framework,version,device,op_name,kernel_source,bmm_dtype,num_tokens,num_heads,latency
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,128,0.0065760001540184024
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,128,0.010569600015878677
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,64,0.004524800181388855
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,64,0.008582399785518646
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,32,0.004524800181388855
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,32,0.006627199798822403
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,16,0.004294399917125702
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,16,0.006617599725723266
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,8,0.004278400167822838
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,8,0.008451200276613235
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,4,0.004614400118589402
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,4,0.007526399940252304
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,2,0.004438399896025657
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,2,0.008463999629020691
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1,1,0.004479999840259552
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1,1,0.007334399968385696
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,128,0.006620799750089645
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,128,0.010678400099277497
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,64,0.004553600028157234
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,64,0.008476799726486206
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,32,0.0045056000351905824
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,32,0.008463999629020691
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,16,0.004377600178122521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,16,0.006585600227117539
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,8,0.004499199986457825
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,8,0.0066592000424861904
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,4,0.00453759990632534
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,4,0.007680000364780426
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,2,0.004508800059556961
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,2,0.008451200276613235
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2,1,0.004560000076889992
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2,1,0.007705599814653396
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,128,0.006499200314283371
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,128,0.01050880029797554
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,64,0.004518400132656098
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,64,0.008495999872684479
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,32,0.004556800052523613
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,32,0.006678400188684463
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,16,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,16,0.006665600091218948
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,8,0.004416000097990036
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,8,0.006691200286149978
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,4,0.004419200122356415
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,4,0.006700800359249115
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,2,0.004339199885725975
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,2,0.008486399799585343
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4,1,0.0044064000248909
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4,1,0.007487999647855759
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,128,0.006623999774456024
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,128,0.010764800012111664
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,64,0.0043935999274253845
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,64,0.008508799970149994
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,32,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,32,0.008483199775218964
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,16,0.004553600028157234
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,16,0.0066431999206542965
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,8,0.004355200007557869
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,8,0.007516799867153168
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,4,0.004185599833726883
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,4,0.006662400066852569
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,2,0.004521600157022476
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,2,0.006540799885988236
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8,1,0.004310400038957596
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8,1,0.008511999994516373
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,128,0.006672000139951706
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,128,0.012604799866676331
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,64,0.004527999833226204
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,64,0.008713600039482117
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,32,0.004403200000524521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,32,0.00865280032157898
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,16,0.004377600178122521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,16,0.00851840004324913
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,8,0.004473600164055824
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,8,0.006601600348949433
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,4,0.004412800073623657
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,4,0.007023999840021134
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,2,0.004412800073623657
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,2,0.006691200286149978
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,16,1,0.0044064000248909
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,16,1,0.006604799628257751
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,128,0.007379200309514999
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,128,0.014636799693107605
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,64,0.00459199994802475
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,64,0.010527999699115753
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,32,0.0045311998575925825
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,32,0.00862400010228157
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,16,0.004620800167322159
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,16,0.008537600189447403
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,8,0.0044319998472929
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,8,0.008483199775218964
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,4,0.004342399910092354
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,4,0.008457600325345992
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,2,0.004073600098490715
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,2,0.006892800331115723
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,32,1,0.004476799815893173
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,32,1,0.006649599969387054
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,128,0.00846719965338707
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,128,0.018918399512767792
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,64,0.004665600135922432
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,64,0.012697599828243256
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,32,0.004630399867892265
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,32,0.01056319996714592
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,16,0.0044064000248909
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,16,0.008476799726486206
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,8,0.004476799815893173
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,8,0.008451200276613235
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,4,0.0047680001705884935
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,4,0.008457600325345992
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,2,0.004342399910092354
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,2,0.008483199775218964
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,48,1,0.004332799836993218
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,48,1,0.006684800237417221
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,128,0.006684800237417221
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,128,0.020803199708461763
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,64,0.0046016000211238865
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,64,0.012780800461769104
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,32,0.0046016000211238865
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,32,0.010496000200510025
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,16,0.005158400163054466
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,16,0.008665599673986436
