885864258
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,6144,128,0,0.4623200098673503
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,6144,2,0,12.152299245198568
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,6144,8,0,6.272607803344727
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,6144,16,0,3.336816151936849
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,6144,32,0,1.8603679339090984
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,6144,64,0,1.1191253662109375
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,6144,128,0,0.736128012339274
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,16,4096,2,0,48.036031087239586
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,6144,4,0,12.196671803792318
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,6144,1,0,23.941205342610676
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,6144,16,0,6.40667724609375
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,6144,2,0,24.10346221923828
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,6144,8,0,12.28595225016276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,6144,32,0,3.4604479471842446
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,6144,64,0,2.003701368967692
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,6144,128,0,1.2354933420817058
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,8192,4,0,5.264250755310059
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,6144,4,0,24.187108357747395
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,8192,8,0,2.809722582499186
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,8192,16,0,1.5846346219380696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,8192,2,0,10.226383845011393
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,8192,32,0,0.9684639771779379
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,8192,64,0,0.6453226804733276
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,8192,128,0,0.542410651842753
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,6144,1,0,47.67951456705729
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,8192,1,0,20.023621877034504
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,8192,8,0,5.303669293721517
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,8192,16,0,2.8541386922200522
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,8192,32,0,1.6245333353678386
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,8192,4,0,10.23800532023112
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,8192,64,0,1.0106560389200847
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,8192,128,0,0.6702773571014404
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,16,4096,1,0,95.76326497395833
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,6144,2,0,47.81234232584635
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,8192,2,0,20.05951436360677
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,8192,32,0,2.940826733907064
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,8192,16,0,5.392666498819987
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,8192,8,0,10.31497065226237
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,8192,64,0,1.7131786346435547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,8192,128,0,1.0877280235290527
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,8192,4,0,20.119717915852863
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,8192,1,0,39.702433268229164
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,8,8192,16,0,10.457845052083334
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,8192,2,0,39.74469248453776
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,8,8192,8,0,20.26902898152669
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,8,8192,32,0,5.55891227722168
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,8,8192,64,0,3.1112534205118814
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,8,8192,128,0,1.8683733940124512
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,6144,1,0,95.25083414713542
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,8,8192,4,0,39.88197326660156
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,10240,8,0,4.104144096374512
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,10240,4,0,7.781007766723633
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,10240,32,0,1.328869342803955
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,10240,16,0,2.2656853993733725
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,10240,64,0,0.8684960206349691
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,10240,128,0,0.6583146651585897
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,10240,2,0,15.139423370361328
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,10240,8,0,7.826735814412435
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,10240,1,0,29.895957946777344
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,10240,4,0,15.18947728474935
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,10240,16,0,4.160773277282715
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,10240,32,0,2.4839305877685547
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,10240,64,0,1.38372802734375
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,8192,1,0,79.3839619954427
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,10240,128,0,0.9061546325683594
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,10240,2,0,29.909952799479168
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,8,8192,2,0,79.36392720540364
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,10240,16,0,7.943109512329102
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,10240,8,0,15.305285135904947
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,10240,32,0,4.268202781677246
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,10240,64,0,2.4268693923950195
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,10240,128,0,1.4915199279785156
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,10240,4,0,29.988372802734375
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,12288,8,0,5.631610870361328
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,10240,1,0,59.37737528483073
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,12288,16,0,3.0660266876220703
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,12288,32,0,1.772213300069173
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,12288,4,0,10.787775675455729
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,12288,128,0,0.7709120114644369
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,12288,64,0,1.134778658548991
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,12288,2,0,21.062048594156902
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,10240,2,0,59.627471923828125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,12288,8,0,10.830384572347006
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,12288,16,0,5.716421127319336
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,12288,4,0,21.11788813273112
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,12288,32,0,3.1302881240844727
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,12288,64,0,1.8436212539672852
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,12288,128,0,1.1843413511912029
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,12288,1,0,41.72265116373698
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,12288,2,0,41.79394022623698
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,8,8192,1,0,158.20519002278647
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,12288,8,0,21.25011698404948
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,12288,32,0,5.837535858154297
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,12288,64,0,3.2630348205566406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,12288,16,0,10.965493520100912
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,12288,128,0,1.9612107276916504
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,12288,4,0,41.91605885823568
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,10240,1,0,118.4893086751302
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,16384,4,0,18.270768483479817
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,16384,8,0,9.423818588256836
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,16384,16,0,5.034480094909668
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,16384,32,0,2.8278398513793945
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,16384,64,0,1.7181493441263835
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,16384,128,0,1.1455093224843342
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,12288,1,0,82.95060221354167
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,16384,2,0,35.889381408691406
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,16384,16,0,9.517199834187826
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,16384,8,0,18.307540893554688
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,16384,32,0,5.116319974263509
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,16384,64,0,2.9133386611938477
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,12288,2,0,83.31752522786458
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,16384,128,0,1.8182346026102703
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,16384,4,0,35.93974812825521
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,16384,1,0,71.21741739908855
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,4,16384,16,0,18.50006357828776
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,16384,2,0,71.20437113444011
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,4,16384,8,0,36.08922576904297
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,4,16384,32,0,9.694565455118815
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,4,16384,64,0,5.287226676940918
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,4,16384,128,0,3.108880043029785
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,4,16384,4,0,71.28616333007812
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,12288,1,0,165.56788126627603
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,1,32768,16,0,17.712655385335285
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,1,32768,8,0,34.29717763264974
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,1,32768,32,0,9.400837580362955
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,1,32768,64,0,5.236181259155273
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,1,32768,128,0,3.183664004007975
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,16384,1,0,142.03288777669272
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,1,32768,4,0,67.54647318522136
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,4,16384,2,0,142.32445271809897
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,8,2,32768,16,0,34.568651835123696
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,1,32768,2,0,133.97852579752603
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,16,2,32768,8,0,67.69506327311198
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,1,2,32768,128,0,5.396634419759114
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,4,2,32768,32,0,17.921802520751953
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,2,2,32768,64,0,9.568842569986979
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,32,2,32768,4,0,134.44140625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,4,16384,1,0,282.9903564453125
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,1,32768,1,0,267.1160074869792
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,64,2,32768,2,0,267.224853515625
VLLM,0.14.0,NVIDIA A100-SXM4-80GB,context_mla,vllm_triton_mla,float16,float16,128,2,32768,1,0,532.953857421875
