852 lines
No EOL
26 KiB
JSON
852 lines
No EOL
26 KiB
JSON
{
|
|
"prompt": "A playful orange tabby cat starts in an ordinary cozy living room in a normal house, afternoon light, sofa and rug. The cat crouches, jumps, and does one clean athletic backflip in slow motion. As the backflip completes there is a sharp cinematic cut: the cat lands perfectly on a glowing neon disco dance floor wearing oversized black sunglasses. Mirror ball reflections, colorful lights, joyful party energy, stylish and funny, clear before-and-after transformation.",
|
|
"model_path": "/models/minimax_h3_fl2va_pruned_nvfp4.safetensors",
|
|
"width": 1344,
|
|
"height": 768,
|
|
"frames": 124,
|
|
"steps": 12,
|
|
"sampler_step": 1,
|
|
"seed": 440420,
|
|
"text_tokens": 100,
|
|
"block_index": 24,
|
|
"attention": "sage2",
|
|
"warmup": 2,
|
|
"iterations": 3,
|
|
"hidden_shape": [
|
|
37810,
|
|
5376
|
|
],
|
|
"output_shape": [
|
|
37810,
|
|
5376
|
|
],
|
|
"segments": [
|
|
[
|
|
0,
|
|
100,
|
|
1
|
|
],
|
|
[
|
|
100,
|
|
514,
|
|
2
|
|
],
|
|
[
|
|
514,
|
|
37810,
|
|
0
|
|
]
|
|
],
|
|
"timings": {
|
|
"norm1": {
|
|
"count": 3,
|
|
"mean_s": 0.003741738328244537,
|
|
"p50_s": 0.003866238985210657,
|
|
"p90_s": 0.003901209402829409,
|
|
"p95_s": 0.0039055807050317527,
|
|
"p99_s": 0.003909077746793627,
|
|
"min_s": 0.0034490239922888577,
|
|
"max_s": 0.0039099520072340965
|
|
},
|
|
"modulate_msa": {
|
|
"count": 3,
|
|
"mean_s": 0.011052168323658407,
|
|
"p50_s": 0.010806893988046795,
|
|
"p90_s": 0.012067155598197134,
|
|
"p95_s": 0.012224688299465925,
|
|
"p99_s": 0.012350714460480958,
|
|
"min_s": 0.009967389982193708,
|
|
"max_s": 0.012382221000734717
|
|
},
|
|
"linear.attn_qkv_proj.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 5.536014214158058e-06,
|
|
"p50_s": 4.143978003412485e-06,
|
|
"p90_s": 7.62563431635499e-06,
|
|
"p95_s": 8.060841355472804e-06,
|
|
"p99_s": 8.409006986767054e-06,
|
|
"min_s": 3.968016244471073e-06,
|
|
"max_s": 8.496048394590616e-06
|
|
},
|
|
"linear.attn_qkv_proj.pre_quant_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0,
|
|
"p50_s": 0.0,
|
|
"p90_s": 0.0,
|
|
"p95_s": 0.0,
|
|
"p99_s": 0.0,
|
|
"min_s": 0.0,
|
|
"max_s": 0.0
|
|
},
|
|
"linear.attn_qkv_proj.packed_weight_wrapper": {
|
|
"count": 3,
|
|
"mean_s": 1.4229328371584415e-05,
|
|
"p50_s": 8.367991540580988e-06,
|
|
"p90_s": 2.274239668622613e-05,
|
|
"p95_s": 2.4539197329431773e-05,
|
|
"p99_s": 2.597663784399629e-05,
|
|
"min_s": 7.983995601534843e-06,
|
|
"max_s": 2.6335997972637415e-05
|
|
},
|
|
"linear.attn_qkv_proj.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.509320473919312e-06,
|
|
"p50_s": 3.4240074455738068e-06,
|
|
"p90_s": 3.6927871406078338e-06,
|
|
"p95_s": 3.7263846024870874e-06,
|
|
"p99_s": 3.7532625719904902e-06,
|
|
"min_s": 3.343971911817789e-06,
|
|
"max_s": 3.7599820643663406e-06
|
|
},
|
|
"linear.attn_qkv_proj.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0017316586648424466,
|
|
"p50_s": 0.0017092640046030283,
|
|
"p90_s": 0.0017795104067772627,
|
|
"p95_s": 0.001788291207049042,
|
|
"p99_s": 0.0017953158472664653,
|
|
"min_s": 0.0016886399826034904,
|
|
"max_s": 0.0017970720073208213
|
|
},
|
|
"linear.attn_qkv_proj.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.3424005980292955e-05,
|
|
"p50_s": 1.1440017260611057e-05,
|
|
"p90_s": 1.796798314899206e-05,
|
|
"p95_s": 1.8783978885039683e-05,
|
|
