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": 0,
|
|
"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": 0.0038844093408746025,
|
|
"p50_s": 0.003442010027356446,
|
|
"p90_s": 0.004506341204978525,
|
|
"p95_s": 0.004639382602181286,
|
|
"p99_s": 0.004745815719943493,
|
|
"min_s": 0.003438793995883316,
|
|
"max_s": 0.004772423999384046
|
|
},
|
|
"modulate_msa": {
|
|
"count": 3,
|
|
"mean_s": 0.011301896010991186,
|
|
"p50_s": 0.010341919027268887,
|
|
"p90_s": 0.012674343818798662,
|
|
"p95_s": 0.012965896917739883,
|
|
"p99_s": 0.01319913939689286,
|
|
"min_s": 0.010306318989023566,
|
|
"max_s": 0.013257450016681105
|
|
},
|
|
"linear.attn_qkv_proj.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 5.6800199672579765e-06,
|
|
"p50_s": 4.656030796468258e-06,
|
|
"p90_s": 7.612793706357479e-06,
|
|
"p95_s": 7.982389070093631e-06,
|
|
"p99_s": 8.278065361082553e-06,
|
|
"min_s": 4.032044671475887e-06,
|
|
"max_s": 8.351984433829784e-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": 1.581335285057624e-05,
|
|
"p50_s": 8.608039934188128e-06,
|
|
"p90_s": 2.631038660183549e-05,
|
|
"p95_s": 2.8523179935291406e-05,
|
|
"p99_s": 3.0293414602056147e-05,
|
|
"min_s": 8.096045348793268e-06,
|
|
"max_s": 3.073597326874733e-05
|
|
},
|
|
"linear.attn_qkv_proj.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.535998985171318e-06,
|
|
"p50_s": 3.535998985171318e-06,
|
|
"p90_s": 3.6255922168493273e-06,
|
|
"p95_s": 3.636791370809078e-06,
|
|
"p99_s": 3.645750693976879e-06,
|
|
"min_s": 3.4240074455738068e-06,
|
|
"max_s": 3.6479905247688293e-06
|
|
},
|
|
"linear.attn_qkv_proj.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.002059137332253158,
|
|
"p50_s": 0.0018270040163770318,
|
|
"p90_s": 0.0024354639695957302,
|
|
"p95_s": 0.0025115214637480674,
|
|
"p99_s": 0.0025723674590699373,
|
|
"min_s": 0.0017628290224820375,
|
|
"max_s": 0.002587578957900405
|
|
},
|
|
"linear.attn_qkv_proj.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 2.2490644672264654e-05,
|
|
"p50_s": 2.297596074640751e-05,
|
|
"p90_s": 2.3987190797924995e-05,
|
|
"p95_s": 2.411359455436468e-05,
|
|
"p99_s": 2.421471755951643e-05,
|
|
"min_s": 2.025597495958209e-05,
|
|
"max_s": 2.4239998310804367e-05
|
|
},
|
|
"linear.attn_qkv_proj.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002397505333647132,
|
|
"p50_s": 0.0022308279876597226,
|
|
"p90_s": 0.0026664879755117,
|
|
"p95_s": 0.002720945473993197,
|
|
"p99_s": 0.002764511472778395,
|
|
"min_s": 0.002186285040806979,
|
|
"max_s": 0.0027754029724746943
|
|
},
|
|
"linear.attn_qkv_proj.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 7.466665313889583e-06,
|
|
"p50_s": 7.599999662488699e-06,
|
|
"p90_s": 7.996789645403623e-06,
|
|
"p95_s": 8.046388393267989e-06,
|
|
"p99_s": 8.086067391559481e-06,
|
|
"min_s": 6.704009138047695e-06,
|
|
"max_s": 8.095987141132355e-06
|
|
},
|
|
"linear.attn_qkv_proj.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.039801057001265384,
|
|
"p50_s": 0.025940386985894293,
|
|
"p90_s": 0.059222071792464706,
|
|
"p95_s": 0.063382282393286,
|
|
"p99_s": 0.06671045087394303,
|
|
"min_s": 0.02592029102379456,
|
|
"max_s": 0.0675424929941073
|
|
},
|
|
"linear.attn_qkv_proj.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 6.346691710253556e-06,
|
|
"p50_s": 5.4720439948141575e-06,
