Files
aituner/runs/simulator-tuning-latency-matrix-v0/qwen30-tracepd-u0p01-real-audit.json

1363 lines
43 KiB
JSON

{
"configs": {
"tp1_mns16": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 87355.16047904849,
"pooled_p90_ms": 156177.17167094816,
"pooled_samples": 387,
"trial_mean_of_means_ms": 87355.1604790485,
"trial_mean_of_p90s_ms": 156479.69853299824,
"trial_stddev_of_means_ms": 411.9489877870239
},
"tpot_ms": {
"pooled_mean_ms": 20.58492915882586,
"pooled_p90_ms": 25.316827059621797,
"pooled_samples": 387,
"trial_mean_of_means_ms": 20.58492915882586,
"trial_mean_of_p90s_ms": 25.430729349242615,
"trial_stddev_of_means_ms": 0.07031617503255957
},
"ttft_ms": {
"pooled_mean_ms": 17378.17525871318,
"pooled_p90_ms": 42739.809082006104,
"pooled_samples": 387,
"trial_mean_of_means_ms": 17378.17525871318,
"trial_mean_of_p90s_ms": 43264.360885717906,
"trial_stddev_of_means_ms": 247.95865334755402
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 87177.4532565533,
"p90_ms": 156177.17167094816,
"samples": 129
},
"tpot_ms": {
"mean_ms": 20.657075003239317,
"p90_ms": 25.51118326097509,
"samples": 129
},
"ttft_ms": {
"mean_ms": 17192.74773496445,
"p90_ms": 43084.54651210923,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns16/trial1/results/result.json",
"tp": 1,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 87061.89228269603,
"p90_ms": 156085.63073002733,
"samples": 129
},
"tpot_ms": {
"mean_ms": 20.58111452736139,
"p90_ms": 25.316827059621797,
"samples": 129
},
"ttft_ms": {
"mean_ms": 17281.95525217321,
"p90_ms": 42669.12407299969,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns16/trial2/results/result.json",
"tp": 1,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 87826.13589789615,
"p90_ms": 157176.29319801927,
"samples": 129
},
"tpot_ms": {
"mean_ms": 20.51659794587688,
"p90_ms": 25.464177727130956,
"samples": 129
},
"ttft_ms": {
"mean_ms": 17659.822789001886,
"p90_ms": 44039.4120720448,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns16/trial3/results/result.json",
"tp": 1,
"trial": 3
}
]
},
"tp1_mns32": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 81051.19884917597,
"pooled_p90_ms": 162908.61000702716,
"pooled_samples": 387,
"trial_mean_of_means_ms": 81051.19884917597,
"trial_mean_of_p90s_ms": 163725.26491332488,
"trial_stddev_of_means_ms": 957.6364127854422
},
"tpot_ms": {
"pooled_mean_ms": 23.75712112165005,
"pooled_p90_ms": 29.99736783063116,
"pooled_samples": 387,
"trial_mean_of_means_ms": 23.757121121650055,
"trial_mean_of_p90s_ms": 29.953513044613036,
"trial_stddev_of_means_ms": 0.28945138797341263
},
"ttft_ms": {
"pooled_mean_ms": 620.4562074374239,
"pooled_p90_ms": 1931.474626995623,
"pooled_samples": 387,
"trial_mean_of_means_ms": 620.4562074374239,
"trial_mean_of_p90s_ms": 1952.5051703676581,
"trial_stddev_of_means_ms": 4.034284336592585
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 82155.01919551753,
"p90_ms": 165321.14125497174,
"samples": 129
},
"tpot_ms": {
"mean_ms": 24.088935365685685,
"p90_ms": 30.620627052829327,
"samples": 129
},
"ttft_ms": {
"mean_ms": 624.0473438989255,
"p90_ms": 1953.5144560504705,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns32/trial1/results/result.json",
"tp": 1,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 80442.25376262421,