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,8,0.004470400139689445
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,8,0.008479999750852585
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,4,0.004454400017857551
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,4,0.006636799871921539
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,2,0.004329600185155868
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,2,0.008470399677753449
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,64,1,0.00445760004222393
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,64,1,0.00687360018491745
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,128,0.007430399954319
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,128,0.02285760045051575
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,64,0.0064640000462532045
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,64,0.014681600034236908
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,32,0.004595199972391129
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,32,0.012166400253772736
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,16,0.004444799944758415
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,16,0.008739200234413148
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,8,0.004473600164055824
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,8,0.0086496002972126
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,4,0.004470400139689445
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,4,0.008483199775218964
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,2,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,2,0.008505599945783615
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,80,1,0.004464000090956688
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,80,1,0.006604799628257751
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,128,0.007648000121116638
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,128,0.02502720057964325
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,64,0.0064351998269557955
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,64,0.01666560024023056
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,32,0.005635200068354607
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,32,0.012652799487113953
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,16,0.004483199864625931
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,16,0.008691199868917466
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,8,0.0044895999133586885
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,8,0.00862400010228157
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,4,0.004464000090956688
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,4,0.00856959968805313
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,2,0.004416000097990036
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,2,0.006601600348949433
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,96,1,0.004291199892759323
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,96,1,0.006611199676990509
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,128,0.008713600039482117
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,128,0.031004801392555237
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,64,0.006508799642324448
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,64,0.01889919936656952
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,32,0.006505600363016129
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,32,0.012729600071907043
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,16,0.004396799951791763
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,16,0.010496000200510025
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,8,0.004467200115323066
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,8,0.008659200370311737
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,4,0.004419200122356415
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,4,0.008511999994516373
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,2,0.004441599920392036
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,2,0.006630399823188781
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,128,1,0.0044895999133586885
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,128,1,0.00852160006761551
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,128,0.008687999844551087
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,128,0.03573119938373566
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,64,0.006486400216817856
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,64,0.021401600539684297
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,32,0.006406400352716446
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,32,0.014825600385665893
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,16,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,16,0.012243200093507767
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,8,0.004447999969124794
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,8,0.008716800063848496
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,4,0.004364800080657006
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,4,0.008515200018882752
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,2,0.004495999962091446
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,2,0.008499199897050858
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,160,1,0.004287999868392944
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,160,1,0.00843840017914772
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,128,0.010777600109577179
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,128,0.041222399473190306
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,64,0.006665600091218948
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,64,0.0248416006565094
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,32,0.006585600227117539
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,32,0.016752000153064727
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,16,0.004387199878692627
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,16,0.012643200159072877
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,8,0.0044511999934911724
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,8,0.008752000331878663
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,4,0.00448639988899231
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,4,0.00870719999074936
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,2,0.004339199885725975
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,2,0.008502399921417237
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,192,1,0.0044319998472929
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,192,1,0.006700800359249115
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,128,0.010911999642848969
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,128,0.05162879824638367
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,64,0.008473599702119828
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,64,0.031027200818061828
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,32,0.00663359984755516
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,32,0.018908800184726716
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,16,0.004614400118589402
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,16,0.012796799838542938
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,8,0.00448639988899231
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,8,0.010438399761915207
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,4,0.004575999826192856
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,4,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,2,0.004291199892759323
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,2,0.008534400165081025
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,256,1,0.00445760004222393
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,256,1,0.006681600213050842
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,128,0.012748800218105316
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,128,0.06369919776916504
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,64,0.008550400286912918
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,64,0.03612479865550995
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,32,0.006492800265550614
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,32,0.021401600539684297
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,16,0.006207999959588051
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,16,0.014793600142002105
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,8,0.004575999826192856
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,8,0.010694400221109391
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,4,0.004380799829959869
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,4,0.008614400029182434
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,2,0.004508800059556961
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,2,0.00854720026254654
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,320,1,0.0044319998472929
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,320,1,0.008489599823951722
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,128,0.014803199470043183
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,128,0.0740447998046875