|
|
"p90_s": 7.63522693887353e-06,
|
|
"p95_s": 7.90562480688095e-06,
|
|
"p99_s": 8.121943101286887e-06,
|
|
"min_s": 5.39200846105814e-06,
|
|
"max_s": 8.176022674888372e-06
|
|
},
|
|
"attn_qkv_proj": {
|
|
"count": 3,
|
|
"mean_s": 0.04436540632741526,
|
|
"p50_s": 0.03002823598217219,
|
|
"p90_s": 0.06444033759180456,
|
|
"p95_s": 0.06874185029300861,
|
|
"p99_s": 0.07218306045397184,
|
|
"min_s": 0.030024620005860925,
|
|
"max_s": 0.07304336299421266
|
|
},
|
|
"attn_qkv_split_view": {
|
|
"count": 3,
|
|
"mean_s": 1.6122687763224047e-05,
|
|
"p50_s": 1.315202098339796e-05,
|
|
"p90_s": 2.1305610425770285e-05,
|
|
"p95_s": 2.2324809106066822e-05,
|
|
"p99_s": 2.3140168050304055e-05,
|
|
"min_s": 1.1872034519910812e-05,
|
|
"max_s": 2.3344007786363363e-05
|
|
},
|
|
"attn_qk_rms_rope": {
|
|
"count": 3,
|
|
"mean_s": 0.012314411326466749,
|
|
"p50_s": 0.012499674980062991,
|
|
"p90_s": 0.012545524595770985,
|
|
"p95_s": 0.012551255797734485,
|
|
"p99_s": 0.012555840759305284,
|
|
"min_s": 0.011886571999639273,
|
|
"max_s": 0.012556986999697983
|
|
},
|
|
"attn_q_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.005020225983268271,
|
|
"p50_s": 0.005004551960155368,
|
|
"p90_s": 0.005060359183698892,
|
|
"p95_s": 0.0050673350866418335,
|
|
"p99_s": 0.005072915808996186,
|
|
"min_s": 0.004981815000064671,
|
|
"max_s": 0.005074310989584774
|
|
},
|
|
"attn_k_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.005056855365789185,
|
|
"p50_s": 0.005042263015639037,
|
|
"p90_s": 0.0050791406421922146,
|
|
"p95_s": 0.0050837503455113614,
|
|
"p99_s": 0.005087438108166679,
|
|
"min_s": 0.005039943032898009,
|
|
"max_s": 0.005088360048830509
|
|
},
|
|
"attn_v_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.005133500671945512,
|
|
"p50_s": 0.005174166988581419,
|
|
"p90_s": 0.0052229606080800295,
|
|
"p95_s": 0.005229059810517356,
|
|
"p99_s": 0.005233939172467217,
|
|
"min_s": 0.004991176014300436,
|
|
"max_s": 0.005235159012954682
|
|
},
|
|
"attention_kernel": {
|
|
"count": 3,
|
|
"mean_s": 0.2589895313140005,
|
|
"p50_s": 0.25650634698104113,
|
|
"p90_s": 0.26326487499754875,
|
|
"p95_s": 0.26410969099961223,
|
|
"p99_s": 0.264785543801263,
|
|
"min_s": 0.25550773995928466,
|
|
"max_s": 0.26495450700167567
|
|
},
|
|
"attn_output_reshape": {
|
|
"count": 3,
|
|
"mean_s": 0.0046977306677338975,
|
|
"p50_s": 0.004640695988200605,
|
|
"p90_s": 0.004778039990924299,
|
|
"p95_s": 0.0047952079912647605,
|
|
"p99_s": 0.00480894239153713,
|
|
"min_s": 0.004640120023395866,
|
|
"max_s": 0.004812375991605222
|
|
},
|
|
"linear.attn_out_proj.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.453351721167564e-06,
|
|
"p50_s": 4.191999323666096e-06,
|
|
"p90_s": 4.870421253144741e-06,
|
|
"p95_s": 4.955223994329571e-06,
|
|
"p99_s": 5.023066187277436e-06,
|
|
"min_s": 4.128029104322195e-06,
|
|
"max_s": 5.040026735514402e-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": 1.0901324761410555e-05,
|
|
"p50_s": 9.343959391117096e-06,
|
|
"p90_s": 1.3158377259969712e-05,
|
|
"p95_s": 1.3635179493576289e-05,
|
|
"p99_s": 1.401662128046155e-05,
|
|
"min_s": 9.248033165931702e-06,
|
|
"max_s": 1.4111981727182865e-05
|
|
},
|
|
"linear.attn_out_proj.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.471989960720142e-06,
|
|
"p50_s": 3.3919932320713997e-06,
|