"p90_ms": 162710.11894091498,
"samples": 129
},
"tpot_ms": {
"mean_ms": 23.556478044326088,
"p90_ms": 29.607802025608333,
"samples": 129
},
"ttft_ms": {
"mean_ms": 616.0909594027614,
"p90_ms": 1931.474626995623,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns32/trial2/results/result.json",
"tp": 1,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 80556.32358938617,
"p90_ms": 163144.53454408795,
"samples": 129
},
"tpot_ms": {
"mean_ms": 23.62594995493838,
"p90_ms": 29.632110055401448,
"samples": 129
},
"ttft_ms": {
"mean_ms": 621.2303190105847,
"p90_ms": 1972.5264280568808,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns32/trial3/results/result.json",
"tp": 1,
"trial": 3
}
]
},
"tp1_mns64": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 82253.7312629709,
"pooled_p90_ms": 165355.2623600699,
"pooled_samples": 387,
"trial_mean_of_means_ms": 82253.73126297089,
"trial_mean_of_p90s_ms": 166168.38832168528,
"trial_stddev_of_means_ms": 276.04080210780324
},
"tpot_ms": {
"pooled_mean_ms": 24.114592194742286,
"pooled_p90_ms": 30.355145227612027,
"pooled_samples": 387,
"trial_mean_of_means_ms": 24.114592194742286,
"trial_mean_of_p90s_ms": 30.441765467132914,
"trial_stddev_of_means_ms": 0.09945812871448045
},
"ttft_ms": {
"pooled_mean_ms": 619.7960944975623,
"pooled_p90_ms": 1956.7296199966222,
"pooled_samples": 387,
"trial_mean_of_means_ms": 619.7960944975623,
"trial_mean_of_p90s_ms": 1975.1493813625227,
"trial_stddev_of_means_ms": 3.103486103420667
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 82496.79665562836,
"p90_ms": 166090.5934929615,
"samples": 129
},
"tpot_ms": {
"mean_ms": 24.210340237338897,
"p90_ms": 30.310964708207926,
"samples": 129
},
"ttft_ms": {
"mean_ms": 616.759402438269,
"p90_ms": 1956.7296199966222,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns64/trial1/results/result.json",
"tp": 1,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 81953.6266326841,
"p90_ms": 165355.2623600699,
"samples": 129
},
"tpot_ms": {
"mean_ms": 24.011798653948865,
"p90_ms": 30.622402682139015,
"samples": 129
},
"ttft_ms": {
"mean_ms": 622.9623185557335,
"p90_ms": 1988.6804190464318,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns64/trial2/results/result.json",
"tp": 1,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 82310.77050060022,
"p90_ms": 167059.3091120245,
"samples": 129
},
"tpot_ms": {
"mean_ms": 24.121637692939093,
"p90_ms": 30.391929011051804,
"samples": 129
},
"ttft_ms": {
"mean_ms": 619.6665624986844,
"p90_ms": 1980.038105044514,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns64/trial3/results/result.json",
"tp": 1,
"trial": 3
}
]
},
"tp1_mns8": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 137996.3827215202,
"pooled_p90_ms": 222136.04307791684,
"pooled_samples": 387,
"trial_mean_of_means_ms": 137996.38272152017,
"trial_mean_of_p90s_ms": 223625.49974532644,
"trial_stddev_of_means_ms": 2140.2325979814923
},
"tpot_ms": {
"pooled_mean_ms": 14.282724020345517,
"pooled_p90_ms": 16.331995667962982,
"pooled_samples": 387,
"trial_mean_of_means_ms": 14.282724020345519,
"trial_mean_of_p90s_ms": 16.30076252284316,
"trial_stddev_of_means_ms": 0.06621858834469496
},
"ttft_ms": {
"pooled_mean_ms": 88479.94003801016,