"p99_s": 1.9436775473877788e-05,
|
|
"min_s": 9.232026059180498e-06,
|
|
"max_s": 1.9599974621087313e-05
|
|
},
|
|
"linear.attn_qkv_proj.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002303071009616057,
|
|
"p50_s": 0.0022418550215661526,
|
|
"p90_s": 0.002481432631611824,
|
|
"p95_s": 0.002511379832867533,
|
|
"p99_s": 0.0025353375938721003,
|
|
"min_s": 0.0021260309731587768,
|
|
"max_s": 0.002541327034123242
|
|
},
|
|
"linear.attn_qkv_proj.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.789341568946838e-06,
|
|
"p50_s": 6.608024705201387e-06,
|
|
"p90_s": 7.401604671031237e-06,
|
|
"p95_s": 7.500802166759968e-06,
|
|
"p99_s": 7.580160163342953e-06,
|
|
"min_s": 6.160000339150429e-06,
|
|
"max_s": 7.599999662488699e-06
|
|
},
|
|
"linear.attn_qkv_proj.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.029993231679933768,
|
|
"p50_s": 0.026411275030113757,
|
|
"p90_s": 0.03558764460030943,
|
|
"p95_s": 0.03673469079658389,
|
|
"p99_s": 0.03765232775360346,
|
|
"min_s": 0.025686683016829193,
|
|
"max_s": 0.03788173699285835
|
|
},
|
|
"linear.attn_qkv_proj.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 6.229345065852006e-06,
|
|
"p50_s": 5.4559786804020405e-06,
|
|
"p90_s": 7.32481712475419e-06,
|
|
"p95_s": 7.558421930298209e-06,
|
|
"p99_s": 7.745305774733424e-06,
|
|
"min_s": 5.44002978131175e-06,
|
|
"max_s": 7.792026735842228e-06
|
|
},
|
|
"attn_qkv_proj": {
|
|
"count": 3,
|
|
"mean_s": 0.03412259600978965,
|
|
"p50_s": 0.030302010010927916,
|
|
"p90_s": 0.039898885227739815,
|
|
"p95_s": 0.041098494629841296,
|
|
"p99_s": 0.04205818215152249,
|
|
"min_s": 0.029767673986498266,
|
|
"max_s": 0.042298104031942785
|
|
},
|
|
"attn_qkv_split_view": {
|
|
"count": 3,
|
|
"mean_s": 1.5023998760928711e-05,
|
|
"p50_s": 1.2496020644903183e-05,
|
|
"p90_s": 1.977919600903988e-05,
|
|
"p95_s": 2.0689592929556962e-05,
|
|
"p99_s": 2.1417910465970633e-05,
|
|
"min_s": 1.0975985787808895e-05,
|
|
"max_s": 2.1599989850074053e-05
|
|
},
|
|
"attn_qk_rms_rope": {
|
|
"count": 3,
|
|
"mean_s": 0.012241016685341796,
|
|
"p50_s": 0.01193398202303797,
|
|
"p90_s": 0.012705834838561714,
|
|
"p95_s": 0.012802316440502183,
|
|
"p99_s": 0.012879501722054557,
|
|
"min_s": 0.011890269990544766,
|
|
"max_s": 0.01289879804244265
|
|
},
|
|
"attn_q_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.014036514330655336,
|
|
"p50_s": 0.0053002709755674005,
|
|
"p90_s": 0.026539997407235207,
|
|
"p95_s": 0.029194963211193678,
|
|
"p99_s": 0.031318935854360457,
|
|
"min_s": 0.004959343001246452,
|
|
"max_s": 0.031849929015152156
|
|
},
|
|
"attn_k_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.007838361353302995,
|
|
"p50_s": 0.005219135025981814,
|
|
"p90_s": 0.011728727829176933,
|
|
"p95_s": 0.012542426929576323,
|
|
"p99_s": 0.013193386209895835,
|
|
"min_s": 0.00493982300395146,
|
|
"max_s": 0.013356126029975712
|
|
},
|
|
"attn_v_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.01099991767356793,
|
|
"p50_s": 0.005040751013439149,
|
|
"p90_s": 0.019409106194507333,
|
|
"p95_s": 0.021205150592140853,
|
|
"p99_s": 0.02264198611024767,
|
|
"min_s": 0.004957807017490268,
|
|
"max_s": 0.023001194989774376
|
|
},
|
|
"attention_kernel": {
|
|
"count": 3,
|
|
"mean_s": 0.2562586222969306,
|
|
"p50_s": 0.25644913397263736,
|
|
"p90_s": 0.2567798731615767,
|
|
"p95_s": 0.2568212155601941,
|
|
"p99_s": 0.25685428947908806,
|
|