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,64,0.010496000200510025
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,64,0.04122560024261475
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,32,0.006652799993753433
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,32,0.024889600276947022
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,16,0.006483200192451477
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,16,0.016841599345207216
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,8,0.004608000069856644
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,8,0.012595200538635254
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,4,0.004572800174355507
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,4,0.008825600147247314
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,2,0.004659200087189674
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,2,0.008534400165081025
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,384,1,0.0044511999934911724
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,384,1,0.008489599823951722
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,128,0.01863040030002594
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,128,0.0985152006149292
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,64,0.012643200159072877
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,64,0.051520001888275144
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,32,0.008483199775218964
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,32,0.03107840120792389
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,16,0.006409599632024765
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,16,0.018960000574588777
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,8,0.004495999962091446
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,8,0.012748800218105316
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,4,0.0044319998472929
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,4,0.010569600015878677
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,2,0.00453759990632534
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,2,0.008659200370311737
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,512,1,0.004444799944758415
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,512,1,0.008508799970149994
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,128,0.027356800436973572
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,128,0.1430367946624756
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,64,0.014611199498176575
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,64,0.07397440075874329
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,32,0.01072319969534874
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,32,0.041545599699020386
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,16,0.008537600189447403
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,16,0.025043201446533204
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,8,0.0064640000462532045
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,8,0.01666879951953888
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,4,0.00647680014371872
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,4,0.01257600039243698
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,2,0.004441599920392036
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,2,0.010502400249242783
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,768,1,0.004419200122356415
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,768,1,0.008668799698352814
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,128,0.03534719944000244
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,128,0.1888576030731201
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,64,0.018764799833297728
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,64,0.09750720262527465
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,32,0.010729599744081497
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,32,0.051635199785232545
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,16,0.008454400300979614
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,16,0.031174400448799135
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,8,0.006579200178384781
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,8,0.018729600310325622
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,4,0.004630399867892265
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,4,0.012742400169372559
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,2,0.004470400139689445
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,2,0.010527999699115753
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1024,1,0.004358400031924248
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1024,1,0.008640000224113464
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,128,0.04743359982967377
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,128,0.2793600082397461
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,64,0.02686080038547516
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,64,0.14277759790420533
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,32,0.014611199498176575
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,32,0.07417920231819153
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,16,0.010716799646615982
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,16,0.04127359986305237
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,8,0.00867839977145195
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,8,0.0248416006565094
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,4,0.00642239972949028
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,4,0.01681919991970062
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,2,0.006451199948787689
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,2,0.012646399438381195
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,1536,1,0.004361600056290627
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,1536,1,0.008928000181913375
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,128,0.06177279949188232
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,128,0.3703871965408325
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,64,0.034143999218940735
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,64,0.18874560594558715
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,32,0.018768000602722167
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,32,0.09673280119895936
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,16,0.010691200196743012
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,16,0.05171520113945007
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,8,0.008511999994516373
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,8,0.03094080090522766
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,4,0.006447999924421311
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,4,0.018911999464035035
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,2,0.004566400125622749
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,2,0.012803199887275695
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,2048,1,0.004566400125622749
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,2048,1,0.010710400342941285
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,128,0.08770560026168824
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,128,0.5506239891052246
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,64,0.04742400050163269
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,64,0.27877440452575686
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,32,0.026998400688171387
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,32,0.141593599319458
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,16,0.014710399508476257
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,16,0.07411839962005615
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,8,0.010732799768447876
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,8,0.041254401206970215
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,4,0.008710400015115739
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,4,0.02484800070524216
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,2,0.006345599889755249
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,2,0.01669439971446991
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,3072,1,0.004604800045490265
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,3072,1,0.012566399574279786
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,128,0.11305919885635377
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,128,0.7328800201416016
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,64,0.06121600270271301
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,64,0.3693631887435913
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,32,0.03488639891147614
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,32,0.18867199420928954
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,16,0.017868800461292265
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,16,0.09696639776229858
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,8,0.010540799796581268
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,8,0.0516319990158081