"pooled_p90_ms": 168185.5451229494,
"pooled_samples": 387,
"trial_mean_of_means_ms": 88479.94003801014,
"trial_mean_of_p90s_ms": 168658.08321364844,
"trial_stddev_of_means_ms": 1884.4627766230617
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 140225.59553917762,
"p90_ms": 227556.57476198394,
"samples": 129
},
"tpot_ms": {
"mean_ms": 14.352062305808436,
"p90_ms": 16.440255243105753,
"samples": 129
},
"ttft_ms": {
"mean_ms": 90447.42801240031,
"p90_ms": 172230.12473701965,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns8/trial1/results/result.json",
"tp": 1,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 135957.89686511378,
"p90_ms": 219851.20729799382,
"samples": 129
},
"tpot_ms": {
"mean_ms": 14.220143355456331,
"p90_ms": 16.233834194804228,
"samples": 129
},
"ttft_ms": {
"mean_ms": 86691.24682294385,
"p90_ms": 165558.5797809763,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns8/trial2/results/result.json",
"tp": 1,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 137805.65576026912,
"p90_ms": 223468.71717600152,
"samples": 129
},
"tpot_ms": {
"mean_ms": 14.27596639977179,
"p90_ms": 16.22819813061951,
"samples": 129
},
"ttft_ms": {
"mean_ms": 88301.14527868626,
"p90_ms": 168185.5451229494,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp1_mns8/trial3/results/result.json",
"tp": 1,
"trial": 3
}
]
},
"tp2_mns16": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 82936.70731938194,
"pooled_p90_ms": 137726.37015290093,
"pooled_samples": 387,
"trial_mean_of_means_ms": 82936.70731938194,
"trial_mean_of_p90s_ms": 138201.4721223386,
"trial_stddev_of_means_ms": 300.9695124191068
},
"tpot_ms": {
"pooled_mean_ms": 14.09701977962943,
"pooled_p90_ms": 16.149621762221777,
"pooled_samples": 387,
"trial_mean_of_means_ms": 14.097019779629429,
"trial_mean_of_p90s_ms": 16.09430993926139,
"trial_stddev_of_means_ms": 0.009955627205404972
},
"ttft_ms": {
"pooled_mean_ms": 34442.970551674625,
"pooled_p90_ms": 77922.3588249879,
"pooled_samples": 387,
"trial_mean_of_means_ms": 34442.970551674625,
"trial_mean_of_p90s_ms": 77968.50966336206,
"trial_stddev_of_means_ms": 245.29483636828064
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 82589.82040462908,
"p90_ms": 137726.37015290093,
"samples": 129
},
"tpot_ms": {
"mean_ms": 14.086598798707497,
"p90_ms": 16.11459195926328,
"samples": 129
},
"ttft_ms": {
"mean_ms": 34161.05435777441,
"p90_ms": 77612.42442007642,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns16/trial1/results/result.json",
"tp": 2,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 83128.44691530518,
"p90_ms": 138288.39251911268,
"samples": 129
},
"tpot_ms": {
"mean_ms": 14.106433468934316,
"p90_ms": 16.016419614614367,
"samples": 129
},
"ttft_ms": {
"mean_ms": 34560.222078616505,
"p90_ms": 78136.86232501641,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns16/trial2/results/result.json",
"tp": 2,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 83091.85463821157,
"p90_ms": 138589.6536950022,
"samples": 129
},
"tpot_ms": {
"mean_ms": 14.098027071246479,
"p90_ms": 16.151918243906525,
"samples": 129
},
"ttft_ms": {
"mean_ms": 34607.63521863295,
"p90_ms": 78156.24224499334,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns16/trial3/results/result.json",
