{ "combined_t2va_n20": [ { "audit": { "error_line_count": 0, "fatal_error": false, "shape_ok": true, "status": "passed" }, "case": "allgather-dp1-sp8-t2va-50step-n20", "concurrency": 1, "denoise_updates_executed": 49, "dp": 1, "energy_wh_per_video": 68.08272440027358, "engine_init_s_values": [ 326.9554082560353, 322.72203186899424 ], "engine_lifecycles": 2, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 26.365234375, "max_temperature_c": 74.0, "mean_gpu_util_pct": 90.40346996313164, "mean_node_power_w": 7069.559229830329, "mean_sm_clock_mhz": 1821.499615050965, "mean_wave_s": 34.669439355446954, "measured_peak_memory_gib": 26.365234375, "measured_videos": 20, "mode": "allgather", "peak_node_power_w": 8535.58, "request_p50_s": 34.43105830694549, "request_p95_s": 35.141176707076376, "request_samples": 20, "request_values_s": [ 34.43183084088378, 34.112008680123836, 34.09530321205966, 34.70891568507068, 34.05489400285296, 36.0866008840967, 34.10303023084998, 34.967818598030135, 34.41919455095194, 34.593092336086556, 34.35336302802898, 34.52474875305779, 34.76744698896073, 34.4302857730072, 34.13301039696671, 34.35436272015795, 35.09141753986478, 34.064232495147735, 34.56138198100962, 35.07870242907666 ], "source_cases": [ "allgather-dp1-sp8-t2va-50step-r5", "soak-allgather-dp1-sp8-t2va-50step-r15" ], "sp": 8, "steps_requested": 50, "sustained_videos_per_hour": 103.83784874889821, "task": "t2va", "telemetry_coverage_pct": 100.0, "topology": "dp1-sp8", "warmup_wave_s_values": [ 68.23246542993002, 80.27618861896917 ], "wave_cv_pct": 1.3798904664862208, "wave_max_s": 36.21506123198196, "wave_min_s": 34.16365527105518, "wave_p50_s": 34.553416537470184, "wave_p95_s": 35.25245706331916, "wave_samples": 20, "wave_values_s": [ 34.542716522933915, 34.24266396299936, 34.22889453009702, 34.83860694314353, 34.16365527105518, 36.21506123198196, 34.241454757982865, 35.114581926958635, 34.560933315893635, 34.69453239394352, 34.46163991512731, 34.6710405908525, 34.8828783580102, 34.54589975904673, 34.241072338074446, 34.501001303084195, 35.20179368602112, 34.174797968007624, 34.67267275089398, 35.19288958283141 ], "width": 1344 }, { "audit": { "error_line_count": 0, "fatal_error": false, "shape_ok": true, "status": "passed" }, "case": "allgather-dp4-sp2-t2va-50step-n20", "concurrency": 4, "denoise_updates_executed": 49, "dp": 4, "energy_wh_per_video": 51.75644020222868, "engine_init_s_values": [ 311.8539885180071, 310.22449679416604 ], "engine_lifecycles": 2, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 25.107421875, "max_temperature_c": 78.0, "mean_gpu_util_pct": 93.90196124823339, "mean_node_power_w": 7870.185761523024, "mean_sm_clock_mhz": 1615.3053389359984, "mean_wave_s": 94.80295536004705, "measured_peak_memory_gib": 25.107421875, "measured_videos": 80, "mode": "allgather", "peak_node_power_w": 8719.47, "request_p50_s": 94.35054075403605, "request_p95_s": 94.92447212070692, "request_samples": 80, "request_values_s": [ 94.17101976904087, 94.17073291400447, 94.17059959704056, 94.17050093901344, 94.28023565607145, 94.27985115186311, 94.06823320197873, 94.27973123500124, 93.83675427408889, 93.83531144005246, 93.8363644410856, 93.8363570610527, 93.87086566514336, 93.87051565083675, 93.87055804999545, 93.870468033012, 94.0576149970293, 93.9937849259004, 93.99374685715884, 94.0571317251306, 94.47121634706855, 94.66622212808579, 94.66611363901757, 94.66599686001427, 94.85103903105482, 94.85062345000915, 94.8502634849865, 94.63475876813754, 94.92419367516413, 94.92395213898271, 94.72288371506147, 94.92379026301205, 94.93037632503547, 94.92998701287434, 94.92977987602353, 94.92976258601993, 94.3964108589571, 94.39619075111113, 94.3959930029232, 94.39601668296382, 94.35429618996568, 94.35393299395218, 94.35390495508909, 94.14929260499775, 94.35099808499217, 94.35061551816761, 94.35046598990448, 94.34920608694665, 94.43901469791308, 94.43867649207823, 94.43863055412658, 94.4385010660626, 94.18403327884153, 94.18375919293612, 94.18371233390644, 94.18352163606323, 94.33908008597791, 94.54142624605447, 94.54127552895807, 94.54125517886132, 94.27269358700141, 94.2723609409295, 94.27231317199767, 94.27217112388462, 94.59221268096007, 94.5918444271665, 94.59153180080466, 94.39337701397017, 94.23428387497552, 94.03756350209005, 94.23394960979931, 94.23391300090589, 94.43688919418491, 94.64471982698888, 94.6445708789397, 94.64454760914668, 94.29002399696037, 94.2896244241856, 94.28939079819247, 94.07687358511612 ], "source_cases": [ "allgather-dp4-sp2-t2va-50step-r5", "soak-allgather-dp4-sp2-t2va-50step-r15" ], "sp": 2, "steps_requested": 50, "sustained_videos_per_hour": 