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,4,0.008534400165081025
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,4,0.03118399977684021
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,2,0.006467200070619583
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,2,0.018806399405002595
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,4096,1,0.0046271998435258865
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,4096,1,0.01281599998474121
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,128,0.1647744059562683
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,128,1.1062591552734375
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,64,0.08746560215950012
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,64,0.5502848148345947
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,32,0.04661119878292084
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,32,0.27850239276885985
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,16,0.025119999051094057
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,16,0.1416640043258667
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,8,0.014550399780273438
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,8,0.07396799921989441
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,4,0.010784000158309937
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,4,0.043100801110267636
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,2,0.008662399649620057
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,2,0.024899199604988098
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,6144,1,0.006428799778223038
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,6144,1,0.016710400581359863
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,128,0.21689279079437257
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,128,1.4580896377563477
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,64,0.1136512041091919
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,64,0.7322175979614258
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,32,0.06179839968681335
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,32,0.3696127891540527
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,16,0.03446399867534637
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,16,0.18881280422210694
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,8,0.01687999963760376
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,8,0.09689279794692993
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,4,0.010742399841547012
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,4,0.05167040228843689
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,2,0.008531200140714646
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,2,0.031017601490020752
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,float16,8192,1,0.00639680027961731
SGLang,0.5.10,NVIDIA B200,mla_gen_pre,default,fp8,8192,1,0.01876160055398941
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,128,0.004748800024390221
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,128,0.008575999736785888
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,64,0.004521600157022476
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,64,0.008473599702119828
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,32,0.004560000076889992
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,32,0.00857279971241951
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,16,0.004464000090956688
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,16,0.008534400165081025
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,8,0.004595199972391129
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,8,0.008582399785518646
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,4,0.0045311998575925825
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,4,0.006703999638557434
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,2,0.004521600157022476
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,2,0.008508799970149994
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1,1,0.004556800052523613
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1,1,0.00687360018491745
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,128,0.0045855998992919925
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,128,0.008489599823951722
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,64,0.0044895999133586885
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,64,0.008560000360012055
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,32,0.004716800153255462
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,32,0.008553600311279297
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,16,0.004639999940991402
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,16,0.008566399663686752
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,8,0.004620800167322159
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,8,0.008473599702119828
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,4,0.0044895999133586885
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,4,0.00663359984755516
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,2,0.004524800181388855
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,2,0.006678400188684463
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2,1,0.004444799944758415
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2,1,0.008531200140714646
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,128,0.004694399982690811
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,128,0.008601599931716919
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,64,0.005161599814891815
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,64,0.008550400286912918
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,32,0.0046271998435258865
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,32,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,16,0.005065599828958512
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,16,0.008479999750852585
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,8,0.004527999833226204
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,8,0.008479999750852585
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,4,0.0044511999934911724
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,4,0.006726399809122085
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,2,0.004553600028157234
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,2,0.00857279971241951
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4,1,0.004444799944758415
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4,1,0.00663359984755516
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,128,0.004662400111556053
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,128,0.008713600039482117
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,64,0.004540799930691719
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,64,0.008582399785518646
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,32,0.00461760014295578
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,32,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,16,0.004639999940991402
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,16,0.008460800349712371
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,8,0.004521600157022476
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,8,0.008579199761152267
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,4,0.00453759990632534
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,4,0.008607999980449676
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,2,0.004598399996757508
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,2,0.00854720026254654
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8,1,0.004339199885725975
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8,1,0.008115199953317642
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,128,0.006496000289916992
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,128,0.01061440035700798
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,64,0.004483199864625931
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,64,0.008703999966382981
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,32,0.004447999969124794
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,32,0.008495999872684479
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,16,0.004492799937725067
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,16,0.008595199882984161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,8,0.004710400104522705
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,8,0.008537600189447403
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,4,0.004476799815893173
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,4,0.008560000360012055
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,2,0.004441599920392036
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,2,0.006668800115585327
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16,1,0.004444799944758415
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16,1,0.007916799932718276
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,128,0.0065151996910572055
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,128,0.012681600451469422
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,64,0.0046271998435258865
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,64,0.010655999928712846