"tp": 2,
"trial": 3
}
]
},
"tp2_mns32": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 62267.50376182025,
"pooled_p90_ms": 117239.33770309668,
"pooled_samples": 387,
"trial_mean_of_means_ms": 62267.503761820255,
"trial_mean_of_p90s_ms": 117291.74644201218,
"trial_stddev_of_means_ms": 112.21230246579601
},
"tpot_ms": {
"pooled_mean_ms": 17.99748422275332,
"pooled_p90_ms": 21.280696114935413,
"pooled_samples": 387,
"trial_mean_of_means_ms": 17.997484222753318,
"trial_mean_of_p90s_ms": 21.350839715170533,
"trial_stddev_of_means_ms": 0.024026728370897792
},
"ttft_ms": {
"pooled_mean_ms": 1050.6337350342683,
"pooled_p90_ms": 2521.7670870479196,
"pooled_samples": 387,
"trial_mean_of_means_ms": 1050.6337350342683,
"trial_mean_of_p90s_ms": 2558.191083294029,
"trial_stddev_of_means_ms": 3.4732818152181575
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 62313.896939208076,
"p90_ms": 117369.6845669765,
"samples": 129
},
"tpot_ms": {
"mean_ms": 18.008634396299446,
"p90_ms": 21.280696114935413,
"samples": 129
},
"ttft_ms": {
"mean_ms": 1047.8459898673013,
"p90_ms": 2537.152646924369,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns32/trial1/results/result.json",
"tp": 2,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 62139.53429665742,
"p90_ms": 117239.33770309668,
"samples": 129
},
"tpot_ms": {
"mean_ms": 17.969908251938307,
"p90_ms": 21.518877847159697,
"samples": 129
},
"ttft_ms": {
"mean_ms": 1054.5246146724003,
"p90_ms": 2507.6513229869306,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns32/trial2/results/result.json",
"tp": 2,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 62349.08004959526,
"p90_ms": 117266.21705596335,
"samples": 129
},
"tpot_ms": {
"mean_ms": 18.01391002002221,
"p90_ms": 21.252945183416482,
"samples": 129
},
"ttft_ms": {
"mean_ms": 1049.5306005631032,
"p90_ms": 2629.7692799707875,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns32/trial3/results/result.json",
"tp": 2,
"trial": 3
}
]
},
"tp2_mns64": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 62353.63015175185,
"pooled_p90_ms": 118392.33276096638,
"pooled_samples": 387,
"trial_mean_of_means_ms": 62353.63015175186,
"trial_mean_of_p90s_ms": 118379.07191404763,
"trial_stddev_of_means_ms": 167.93391905527702
},
"tpot_ms": {
"pooled_mean_ms": 18.239052597893746,
"pooled_p90_ms": 21.560091553416488,
"pooled_samples": 387,
"trial_mean_of_means_ms": 18.239052597893746,
"trial_mean_of_p90s_ms": 21.583819967748195,
"trial_stddev_of_means_ms": 0.049751307212647984
},
"ttft_ms": {
"pooled_mean_ms": 375.75464864569483,
"pooled_p90_ms": 1148.593939957209,
"pooled_samples": 387,
"trial_mean_of_means_ms": 375.7546486456948,
"trial_mean_of_p90s_ms": 1156.997084307174,
"trial_stddev_of_means_ms": 0.8716131740409625
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 62355.466284980466,
"p90_ms": 118664.41726102494,
"samples": 129
},
"tpot_ms": {
"mean_ms": 18.244751771577278,
"p90_ms": 21.63474199378721,
"samples": 129
},
"ttft_ms": {
"mean_ms": 376.58390670729614,
"p90_ms": 1167.6474569831043,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns64/trial1/results/result.json",
"tp": 2,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 62520.63847564941,
"p90_ms": 118612.22067801282,
"samples": 129
},