"min_s": 0.2554641749593429,
|
|
"max_s": 0.2568625579588115
|
|
},
|
|
"attn_output_reshape": {
|
|
"count": 3,
|
|
"mean_s": 0.00469084734019513,
|
|
"p50_s": 0.004691840033046901,
|
|
"p90_s": 0.004693656810559332,
|
|
"p95_s": 0.004693883907748386,
|
|
"p99_s": 0.0046940655854996296,
|
|
"min_s": 0.0046865909826010466,
|
|
"max_s": 0.00469411100493744
|
|
},
|
|
"linear.attn_out_proj.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.229329836865266e-06,
|
|
"p50_s": 4.1120219975709915e-06,
|
|
"p90_s": 4.470394924283028e-06,
|
|
"p95_s": 4.515191540122033e-06,
|
|
"p99_s": 4.551028832793236e-06,
|
|
"min_s": 4.01597935706377e-06,
|
|
"max_s": 4.559988155961037e-06
|
|
},
|
|
"linear.attn_out_proj.pre_quant_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0,
|
|
"p50_s": 0.0,
|
|
"p90_s": 0.0,
|
|
"p95_s": 0.0,
|
|
"p99_s": 0.0,
|
|
"min_s": 0.0,
|
|
"max_s": 0.0
|
|
},
|
|
"linear.attn_out_proj.packed_weight_wrapper": {
|
|
"count": 3,
|
|
"mean_s": 9.45067343612512e-06,
|
|
"p50_s": 9.440002031624317e-06,
|
|
"p90_s": 1.010557170957327e-05,
|
|
"p95_s": 1.0188767919316889e-05,
|
|
"p99_s": 1.0255324887111782e-05,
|
|
"min_s": 8.640054147690535e-06,
|
|
"max_s": 1.0271964129060507e-05
|
|
},
|
|
"linear.attn_out_proj.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.5146561761697135e-06,
|
|
"p50_s": 3.4240074455738068e-06,
|
|
"p90_s": 3.667175769805908e-06,
|
|
"p95_s": 3.697571810334921e-06,
|
|
"p99_s": 3.7218886427581313e-06,
|
|
"min_s": 3.3919932320713997e-06,
|
|
"max_s": 3.7279678508639336e-06
|
|
},
|
|
"linear.attn_out_proj.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.002265583665575832,
|
|
"p50_s": 0.0022627509897574782,
|
|
"p90_s": 0.002275923010893166,
|
|
"p95_s": 0.002277569513535127,
|
|
"p99_s": 0.0022788867156486956,
|
|
"min_s": 0.00225478399079293,
|
|
"max_s": 0.002279216016177088
|
|
},
|
|
"linear.attn_out_proj.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.0005314834415913e-05,
|
|
"p50_s": 9.807990863919258e-06,
|
|
"p90_s": 1.0319985449314117e-05,
|
|
"p95_s": 1.0383984772488475e-05,
|
|
"p99_s": 1.0435184231027961e-05,
|
|
"min_s": 9.759969543665648e-06,
|
|
"max_s": 1.0447984095662832e-05
|
|
},
|
|
"linear.attn_out_proj.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002860164638453474,
|
|
"p50_s": 0.0028614879702217877,
|
|
"p90_s": 0.002871010371018201,
|
|
"p95_s": 0.002872200671117753,
|
|
"p99_s": 0.0028731529111973944,
|
|
"min_s": 0.0028456149739213288,
|
|
"max_s": 0.0028733909712173045
|
|
},
|
|
"linear.attn_out_proj.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.378686521202326e-06,
|
|
"p50_s": 6.432004738599062e-06,
|
|
"p90_s": 6.508838851004839e-06,
|
|
"p95_s": 6.518443115055561e-06,
|
|
"p99_s": 6.526126526296139e-06,
|
|
"min_s": 6.176007445901632e-06,
|
|
"max_s": 6.528047379106283e-06
|
|
},
|
|
"linear.attn_out_proj.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.007843875675462186,
|
|
"p50_s": 0.007834365998860449,
|
|
"p90_s": 0.007871806004550308,
|
|
"p95_s": 0.00787648600526154,
|
|
"p99_s": 0.007880230005830527,
|
|
"min_s": 0.007816095021553338,
|
|
"max_s": 0.007881166005972773
|
|
},
|
|
"linear.attn_out_proj.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.5093357029060526e-06,
|
|
"p50_s": 5.552021320909262e-06,
|
|
"p90_s": 5.6416145525872706e-06,
|
|
"p95_s": 5.652813706547021e-06,
|
|
"p99_s": 5.661773029714823e-06,
|
|