|
"p90_s": 3.6095967516303063e-06,
|
|
"p95_s": 3.6367971915751694e-06,
|
|
"p99_s": 3.6585575435310603e-06,
|
|
"min_s": 3.3599790185689926e-06,
|
|
"max_s": 3.663997631520033e-06
|
|
},
|
|
"linear.attn_out_proj.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.002189169346820563,
|
|
"p50_s": 0.0021924120374023914,
|
|
"p90_s": 0.002197237592190504,
|
|
"p95_s": 0.0021978407865390183,
|
|
"p99_s": 0.0021983233420178293,
|
|
"min_s": 0.0021766520221717656,
|
|
"max_s": 0.0021984439808875322
|
|
},
|
|
"linear.attn_out_proj.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.1679988044003645e-05,
|
|
"p50_s": 1.0255957022309303e-05,
|
|
"p90_s": 1.3852771371603011e-05,
|
|
"p95_s": 1.4302373165264725e-05,
|
|
"p99_s": 1.4662054600194098e-05,
|
|
"min_s": 1.0032032150775194e-05,
|
|
"max_s": 1.475197495892644e-05
|
|
},
|
|
"linear.attn_out_proj.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002833104692399502,
|
|
"p50_s": 0.0028447150252759457,
|
|
"p90_s": 0.0028486958239227534,
|
|
"p95_s": 0.002849193423753604,
|
|
"p99_s": 0.002849591503618285,
|
|
"min_s": 0.0028049080283381045,
|
|
"max_s": 0.0028496910235844553
|
|
},
|
|
"linear.attn_out_proj.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 7.119980485488971e-06,
|
|
"p50_s": 6.671994924545288e-06,
|
|
"p90_s": 7.926393300294876e-06,
|
|
"p95_s": 8.083193097263575e-06,
|
|
"p99_s": 8.208632934838535e-06,
|
|
"min_s": 6.447953637689352e-06,
|
|
"max_s": 8.239992894232273e-06
|
|
},
|
|
"linear.attn_out_proj.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.008127884687079737,
|
|
"p50_s": 0.007850803027395159,
|
|
"p90_s": 0.008538391825277358,
|
|
"p95_s": 0.008624340425012632,
|
|
"p99_s": 0.008693099304800852,
|
|
"min_s": 0.007822562009096146,
|
|
"max_s": 0.008710289024747908
|
|
},
|
|
"linear.attn_out_proj.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.770668697853883e-06,
|
|
"p50_s": 5.69602707400918e-06,
|
|
"p90_s": 6.054400000721216e-06,
|
|
"p95_s": 6.099196616560221e-06,
|
|
"p99_s": 6.1350339092314244e-06,
|
|
"min_s": 5.471985787153244e-06,
|
|
"max_s": 6.143993232399225e-06
|
|
},
|
|
"attn_out_proj": {
|
|
"count": 3,
|
|
"mean_s": 0.013236238659980396,
|
|
"p50_s": 0.01298735401360318,
|
|
"p90_s": 0.01366794117493555,
|
|
"p95_s": 0.013753014570102095,
|
|
"p99_s": 0.013821073286235334,
|
|
"min_s": 0.012883274001069367,
|
|
"max_s": 0.013838087965268642
|
|
},
|
|
"gate_msa": {
|
|
"count": 3,
|
|
"mean_s": 0.00891906803008169,
|
|
"p50_s": 0.008913073048461229,
|
|
"p90_s": 0.009082570637110621,
|
|
"p95_s": 0.009103757835691794,
|
|
"p99_s": 0.009120707594556734,
|
|
"min_s": 0.00871918600751087,
|
|
"max_s": 0.009124945034272969
|
|
},
|
|
"norm2": {
|
|
"count": 3,
|
|
"mean_s": 0.0035535029989356795,
|
|
"p50_s": 0.0035576900118030608,
|
|
"p90_s": 0.0036055227858014403,
|
|
"p95_s": 0.003611501882551238,
|
|
"p99_s": 0.003616285159951076,
|
|
"min_s": 0.0034853380057029426,
|
|
"max_s": 0.0036174809793010354
|
|
},
|
|
"modulate_mlp": {
|
|
"count": 3,
|
|
"mean_s": 0.010866791980030635,
|
|
"p50_s": 0.010817836970090866,
|
|
"p90_s": 0.010982598597183823,
|
|
"p95_s": 0.011003193800570443,
|
|
"p99_s": 0.011019669963279738,
|
|
"min_s": 0.010758749966043979,
|
|
"max_s": 0.011023789003957063
|
|
},
|
|
"linear.mlp_fc1.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.4586680208643275e-06,