"tpot_ms": {
"mean_ms": 18.285708890842777,
"p90_ms": 21.623841600560585,
"samples": 129
},
"ttft_ms": {
"mean_ms": 374.8460987451046,
"p90_ms": 1154.7498559812084,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns64/trial2/results/result.json",
"tp": 2,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 62184.7856946257,
"p90_ms": 117860.57780310512,
"samples": 129
},
"tpot_ms": {
"mean_ms": 18.18669713126118,
"p90_ms": 21.492876308896783,
"samples": 129
},
"ttft_ms": {
"mean_ms": 375.83394048468375,
"p90_ms": 1148.593939957209,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns64/trial3/results/result.json",
"tp": 2,
"trial": 3
}
]
},
"tp2_mns8": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 128296.94038681783,
"pooled_p90_ms": 215848.26136403717,
"pooled_samples": 387,
"trial_mean_of_means_ms": 128296.94038681783,
"trial_mean_of_p90s_ms": 216133.82774300408,
"trial_stddev_of_means_ms": 253.41927200868048
},
"tpot_ms": {
"pooled_mean_ms": 9.244370614307101,
"pooled_p90_ms": 10.190386958999568,
"pooled_samples": 387,
"trial_mean_of_means_ms": 9.244370614307103,
"trial_mean_of_p90s_ms": 10.191423497770423,
"trial_stddev_of_means_ms": 0.011292566107855785
},
"ttft_ms": {
"pooled_mean_ms": 95920.08463817694,
"pooled_p90_ms": 183878.83744493593,
"pooled_samples": 387,
"trial_mean_of_means_ms": 95920.08463817695,
"trial_mean_of_p90s_ms": 184091.12584633598,
"trial_stddev_of_means_ms": 246.8869690217293
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 128460.99048956431,
"p90_ms": 216078.09481292497,
"samples": 129
},
"tpot_ms": {
"mean_ms": 9.23733635271789,
"p90_ms": 10.187162502093368,
"samples": 129
},
"ttft_ms": {
"mean_ms": 96106.82740485722,
"p90_ms": 184006.2059359625,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns8/trial1/results/result.json",
"tp": 2,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 128005.065388418,
"p90_ms": 215848.26136403717,
"samples": 129
},
"tpot_ms": {
"mean_ms": 9.238379252990907,
"p90_ms": 10.190386958999568,
"samples": 129
},
"ttft_ms": {
"mean_ms": 95640.16961438723,
"p90_ms": 183862.3900989769,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns8/trial2/results/result.json",
"tp": 2,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 128424.76528247118,
"p90_ms": 216475.12705205008,
"samples": 129
},
"tpot_ms": {
"mean_ms": 9.257396237212511,
"p90_ms": 10.19672103221833,
"samples": 129
},
"ttft_ms": {
"mean_ms": 96013.2568952864,
"p90_ms": 184404.78150406852,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp2_mns8/trial3/results/result.json",
"tp": 2,
"trial": 3
}
]
},
"tp4_mns16": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 65779.08377038215,
"pooled_p90_ms": 110374.3328540586,
"pooled_samples": 387,
"trial_mean_of_means_ms": 65779.08377038215,
"trial_mean_of_p90s_ms": 110593.98958502182,
"trial_stddev_of_means_ms": 79.17219724389233
},
"tpot_ms": {
"pooled_mean_ms": 8.806983272811275,
"pooled_p90_ms": 9.626243685920313,
"pooled_samples": 387,
"trial_mean_of_means_ms": 8.806983272811275,
"trial_mean_of_p90s_ms": 9.623114289857485,
"trial_stddev_of_means_ms": 0.00993955000910551
},
"ttft_ms": {
"pooled_mean_ms": 35042.61012925094,
"pooled_p90_ms": 74281.28085809294,
"pooled_samples": 387,
"trial_mean_of_means_ms": 35042.610129250934,