"min_s": 5.311972927302122e-06,
|
|
"max_s": 5.664012860506773e-06
|
|
},
|
|
"attn_out_proj": {
|
|
"count": 3,
|
|
"mean_s": 0.013048968006235858,
|
|
"p50_s": 0.013041116995736957,
|
|
"p90_s": 0.013074909010902048,
|
|
"p95_s": 0.013079133012797683,
|
|
"p99_s": 0.013082512214314192,
|
|
"min_s": 0.013022430008277297,
|
|
"max_s": 0.01308335701469332
|
|
},
|
|
"gate_msa": {
|
|
"count": 3,
|
|
"mean_s": 0.008436798331482956,
|
|
"p50_s": 0.008493119035847485,
|
|
"p90_s": 0.008522289409302174,
|
|
"p95_s": 0.008525935705984011,
|
|
"p99_s": 0.00852885274332948,
|
|
"min_s": 0.008287693955935538,
|
|
"max_s": 0.008529582002665848
|
|
},
|
|
"norm2": {
|
|
"count": 3,
|
|
"mean_s": 0.0034889166515010097,
|
|
"p50_s": 0.0034742390271276236,
|
|
"p90_s": 0.0035121661610901357,
|
|
"p95_s": 0.0035169070528354492,
|
|
"p99_s": 0.0035206997662317004,
|
|
"min_s": 0.0034708629827946424,
|
|
"max_s": 0.0035216479445807636
|
|
},
|
|
"modulate_mlp": {
|
|
"count": 3,
|
|
"mean_s": 0.010772093994698176,
|
|
"p50_s": 0.010788750019855797,
|
|
"p90_s": 0.010858983569778501,
|
|
"p95_s": 0.01086776276351884,
|
|
"p99_s": 0.01087478611851111,
|
|
"min_s": 0.010650990006979555,
|
|
"max_s": 0.010876541957259178
|
|
},
|
|
"linear.mlp_fc1.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.058626169959704e-06,
|
|
"p50_s": 3.951950930058956e-06,
|
|
"p90_s": 4.271953366696835e-06,
|
|
"p95_s": 4.3119536712765695e-06,
|
|
"p99_s": 4.343953914940357e-06,
|
|
"min_s": 3.871973603963852e-06,
|
|
"max_s": 4.351953975856304e-06
|
|
},
|
|
"linear.mlp_fc1.pre_quant_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0,
|
|
"p50_s": 0.0,
|
|
"p90_s": 0.0,
|
|
"p95_s": 0.0,
|
|
"p99_s": 0.0,
|
|
"min_s": 0.0,
|
|
"max_s": 0.0
|
|
},
|
|
"linear.mlp_fc1.packed_weight_wrapper": {
|
|
"count": 3,
|
|
"mean_s": 8.293330514182648e-06,
|
|
"p50_s": 8.01600981503725e-06,
|
|
"p90_s": 8.860812522470951e-06,
|
|
"p95_s": 8.966412860900165e-06,
|
|
"p99_s": 9.050893131643534e-06,
|
|
"min_s": 7.791968528181314e-06,
|
|
"max_s": 9.072013199329376e-06
|
|
},
|
|
"linear.mlp_fc1.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.4186717433234057e-06,
|
|
"p50_s": 3.456021659076214e-06,
|
|
"p90_s": 3.4816330298781395e-06,
|
|
"p95_s": 3.48483445122838e-06,
|
|
"p99_s": 3.4873955883085728e-06,
|
|
"min_s": 3.311957698315382e-06,
|
|
"max_s": 3.488035872578621e-06
|
|
},
|
|
"linear.mlp_fc1.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0017874666567270954,
|
|
"p50_s": 0.0017595840035937726,
|
|
"p90_s": 0.0018278975854627787,
|
|
"p95_s": 0.0018364367831964047,
|
|
"p99_s": 0.0018432681413833051,
|
|
"min_s": 0.0017578399856574833,
|
|
"max_s": 0.0018449759809300303
|
|
},
|
|
"linear.mlp_fc1.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 9.296015681078037e-06,
|
|
"p50_s": 9.328010492026806e-06,
|
|
"p90_s": 9.340816177427769e-06,
|
|
"p95_s": 9.342416888102889e-06,
|
|
"p99_s": 9.343697456642985e-06,
|
|
"min_s": 9.216018952429295e-06,
|
|
"max_s": 9.34401759877801e-06
|
|
},
|
|
"linear.mlp_fc1.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002177828651232024,
|
|
"p50_s": 0.002151614986360073,
|
|
"p90_s": 0.0022210165858268737,
|
|
"p95_s": 0.002229691785760224,
|
|
"p99_s": 0.002236631945706904,
|
|
"min_s": 0.002143503981642425,
|
|
"max_s": 0.002238366985693574
|
|
},
|
|
"linear.mlp_fc1.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.623993006845315e-06,