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,32,0.00456320010125637
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,32,0.008591999858617782
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,16,0.004460800066590309
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,16,0.008531200140714646
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,8,0.004495999962091446
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,8,0.008489599823951722
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,4,0.004441599920392036
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,4,0.00854720026254654
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,2,0.00448639988899231
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,2,0.0084927998483181
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,32,1,0.004534399881958961
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,32,1,0.006716799736022949
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,128,0.006492800265550614
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,128,0.016681599617004394
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,64,0.004608000069856644
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,64,0.010966400057077408
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,32,0.004518400132656098
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,32,0.010540799796581268
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,16,0.00445760004222393
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,16,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,8,0.004447999969124794
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,8,0.00846719965338707
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,4,0.004521600157022476
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,4,0.00856959968805313
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,2,0.004633599892258644
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,2,0.00857279971241951
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,48,1,0.004438399896025657
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,48,1,0.006729599833488464
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,128,0.0066592000424861904
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,128,0.01879040002822876
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,64,0.004633599892258644
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,64,0.012758399546146392
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,32,0.004608000069856644
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,32,0.01051200032234192
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,16,0.004470400139689445
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,16,0.008643200248479843
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,8,0.004515200108289719
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,8,0.008556800335645676
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,4,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,4,0.00854720026254654
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,2,0.004473600164055824
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,2,0.007091200351715088
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,64,1,0.004694399982690811
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,64,1,0.00846719965338707
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,128,0.006508799642324448
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,128,0.020771199464797975
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,64,0.0046016000211238865
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,64,0.014640000462532044
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,32,0.00461760014295578
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,32,0.010611200332641601
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,16,0.004630399867892265
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,16,0.008710400015115739
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,8,0.00461760014295578
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,8,0.008761599659919739
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,4,0.004454400017857551
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,4,0.008630400151014328
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,2,0.004495999962091446
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,2,0.00852160006761551
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,80,1,0.0045311998575925825
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,80,1,0.006649599969387054
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,128,0.0066431999206542965
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,128,0.02306559979915619
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,64,0.004662400111556053
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,64,0.015321600437164306
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,32,0.004598399996757508
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,32,0.010780800133943558
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,16,0.0046431999653577805
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,16,0.010284800082445145
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,8,0.005340800061821938
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,8,0.008550400286912918
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,4,0.0045056000351905824
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,4,0.008486399799585343
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,2,0.00451200008392334
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,2,0.00851840004324913
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,96,1,0.004499199986457825
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,96,1,0.006662400066852569
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,128,0.006703999638557434
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,128,0.028992000222206115
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,64,0.006534399837255478
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,64,0.018777599930763243
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,32,0.004668800160288811
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,32,0.012668800354003907
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,16,0.004560000076889992
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,16,0.01066880002617836
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,8,0.004659200087189674
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,8,0.008511999994516373
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,4,0.005014400184154511
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,4,0.008556800335645676
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,2,0.00445760004222393
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,2,0.008591999858617782
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,128,1,0.004755200073122978
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,128,1,0.00859839990735054
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,128,0.006748799979686737
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,128,0.03320960104465485
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,64,0.006441599875688553
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,64,0.02083519995212555
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,32,0.004630399867892265
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,32,0.014703999459743499
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,16,0.004633599892258644
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,16,0.010969600081443787
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,8,0.004691199958324432
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,8,0.008790399879217148
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,4,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,4,0.00856959968805313
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,2,0.00448639988899231
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,2,0.008528000116348267
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,160,1,0.004966399818658829
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,160,1,0.008463999629020691
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,128,0.008515200018882752
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,128,0.03919680118560791
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,64,0.006460800021886826
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,64,0.02295999974012375
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,32,0.0046720001846551895
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,32,0.014895999431610107
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,16,0.004659200087189674
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,16,0.012624000012874604
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,8,0.004524800181388855
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,8,0.00878399983048439
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,4,0.004521600157022476
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,4,0.008550400286912918
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,2,0.004515200108289719