"trial_mean_of_p90s_ms": 74402.09861169569,
"trial_stddev_of_means_ms": 73.61057537965252
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 65867.343166196,
"p90_ms": 110630.9638969833,
"samples": 129
},
"tpot_ms": {
"mean_ms": 8.809118914012139,
"p90_ms": 9.634016542784963,
"samples": 129
},
"ttft_ms": {
"mean_ms": 35125.26244937818,
"p90_ms": 74573.83236999158,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns16/trial1/results/result.json",
"tp": 4,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 65755.59253424818,
"p90_ms": 110776.67200402357,
"samples": 129
},
"tpot_ms": {
"mean_ms": 8.815681410017397,
"p90_ms": 9.632762871791096,
"samples": 129
},
"ttft_ms": {
"mean_ms": 34984.109502939085,
"p90_ms": 74351.18260700256,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns16/trial2/results/result.json",
"tp": 4,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 65714.31561070224,
"p90_ms": 110374.3328540586,
"samples": 129
},
"tpot_ms": {
"mean_ms": 8.79614949440429,
"p90_ms": 9.602563454996394,
"samples": 129
},
"ttft_ms": {
"mean_ms": 35018.45843543555,
"p90_ms": 74281.28085809294,
"samples": 129
}
},
"mns": 16,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns16/trial3/results/result.json",
"tp": 4,
"trial": 3
}
]
},
"tp4_mns32": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 47944.52887711636,
"pooled_p90_ms": 83825.7212981116,
"pooled_samples": 387,
"trial_mean_of_means_ms": 47944.52887711636,
"trial_mean_of_p90s_ms": 84201.27462669431,
"trial_stddev_of_means_ms": 314.84689084624006
},
"tpot_ms": {
"pooled_mean_ms": 12.181116414313188,
"pooled_p90_ms": 13.75045588274547,
"pooled_samples": 387,
"trial_mean_of_means_ms": 12.181116414313186,
"trial_mean_of_p90s_ms": 13.764242049706512,
"trial_stddev_of_means_ms": 0.05312117190893007
},
"ttft_ms": {
"pooled_mean_ms": 6296.228585964807,
"pooled_p90_ms": 19159.870272967964,
"pooled_samples": 387,
"trial_mean_of_means_ms": 6296.228585964807,
"trial_mean_of_p90s_ms": 19163.547096347127,
"trial_stddev_of_means_ms": 141.30091703005428
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 48268.361499181105,
"p90_ms": 84625.69403997622,
"samples": 129
},
"tpot_ms": {
"mean_ms": 12.233328917928599,
"p90_ms": 13.809786885754699,
"samples": 129
},
"ttft_ms": {
"mean_ms": 6448.833544885689,
"p90_ms": 19575.953547027893,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns32/trial1/results/result.json",
"tp": 4,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 47639.51159005465,
"p90_ms": 83825.7212981116,
"samples": 129
},
"tpot_ms": {
"mean_ms": 12.127130963187064,
"p90_ms": 13.718962375832668,
"samples": 129
},
"ttft_ms": {
"mean_ms": 6169.928666188873,
"p90_ms": 18779.106684960425,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns32/trial2/results/result.json",
"tp": 4,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 47925.713542113335,
"p90_ms": 84152.4085419951,
"samples": 129
},
"tpot_ms": {
"mean_ms": 12.182889361823895,
"p90_ms": 13.763976887532172,
"samples": 129
},
"ttft_ms": {
"mean_ms": 6269.923546819861,
"p90_ms": 19135.58105705306,
"samples": 129
}
},
"mns": 32,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns32/trial3/results/result.json",
"tp": 4,
"trial": 3
}
]
},