|
|
"p50_s": 6.607966497540474e-06,
|
|
"p90_s": 6.659189239144325e-06,
|
|
"p95_s": 6.665592081844807e-06,
|
|
"p99_s": 6.670714356005192e-06,
|
|
"min_s": 6.592017598450184e-06,
|
|
"max_s": 6.671994924545288e-06
|
|
},
|
|
"linear.mlp_fc1.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.03496598266065121,
|
|
"p50_s": 0.03502911300165579,
|
|
"p90_s": 0.03507212101249024,
|
|
"p95_s": 0.03507749701384455,
|
|
"p99_s": 0.035081797814928,
|
|
"min_s": 0.03478596196509898,
|
|
"max_s": 0.03508287301519886
|
|
},
|
|
"linear.mlp_fc1.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.49866429840525e-06,
|
|
"p50_s": 5.4559786804020405e-06,
|
|
"p90_s": 5.622406024485826e-06,
|
|
"p95_s": 5.643209442496299e-06,
|
|
"p99_s": 5.659852176904679e-06,
|
|
"min_s": 5.376001354306936e-06,
|
|
"max_s": 5.664012860506773e-06
|
|
},
|
|
"mlp_fc1": {
|
|
"count": 3,
|
|
"mean_s": 0.03901222701339672,
|
|
"p50_s": 0.03901776799466461,
|
|
"p90_s": 0.039129883213900034,
|
|
"p95_s": 0.03914389761630446,
|
|
"p99_s": 0.039155109138228,
|
|
"min_s": 0.038861001026816666,
|
|
"max_s": 0.039157912018708885
|
|
},
|
|
"mlp_swiglu": {
|
|
"count": 3,
|
|
"mean_s": 0.02408325932143877,
|
|
"p50_s": 0.0241031949990429,
|
|
"p90_s": 0.024252827803138644,
|
|
"p95_s": 0.02427153190365061,
|
|
"p99_s": 0.024286495184060185,
|
|
"min_s": 0.02385634696111083,
|
|
"max_s": 0.02429023600416258
|
|
},
|
|
"linear.mlp_fc2.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.234665539115667e-06,
|
|
"p50_s": 4.01597935706377e-06,
|
|
"p90_s": 4.617613740265369e-06,
|
|
"p95_s": 4.692818038165569e-06,
|
|
"p99_s": 4.7529814764857296e-06,
|
|
"min_s": 3.9199949242174625e-06,
|
|
"max_s": 4.768022336065769e-06
|
|
},
|
|
"linear.mlp_fc2.pre_quant_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0,
|
|
"p50_s": 0.0,
|
|
"p90_s": 0.0,
|
|
"p95_s": 0.0,
|
|
"p99_s": 0.0,
|
|
"min_s": 0.0,
|
|
"max_s": 0.0
|
|
},
|
|
"linear.mlp_fc2.packed_weight_wrapper": {
|
|
"count": 3,
|
|
"mean_s": 7.589328257987897e-06,
|
|
"p50_s": 7.71199120208621e-06,
|
|
"p90_s": 7.929594721645116e-06,
|
|
"p95_s": 7.95679516158998e-06,
|
|
"p99_s": 7.978555513545871e-06,
|
|
"min_s": 7.071997970342636e-06,
|
|
"max_s": 7.983995601534843e-06
|
|
},
|
|
"linear.mlp_fc2.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.722670953720808e-06,
|
|
"p50_s": 3.503984771668911e-06,
|
|
"p90_s": 4.195212386548519e-06,
|
|
"p95_s": 4.28161583840847e-06,
|
|
"p99_s": 4.3507385998964315e-06,
|
|
"min_s": 3.296008799225092e-06,
|
|
"max_s": 4.368019290268421e-06
|
|
},
|
|
"linear.mlp_fc2.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.004541535027480374,
|
|
"p50_s": 0.004574159043841064,
|
|
"p90_s": 0.004587253439240158,
|
|
"p95_s": 0.004588890238665044,
|
|
"p99_s": 0.004590199678204954,
|
|
"min_s": 0.004459919000510126,
|
|
"max_s": 0.004590527038089931
|
|
},
|
|
"linear.mlp_fc2.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 8.880005528529486e-06,
|
|
"p50_s": 8.608039934188128e-06,
|
|
"p90_s": 9.3247857876122e-06,
|
|
"p95_s": 9.414379019290208e-06,
|
|
"p99_s": 9.486053604632616e-06,
|
|
"min_s": 8.52800440043211e-06,
|
|
"max_s": 9.503972250968218e-06
|
|
},
|
|
"linear.mlp_fc2.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.005774937667107831,
|
|
"p50_s": 0.005734590988140553,
|
|
"p90_s": 0.0058928117970936,
|
|
"p95_s": 0.005912589398212731,
|
|