|
|
"p50_s": 4.352012183517218e-06,
|
|
"p90_s": 4.8128305934369566e-06,
|
|
"p95_s": 4.870432894676924e-06,
|
|
"p99_s": 4.916514735668898e-06,
|
|
"min_s": 4.0959566831588745e-06,
|
|
"max_s": 4.928035195916891e-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": 9.039998985826969e-06,
|
|
"p50_s": 8.944014552980661e-06,
|
|
"p90_s": 9.775965008884668e-06,
|
|
"p95_s": 9.879958815872669e-06,
|
|
"p99_s": 9.963153861463071e-06,
|
|
"min_s": 8.192029781639576e-06,
|
|
"max_s": 9.98395262286067e-06
|
|
},
|
|
"linear.mlp_fc1.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.536018387724956e-06,
|
|
"p50_s": 3.648048732429743e-06,
|
|
"p90_s": 3.660807851701975e-06,
|
|
"p95_s": 3.662402741611004e-06,
|
|
"p99_s": 3.663678653538227e-06,
|
|
"min_s": 3.296008799225092e-06,
|
|
"max_s": 3.663997631520033e-06
|
|
},
|
|
"linear.mlp_fc1.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0017847650063534577,
|
|
"p50_s": 0.0017553410143591464,
|
|
"p90_s": 0.0018440961954183876,
|
|
"p95_s": 0.0018551905930507928,
|
|
"p99_s": 0.001864066111156717,
|
|
"min_s": 0.001732669014018029,
|
|
"max_s": 0.001866284990683198
|
|
},
|
|
"linear.mlp_fc1.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.0303997745116552e-05,
|
|
"p50_s": 1.047999830916524e-05,
|
|
"p90_s": 1.0748824570327997e-05,
|
|
"p95_s": 1.0782427852973343e-05,
|
|
"p99_s": 1.0809310479089617e-05,
|
|
"min_s": 9.615963790565729e-06,
|
|
"max_s": 1.0816031135618687e-05
|
|
},
|
|
"linear.mlp_fc1.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.0021633559954352677,
|
|
"p50_s": 0.0021477399859577417,
|
|
"p90_s": 0.002185807190835476,
|
|
"p95_s": 0.0021905655914451925,
|
|
"p99_s": 0.0021943723119329664,
|
|
"min_s": 0.002147004008293152,
|
|
"max_s": 0.0021953239920549095
|
|
},
|
|
"linear.mlp_fc1.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.517337169498205e-06,
|
|
"p50_s": 6.352027412503958e-06,
|
|
"p90_s": 6.799993570894003e-06,
|
|
"p95_s": 6.8559893406927586e-06,
|
|
"p99_s": 6.900785956531763e-06,
|
|
"min_s": 6.287998985499144e-06,
|
|
"max_s": 6.911985110491514e-06
|
|
},
|
|
"linear.mlp_fc1.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.03497030300786719,
|
|
"p50_s": 0.035002851975150406,
|
|
"p90_s": 0.035098020010627805,
|
|
"p95_s": 0.035109916015062484,
|
|
"p99_s": 0.03511943281861022,
|
|
"min_s": 0.034786245028954,
|
|
"max_s": 0.035121812019497156
|
|
},
|
|
"linear.mlp_fc1.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.631998647004366e-06,
|
|
"p50_s": 5.536014214158058e-06,
|
|
"p90_s": 5.855970084667206e-06,
|
|
"p95_s": 5.895964568480849e-06,
|
|
"p99_s": 5.927960155531764e-06,
|
|
"min_s": 5.424022674560547e-06,
|
|
"max_s": 5.935959052294493e-06
|
|
},
|
|
"mlp_fc1": {
|
|
"count": 3,
|
|
"mean_s": 0.03900541601857791,
|
|
"p50_s": 0.038999997021164745,
|
|
"p90_s": 0.03917357780737803,
|
|
"p95_s": 0.039195275405654685,
|
|
"p99_s": 0.03921263348427601,
|
|
"min_s": 0.03879927803063765,
|
|
"max_s": 0.039216973003931344
|
|
},
|
|
"mlp_swiglu": {
|
|
"count": 3,
|
|
"mean_s": 0.0237522956643564,
|
|
"p50_s": 0.023783078999258578,
|
|
"p90_s": 0.023799194977618756,
|
|
"p95_s": 0.023801209474913775,
|
|
"p99_s": 0.023802821072749794,
|
|
"min_s": 0.023670584021601826,
|
|
"max_s": 0.023803223972208798
|
|