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,2,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,192,1,0.004422400146722794
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,192,1,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,128,0.008524800091981888
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,128,0.04967679977416992
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,64,0.006467200070619583
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,64,0.028937599062919615
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,32,0.006470400094985962
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,32,0.018825599551200868
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,16,0.00461760014295578
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,16,0.012812800705432892
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,8,0.004454400017857551
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,8,0.010569600015878677
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,4,0.0045855998992919925
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,4,0.008550400286912918
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,2,0.004467200115323066
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,2,0.00857279971241951
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,256,1,0.0044096000492572784
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,256,1,0.00852160006761551
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,128,0.008755200356245042
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,128,0.06155520081520081
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,64,0.006553599983453751
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,64,0.033155199885368344
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,32,0.006400000303983688
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,32,0.020870399475097657
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,16,0.004639999940991402
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,16,0.014860799908638
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,8,0.004428799822926521
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,8,0.010841599851846694
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,4,0.0044351998716592785
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,4,0.00873280018568039
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,2,0.004419200122356415
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,2,0.00854720026254654
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,320,1,0.004447999969124794
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,320,1,0.008537600189447403
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,128,0.010652799904346467
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,128,0.07405440211296081
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,64,0.006620799750089645
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,64,0.03922879993915558
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,32,0.006540799885988236
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,32,0.022924800217151643
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,16,0.004598399996757508
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,16,0.016681599617004394
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,8,0.004598399996757508
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,8,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,4,0.004636799916625023
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,4,0.009062399715185165
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,2,0.004540799930691719
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,2,0.008665599673986436
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,384,1,0.004412800073623657
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,384,1,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,128,0.01465280055999756
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,128,0.09689599871635438
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,64,0.008556800335645676
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,64,0.04962559938430786
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,32,0.006691200286149978
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,32,0.02897599935531616
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,16,0.004675199836492538
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,16,0.018742400407791137
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,8,0.004630399867892265
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,8,0.012726399302482604
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,4,0.0046847999095916745
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,4,0.01053759977221489
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,2,0.004569600149989128
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,2,0.008703999966382981
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,512,1,0.0046271998435258865
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,512,1,0.008582399785518646
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,128,0.027215999364852906
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,128,0.1436095952987671
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,64,0.010550399869680404
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,64,0.07248640060424805
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,32,0.008544000238180161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,32,0.03925760090351105
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,16,0.006447999924421311
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,16,0.02290560007095337
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,8,0.004579199850559235
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,8,0.014828799664974213
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,4,0.004748800024390221
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,4,0.01138240024447441
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,2,0.0044895999133586885
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,2,0.00995519980788231
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,768,1,0.0044895999133586885
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,768,1,0.008515200018882752
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,128,0.035129600763320924
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,128,0.1888543963432312
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,64,0.01281599998474121
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,64,0.09662079811096191
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,32,0.008742400258779526
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,32,0.049667200446128844
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,16,0.0066880002617835995
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,16,0.029187199473381043
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,8,0.0065600000321865085
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,8,0.018748800456523895
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,4,0.005040000006556511
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,4,0.012787200510501862
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,2,0.004556800052523613
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,2,0.010627199709415436
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1024,1,0.0045855998992919925
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1024,1,0.008566399663686752
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,128,0.05076479911804199
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,128,0.2794303894042969
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,64,0.027004799246788024
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,64,0.14188799858093262
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,32,0.010566399991512298
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,32,0.07219840288162231
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,16,0.008479999750852585
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,16,0.03936960101127625
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,8,0.006444799900054932
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,8,0.022912000119686127
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,4,0.004675199836492538
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,4,0.014864000678062438
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,2,0.004604800045490265
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,2,0.01080000028014183
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,1536,1,0.004579199850559235
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,1536,1,0.009785600006580353
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,128,0.0685151994228363
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,128,0.3706975936889648
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,64,0.03557760119438171
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,64,0.18685760498046874
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,32,0.012639999389648438
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,32,0.09658880233764648
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,16,0.008697599917650223