"p99_s": 0.005928411479108036,
|
|
"min_s": 0.005657855013851076,
|
|
"max_s": 0.005932366999331862
|
|
},
|
|
"linear.mlp_fc2.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.197350254903237e-06,
|
|
"p50_s": 6.128044333308935e-06,
|
|
"p90_s": 6.396824028342963e-06,
|
|
"p95_s": 6.430421490222216e-06,
|
|
"p99_s": 6.457299459725618e-06,
|
|
"min_s": 5.999987479299307e-06,
|
|
"max_s": 6.464018952101469e-06
|
|
},
|
|
"linear.mlp_fc2.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.05555110399533684,
|
|
"p50_s": 0.05553032597526908,
|
|
"p90_s": 0.05586563399992883,
|
|
"p95_s": 0.0559075475030113,
|
|
"p99_s": 0.055941078305477274,
|
|
"min_s": 0.05517352500464767,
|
|
"max_s": 0.05594946100609377
|
|
},
|
|
"linear.mlp_fc2.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.4613337852060795e-06,
|
|
"p50_s": 5.4879928939044476e-06,
|
|
"p90_s": 5.6543736718595025e-06,
|
|
"p95_s": 5.6751712691038845e-06,
|
|
"p99_s": 5.691809346899391e-06,
|
|
"min_s": 5.200039595365524e-06,
|
|
"max_s": 5.695968866348267e-06
|
|
},
|
|
"mlp_fc2": {
|
|
"count": 3,
|
|
"mean_s": 0.06594285130267963,
|
|
"p50_s": 0.06600016396259889,
|
|
"p90_s": 0.06628018877236172,
|
|
"p95_s": 0.06631519187358208,
|
|
"p99_s": 0.06634319435455836,
|
|
"min_s": 0.06547819497063756,
|
|
"max_s": 0.06635019497480243
|
|
},
|
|
"gate_mlp": {
|
|
"count": 3,
|
|
"mean_s": 0.013789538682127992,
|
|
"p50_s": 0.00966468604747206,
|
|
"p90_s": 0.019895839621312916,
|
|
"p95_s": 0.02117473381804302,
|
|
"p99_s": 0.022197849175427106,
|
|
"min_s": 0.009250301984138787,
|
|
"max_s": 0.02245362801477313
|
|
},
|
|
"block_total": {
|
|
"count": 3,
|
|
"mean_s": 0.5336582703360667,
|
|
"p50_s": 0.5072318860329688,
|
|
"p90_s": 0.5710783508140594,
|
|
"p95_s": 0.5790591589116957,
|
|
"p99_s": 0.5854438053898048,
|
|
"min_s": 0.5067029579658993,
|
|
"max_s": 0.587039967009332
|
|
}
|
|
},
|
|
"module_forward": {
|
|
"count": 3,
|
|
"mean_s": 0.47701150766806677,
|
|
"p50_s": 0.4728726129978895,
|
|
"p90_s": 0.48509467700496317,
|
|
"p95_s": 0.48662243500584734,
|
|
"p99_s": 0.48784464140655476,
|
|
"min_s": 0.4700117169995792,
|
|
"max_s": 0.48815019300673157
|
|
},
|
|
"module_forward_checksum": 296883328.0,
|
|
"fused_elementwise": true,
|
|
"profiler_summary": {
|
|
"profiled_iterations": 1,
|
|
"runtime_kernel_launches": 0,
|
|
"runtime_kernel_launches_per_block": 0.0,
|
|
"positive_self_device_allocated_bytes": 7351755264
|
|
},
|
|
"profiler_top_events": [
|
|
{
|
|
"key": "comfy_kitchen::rms_rope_split_half_",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 581.328,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[1, 37810, 56, 128], [1, 37810, 56, 128], [1, 37810, 1, 48, 2, 2], [128], [128], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::linear",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 512.4480000000003,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 7168], [5376, 7168], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 422.6239999999998,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": -512,
|
|
"input_shapes": "[[37824, 3584], [5376, 3584], [], [], [37888, 448], [5376, 448], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::_to_copy",
|
|
"count": 8,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 247.44000000000028,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], [], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::to",