"tp4_mns64": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 44985.13828221082,
"pooled_p90_ms": 83763.61756504048,
"pooled_samples": 387,
"trial_mean_of_means_ms": 44985.13828221082,
"trial_mean_of_p90s_ms": 83934.7783580112,
"trial_stddev_of_means_ms": 124.87733418991748
},
"tpot_ms": {
"pooled_mean_ms": 13.178025610291787,
"pooled_p90_ms": 15.389331985205848,
"pooled_samples": 387,
"trial_mean_of_means_ms": 13.178025610291789,
"trial_mean_of_p90s_ms": 15.426827041851524,
"trial_stddev_of_means_ms": 0.04180975339667582
},
"ttft_ms": {
"pooled_mean_ms": 245.9527667526406,
"pooled_p90_ms": 685.5076659703627,
"pooled_samples": 387,
"trial_mean_of_means_ms": 245.9527667526406,
"trial_mean_of_p90s_ms": 693.9916636717195,
"trial_stddev_of_means_ms": 2.5068538619550047
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 45089.82795085072,
"p90_ms": 84275.81048000138,
"samples": 129
},
"tpot_ms": {
"mean_ms": 13.210321995984648,
"p90_ms": 15.465764498700636,
"samples": 129
},
"ttft_ms": {
"mean_ms": 245.75881352882774,
"p90_ms": 685.5076659703627,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns64/trial1/results/result.json",
"tp": 4,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 44846.91910041485,
"p90_ms": 83676.52938805986,
"samples": 129
},
"tpot_ms": {
"mean_ms": 13.130800688491455,
"p90_ms": 15.375762174125736,
"samples": 129
},
"ttft_ms": {
"mean_ms": 243.54852308327867,
"p90_ms": 689.600904006511,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns64/trial2/results/result.json",
"tp": 4,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 45018.66779536688,
"p90_ms": 83851.99520597234,
"samples": 129
},
"tpot_ms": {
"mean_ms": 13.192954146399261,
"p90_ms": 15.438954452728199,
"samples": 129
},
"ttft_ms": {
"mean_ms": 248.5509636458154,
"p90_ms": 706.8664210382849,
"samples": 129
}
},
"mns": 64,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns64/trial3/results/result.json",
"tp": 4,
"trial": 3
}
]
},
"tp4_mns8": {
"metrics": {
"e2e_ms": {
"pooled_mean_ms": 126405.41922729714,
"pooled_p90_ms": 209092.4072249327,
"pooled_samples": 387,
"trial_mean_of_means_ms": 126405.41922729714,
"trial_mean_of_p90s_ms": 209266.24333831327,
"trial_stddev_of_means_ms": 13.636256279124204
},
"tpot_ms": {
"pooled_mean_ms": 6.757398869596114,
"pooled_p90_ms": 7.242299378971089,
"pooled_samples": 387,
"trial_mean_of_means_ms": 6.757398869596114,
"trial_mean_of_p90s_ms": 7.256281134648607,
"trial_stddev_of_means_ms": 0.0021221970115212385
},
"ttft_ms": {
"pooled_mean_ms": 102605.87212059843,
"pooled_p90_ms": 193359.7398570273,
"pooled_samples": 387,
"trial_mean_of_means_ms": 102605.87212059843,
"trial_mean_of_p90s_ms": 193580.62508733323,
"trial_stddev_of_means_ms": 31.928988273539215
}
},
"trials": [
{
"metrics": {
"e2e_ms": {
"mean_ms": 126396.78182327477,
"p90_ms": 209295.79258395825,
"samples": 129
},
"tpot_ms": {
"mean_ms": 6.757000882714381,
"p90_ms": 7.275805782008363,
"samples": 129
},
"ttft_ms": {
"mean_ms": 102581.32991039865,
"p90_ms": 193756.58583000768,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns8/trial1/results/result.json",
"tp": 4,
"trial": 1
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 126421.1394171115,
"p90_ms": 209410.53020604886,
"samples": 129
},
"tpot_ms": {
"mean_ms": 6.759691884285006,