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,16,0.049609598517417905
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,8,0.006639999896287918
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,8,0.028998398780822755
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,4,0.0063008002936840056
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,4,0.018825599551200868
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,2,0.004623999819159508
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,2,0.012825599312782288
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,2048,1,0.004566400125622749
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,2048,1,0.010595200210809707
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,128,0.09055039882659913
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,128,0.558457612991333
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,64,0.049807998538017276
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,64,0.27767999172210694
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,32,0.02500160038471222
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,32,0.1416095972061157
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,16,0.01061440035700798
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,16,0.07197120189666747
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,8,0.008553600311279297
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,8,0.039392000436782836
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,4,0.006694400310516357
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,4,0.023049600422382355
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,2,0.00461760014295578
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,2,0.015801599621772765
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,3072,1,0.004553600028157234
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,3072,1,0.010976000130176545
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,128,0.11920319795608521
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,128,0.7630943775177002
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,64,0.06765440106391907
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,64,0.3699104070663452
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,32,0.031088000535964964
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,32,0.18752959966659546
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,16,0.012800000607967377
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,16,0.09680320024490356
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,8,0.00873280018568039
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,8,0.049481600522994995
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,4,0.00663359984755516
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,4,0.0290367990732193
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,2,0.006431999802589417
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,2,0.018751999735832213
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,4096,1,0.0046271998435258865
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,4096,1,0.012851199507713318
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,128,0.18199679851531983
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,128,1.2087488174438477
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,64,0.08791999816894532
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,64,0.5569727897644043
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,32,0.04539200067520142
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,32,0.27924480438232424
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,16,0.024873599410057068
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,16,0.1418303966522217
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,8,0.010592000186443329
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,8,0.07240960001945496
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,4,0.008489599823951722
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,4,0.03922879993915558
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,2,0.006454399973154068
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,2,0.023158399760723113
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,6144,1,0.004966399818658829
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,6144,1,0.014847999811172486
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,128,0.23002560138702394
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,128,1.6492864608764648
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,64,0.11405119895935059
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,64,0.7651711940765381
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,32,0.06200640201568604
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,32,0.37075200080871584
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,16,0.031036800146102904
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,16,0.1879647970199585
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,8,0.012601600587368011
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,8,0.09655359983444214
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,4,0.008691199868917466
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,4,0.04969600141048432
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,2,0.00663359984755516
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,2,0.02898559868335724
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,8192,1,0.00647680014371872
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,8192,1,0.018723200261592864
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,128,0.3457920074462891
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,128,2.552892875671387
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,64,0.17227519750595094
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,64,1.2068960189819335
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,32,0.0830847978591919
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,32,0.563212776184082
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,16,0.045731198787689206
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,16,0.2797152042388916
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,8,0.024931199848651886
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,8,0.1418303966522217
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,4,0.010620799660682679
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,4,0.07219840288162231
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,2,0.008550400286912918
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,2,0.03936319947242737
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,12288,1,0.006527999788522721
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,12288,1,0.02306559979915619
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,128,0.460313606262207
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,128,3.396963119506836
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,64,0.22793920040130616
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,64,1.6653343200683595
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,32,0.11521919965744018
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,32,0.775219202041626
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,16,0.06181439757347107
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,16,0.3703263998031616
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,8,0.03110400140285492
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,8,0.1871135950088501
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,4,0.012639999389648438
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,4,0.0965503990650177
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,2,0.008710400015115739
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,2,0.04951359927654266
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,16384,1,0.006617599725723266
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,16384,1,0.029177600145339967
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,128,0.5737215995788574
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,128,4.261315155029297
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,64,0.2753567934036255
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,64,2.1012447357177733
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,32,0.14132800102233886
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,32,0.9912063598632812
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,16,0.07448319792747497
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,16,0.46726398468017577
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,8,0.039113599061965945
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,8,0.23354880809783934
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,4,0.018799999356269838
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,4,0.11923199892044067
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,2,0.008595199882984161
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,2,0.06034240126609802
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,float16,20480,1,0.006553599983453751
SGLang,0.5.10,NVIDIA B200,mla_gen_post,default,fp8,20480,1,0.03326080143451691