},
|
|
"linear.mlp_fc2.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.831973152856032e-06,
|
|
"p50_s": 5.007954314351082e-06,
|
|
"p90_s": 5.251169204711914e-06,
|
|
"p95_s": 5.281571066007018e-06,
|
|
"p99_s": 5.305892555043101e-06,
|
|
"min_s": 4.175992216914892e-06,
|
|
"max_s": 5.311972927302122e-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": 8.869334124028683e-06,
|
|
"p50_s": 8.800008799880743e-06,
|
|
"p90_s": 9.478384163230659e-06,
|
|
"p95_s": 9.563181083649398e-06,
|
|
"p99_s": 9.63101861998439e-06,
|
|
"min_s": 8.160015568137169e-06,
|
|
"max_s": 9.647978004068136e-06
|
|
},
|
|
"linear.mlp_fc2.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.3919932320713997e-06,
|
|
"p50_s": 3.3760443329811096e-06,
|
|
"p90_s": 3.452785313129425e-06,
|
|
"p95_s": 3.4623779356479647e-06,
|
|
"p99_s": 3.470052033662796e-06,
|
|
"min_s": 3.3279648050665855e-06,
|
|
"max_s": 3.471970558166504e-06
|
|
},
|
|
"linear.mlp_fc2.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.004335538988622527,
|
|
"p50_s": 0.004312007979024202,
|
|
"p90_s": 0.004394120781216771,
|
|
"p95_s": 0.004404384881490842,
|
|
"p99_s": 0.004412596161710098,
|
|
"min_s": 0.004279960005078465,
|
|
"max_s": 0.004414648981764913
|
|
},
|
|
"linear.mlp_fc2.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.5887975071867306e-05,
|
|
"p50_s": 1.0015966836363077e-05,
|
|
"p90_s": 2.4607975501567125e-05,
|
|
"p95_s": 2.643197658471763e-05,
|
|
"p99_s": 2.7891177451238038e-05,
|
|
"min_s": 9.391980711370707e-06,
|
|
"max_s": 2.8255977667868137e-05
|
|
},
|
|
"linear.mlp_fc2.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.005852811639973273,
|
|
"p50_s": 0.005791350966319442,
|
|
"p90_s": 0.005944655789062381,
|
|
"p95_s": 0.005963818891905248,
|
|
"p99_s": 0.005979149374179542,
|
|
"min_s": 0.005784101958852261,
|
|
"max_s": 0.0059829819947481155
|
|
},
|
|
"linear.mlp_fc2.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.6400001135965186e-06,
|
|
"p50_s": 6.6400389187037945e-06,
|
|
"p90_s": 6.742391269654036e-06,
|
|
"p95_s": 6.7551853135228155e-06,
|
|
"p99_s": 6.7654205486178394e-06,
|
|
"min_s": 6.511982064694166e-06,
|
|
"max_s": 6.767979357391596e-06
|
|
},
|
|
"linear.mlp_fc2.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.056130341331784926,
|
|
"p50_s": 0.0560961120063439,
|
|
"p90_s": 0.0565045592142269,
|
|
"p95_s": 0.056555615115212274,
|
|
"p99_s": 0.056596459836000576,
|
|
"min_s": 0.05568824097281322,
|
|
"max_s": 0.05660667101619765
|
|
},
|
|
"linear.mlp_fc2.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.583996729304393e-06,
|
|
"p50_s": 5.599984433501959e-06,
|
|
"p90_s": 5.676771979779005e-06,
|
|
"p95_s": 5.686370423063636e-06,
|
|
"p99_s": 5.694049177691341e-06,
|
|
"min_s": 5.456036888062954e-06,
|
|
"max_s": 5.695968866348267e-06
|
|
},
|
|
"mlp_fc2": {
|
|
"count": 3,
|
|
"mean_s": 0.06640448931527014,
|
|
"p50_s": 0.06639227096457034,
|
|
"p90_s": 0.06670863500330597,
|
|
"p95_s": 0.06674818050814792,
|
|
"p99_s": 0.06677981691202148,
|
|
"min_s": 0.06603347096825019,
|
|
"max_s": 0.06678772601298988
|
|
},
|
|
"gate_mlp": {
|
|
"count": 3,
|
|
"mean_s": 0.01474788433794553,
|
|
"p50_s": 0.008609394019003958,
|
|
"p90_s": 0.023345611605327575,
|
|
"p95_s": 0.025187638803618028,
|
|
"p99_s": 0.02666126056225039,
|
|
"min_s": 0.008604592992924154,