"p90_ms": 7.268115562828778,
"samples": 129
},
"ttft_ms": {
"mean_ms": 102641.97010612892,
"p90_ms": 193625.54957496468,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns8/trial2/results/result.json",
"tp": 4,
"trial": 2
},
{
"metrics": {
"e2e_ms": {
"mean_ms": 126398.33644150518,
"p90_ms": 209092.4072249327,
"samples": 129
},
"tpot_ms": {
"mean_ms": 6.755503841788953,
"p90_ms": 7.224922059108682,
"samples": 129
},
"ttft_ms": {
"mean_ms": 102594.31634526773,
"p90_ms": 193359.7398570273,
"samples": 129
}
},
"mns": 8,
"requests": 129,
"result_path": "/home/admin/cpfs/wjh/aituner/simulator-tuning-latency-q30-tp-normalized-u0p01-20260717/real/tp4_mns8/trial3/results/result.json",
"tp": 4,
"trial": 3
}
]
}
},
"schema": "qwen30-tp-normalized-real-surface-audit-v1",
"trace_manifests": {
"tp1": {
"global_offered_request_rate": 0.215,
"normalized_row_vector_sha256": "e0c0f9b327473dd2b34149eb478a13b464c26df2201ec8e43c6493c207b08786",
"per_gpu_offered_request_rate": 0.215,
"private_jsonl_sha256": "ba61259d2650b225eb3b3b2b6b9dec29503f9f616b4c573869f3b69881cc6e13",
"requests": 129
},
"tp2": {
"global_offered_request_rate": 0.43,
"normalized_row_vector_sha256": "b435481838bee09b18d0b51c23901488a4fe2f08aee4e9b5a46d2f9c160717fc",
"per_gpu_offered_request_rate": 0.215,
"private_jsonl_sha256": "b30ffdb42ee52e17f099d11e47d9f841673e18a05defccc64bf03ce3802d3576",
"requests": 129
},
"tp4": {
"global_offered_request_rate": 0.86,
"normalized_row_vector_sha256": "0eefd2d58674c741a7680b90554ed89cc1655478a709c49670a93d09f156ff2c",
"per_gpu_offered_request_rate": 0.215,
"private_jsonl_sha256": "a8c1d9641107978e960ecc78cd3d43fdce925e3b23d7635ccf0633fcadc3d812",
"requests": 129
}
},
"winners": {
"e2e_ms:pooled_mean_ms": {
"ranking": [
"tp4_mns64",
"tp4_mns32",
"tp2_mns32",
"tp2_mns64",
"tp4_mns16",
"tp1_mns32",
"tp1_mns64",
"tp2_mns16",
"tp1_mns16",
"tp4_mns8",
"tp2_mns8",
"tp1_mns8"
],
"winner": "tp4_mns64",
"winner_value_ms": 44985.13828221082
},
"e2e_ms:pooled_p90_ms": {
"ranking": [
"tp4_mns64",
"tp4_mns32",
"tp4_mns16",
"tp2_mns32",
"tp2_mns64",
"tp2_mns16",
"tp1_mns16",
"tp1_mns32",
"tp1_mns64",
"tp4_mns8",
"tp2_mns8",
"tp1_mns8"
],
"winner": "tp4_mns64",
"winner_value_ms": 83763.61756504048
},
"tpot_ms:pooled_mean_ms": {
"ranking": [
"tp4_mns8",
"tp4_mns16",
"tp2_mns8",
"tp4_mns32",
"tp4_mns64",
"tp2_mns16",
"tp1_mns8",
"tp2_mns32",
"tp2_mns64",
"tp1_mns16",
"tp1_mns32",
"tp1_mns64"
],
"winner": "tp4_mns8",
"winner_value_ms": 6.757398869596114
},
"tpot_ms:pooled_p90_ms": {
"ranking": [
"tp4_mns8",
"tp4_mns16",
"tp2_mns8",
"tp4_mns32",
"tp4_mns64",
"tp2_mns16",
"tp1_mns8",
"tp2_mns32",
"tp2_mns64",
"tp1_mns16",
"tp1_mns32",
"tp1_mns64"
],
"winner": "tp4_mns8",
"winner_value_ms": 7.242299378971089
},
"ttft_ms:pooled_mean_ms": {
"ranking": [
"tp4_mns64",
"tp2_mns64",
"tp1_mns64",
"tp1_mns32",
"tp2_mns32",
"tp4_mns32",
"tp1_mns16",
"tp2_mns16",
"tp4_mns16",
"tp1_mns8",
"tp2_mns8",
"tp4_mns8"
],
"winner": "tp4_mns64",
"winner_value_ms": 245.9527667526406
},
"ttft_ms:pooled_p90_ms": {
"ranking": [
"tp4_mns64",
"tp2_mns64",
"tp1_mns32",
"tp1_mns64",
"tp2_mns32",
"tp4_mns32",
"tp1_mns16",
"tp4_mns16",
"tp2_mns16",
"tp1_mns8",
"tp2_mns8",
"tp4_mns8"
],
"winner": "tp4_mns64",
"winner_value_ms": 685.5076659703627
}
}
}