|
|
"count": 13,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 218.31999999999857,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::quantize_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 209.37599999999975,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 7168], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::rms_norm",
|
|
"count": 2,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 187.68000000000043,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": -303104,
|
|
"input_shapes": "[[37810, 5376], [], [5376], []]"
|
|
},
|
|
{
|
|
"key": "aten::_fused_rms_norm",
|
|
"count": 2,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 175.6159999999994,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 5376], [], [5376], []]"
|
|
},
|
|
{
|
|
"key": "aten::copy_",
|
|
"count": 8,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 164.1119999999994,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::linear",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 139.95200000000023,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 14336], [5376, 14336], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 123.92000000000007,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": -512,
|
|
"input_shapes": "[[37824, 7168], [5376, 7168], [], [], [37888, 896], [5376, 896], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::empty",
|
|
"count": 29,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 123.24799999999709,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 6267507200,
|
|
"input_shapes": "[[], [], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::linear",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 121.84000000000015,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 5376], [28672, 5376], []]"
|
|
},
|
|
{
|
|
"key": "aten::zeros",
|
|
"count": 3,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 106.24000000000069,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 100.2559999999994,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": -512,
|
|
"input_shapes": "[[37824, 2688], [28672, 2688], [], [], [37888, 336], [28672, 336], [], [], []]"
|
|
},
|
|
{
|
|
"key": "sageattention_sm89::qk_int8_sv_f8_accum_f16_fuse_v_scale_attn_inst_buf",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 95.00799999999981,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[1, 37810, 56, 128], [1, 37810, 56, 128], [1, 128, 56, 37824], [1, 37810, 56, 128], [1, 56, 1184], [1, 56, 591], [1, 56, 128], [], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::mean",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 88.44800000000032,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 14336,
|
|
"input_shapes": "[[1, 37810, 56, 128], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::mul",
|
|
"count": 4,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 76.6240000000007,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 2048,
|
|
"input_shapes": "[[], []]"
|
|
},
|
|
{
|
|
"key": "aten::silu",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 64.27199999999993,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 1084227584,
|
|
"input_shapes": "[[37810, 14336]]"
|
|
},
|
|
{
|
|
"key": "aten::copy_",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 63.455999999999904,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[1], [], []]"
|
|
}
|
|
]
|
|
} |