|
|
"max_s": 0.02702966600190848
|
|
},
|
|
"block_total": {
|
|
"count": 3,
|
|
"mean_s": 0.5313544006785378,
|
|
"p50_s": 0.52417680202052,
|
|
"p90_s": 0.5474372819997371,
|
|
"p95_s": 0.5503448419971392,
|
|
"p99_s": 0.5526708899950609,
|
|
"min_s": 0.5166339980205521,
|
|
"max_s": 0.5532524019945413
|
|
}
|
|
},
|
|
"module_forward": {
|
|
"count": 3,
|
|
"mean_s": 0.46708817364803207,
|
|
"p50_s": 0.4669776560040191,
|
|
"p90_s": 0.467878839164041,
|
|
"p95_s": 0.4679914870590437,
|
|
"p99_s": 0.4680816053750459,
|
|
"min_s": 0.4661827299860306,
|
|
"max_s": 0.46810413495404646
|
|
},
|
|
"module_forward_checksum": 606380672.0,
|
|
"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": 512.799,
|
|
"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": 465.9349999999995,
|
|
"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": 369.58299999999963,
|
|
"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::rms_norm",
|
|
"count": 2,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 335.5350000000003,
|
|
"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": 325.1509999999997,
|
|
"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::_to_copy",
|
|
"count": 8,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 236.92700000000127,
|
|
"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": 206.33499999999913,
|
|
"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": 194.91200000000026,
|
|
"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::copy_",
|
|
"count": 8,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 159.58399999999983,
|
|
"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": 136.27199999999993,
|
|
"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::empty",
|
|
"count": 29,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 133.95199999999943,
|
|
"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": 133.83899999999994,
|
|
"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": "aten::mean",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 125.10400000000027,
|
|
"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": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 112.76800000000003,
|
|
"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": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 111.69599999999991,
|
|
"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": "aten::zeros",
|
|
"count": 3,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 102.35199999999986,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], [], [], []]"
|
|
},
|
|
{
|
|
"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": 87.96799999999985,
|
|
"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::mul",
|
|
"count": 4,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 80.92799999999943,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 2048,
|
|
"input_shapes": "[[], []]"
|
|
},
|
|
{
|
|
"key": "aten::slice",
|
|
"count": 19,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 80.81600000000071,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 5376], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::quantize_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 65.34299999999985,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 5376], [], [], [], []]"
|
|
}
|
|
]
|
|
} |