151.89399892979088, "task": "t2va", "telemetry_coverage_pct": 99.88951428485835, "topology": "dp4-sp2", "warmup_wave_s_values": [ 111.8642603319604, 108.46901519596577 ], "wave_cv_pct": 0.3459079436090338, "wave_max_s": 95.39917947002687, "wave_min_s": 94.15086788102053, "wave_p50_s": 94.72865000448655, "wave_p95_s": 95.30612910840428, "wave_samples": 20, "wave_values_s": [ 94.60343311587349, 94.71106801903807, 94.15086788102053, 94.18365128501318, 94.37644968321547, 95.1047255678568, 95.30123172095045, 95.39917947002687, 95.26087002479471, 94.71903971489519, 94.82264852989465, 94.65760229318403, 94.90892298403196, 94.6725157070905, 95.02406334900297, 94.7351494859904, 95.04445270099677, 94.70667615602724, 94.95440898905508, 94.7221505229827 ], "width": 1344 }, { "audit": { "error_line_count": 2, "fatal_error": false, "shape_ok": true, "status": "passed" }, "case": "ranklocal-dp8-sp1-t2va-50step-n20", "concurrency": 8, "denoise_updates_executed": 49, "dp": 8, "energy_wh_per_video": 43.96698965041099, "engine_init_s_values": [ 420.8009794289246, 420.9326464571059 ], "engine_lifecycles": 2, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 20.0546875, "max_temperature_c": 78.0, "mean_gpu_util_pct": 94.54279822310534, "mean_node_power_w": 8080.259023153472, "mean_sm_clock_mhz": 1520.5975722862177, "mean_wave_s": 156.70895906535443, "measured_peak_memory_gib": 20.0546875, "measured_videos": 160, "mode": "ranklocal", "peak_node_power_w": 8780.83, "request_p50_s": 155.9675515484996, "request_p95_s": 156.22881248188204, "request_samples": 160, "request_values_s": [ 156.09283326216973, 156.09258772502653, 156.0923035289161, 156.09241906786337, 156.0924201279413, 156.0921921711415, 156.09218069096096, 155.62114030402154, 156.3031671489589, 156.3027737857774, 156.3027617270127, 156.30260170786642, 156.3024390710052, 156.3026372389868, 156.3023941020947, 156.30238190293312, 156.0568288420327, 156.05649811797775, 155.85895555000752, 156.05631653987803, 156.05636846879497, 155.85741406399757, 156.05628082016483, 156.05625689099543, 155.96664117788896, 155.96618592413142, 155.96604107692838, 155.9659096391406, 155.50678721209988, 155.96590561117046, 155.9659023489803, 155.96573955309577, 156.05211325804703, 156.05160844698548, 156.0514844378922, 155.56675495905802, 156.05146950902417, 156.05151317897253, 156.0514317189809, 156.0512709619943, 156.14487677114084, 156.14460928505287, 156.14467092510313, 155.68986208294518, 156.14449365786277, 156.14446876686998, 156.14434288116172, 156.14415347389877, 155.99486353690736, 155.99475062894635, 155.9946401210036, 155.9946259309072, 155.99446358392015, 155.9942625781987, 155.51326021691784, 155.99422487802804, 156.04613001295365, 156.04586411896162, 156.0458164010197, 156.04581121099181, 155.55696337507106, 156.04566201288253, 156.04547558678314, 156.0454567570705, 155.88891655998304, 155.8886938248761, 155.88875959487632, 155.88862597499974, 155.88852124894038, 155.88839265191928, 155.43807193008251, 155.88844056101516, 155.96787663619034, 155.96768250106834, 155.83641192200594, 155.9673461469356, 155.9674205959309, 155.96740144491196, 155.9672238889616, 155.83728947513737, 155.75283777993172, 156.2249404070899, 156.2248029599432, 156.2246645039413, 156.22477230080403, 156.22466129413806, 156.2245281659998, 156.22436873987317, 156.13527295482345, 156.13124554906972, 156.13495098007843, 156.13482284289785, 156.1348655319307, 156.13483811286278, 156.1346890348941, 156.1345515579451, 155.98085280298255, 155.46119379089214, 155.98034409293905, 155.9800725770183, 155.9797734420281, 155.97966832504608, 155.97964327596128, 155.97952164895833, 155.6094499239698, 156.10996113182046, 156.1098128540907, 156.10971948807128, 156.10958507889882, 156.1094855510164, 156.10963307786733, 156.10947553999722, 155.7665666458197, 155.7662918008864, 155.76621977286413, 155.28089245292358, 155.76613510516472, 155.76593491900712, 155.7657902110368, 155.76586777996272, 155.81967494799756, 155.41064266185276, 155.8191832168959, 155.8190922990907, 155.81780481315218, 155.41156434593722, 155.81888041296043, 155.81879498600028, 155.69791654706933, 155.69738041702658, 155.69736773706973, 155.69727411912754, 155.69727988890372, 155.69727715896443, 155.19995941780508, 155.69695729506202, 155.77147456491366, 155.77101823408157, 155.77099137404002, 155.77102703391574, 155.7709470440168, 155.7708428769838, 155.77081237686798, 155.27326258085668, 155.9037292469293, 155.9033278040588, 155.40905180596747, 155.90311726694927, 155.90291799092665, 155.9029618599452, 155.90287275216542, 155.90292202099226, 155.46773529495113, 155.92744814488105, 155.9275860011112, 155.92731187702157, 155.92731220601127, 155.92715210001916, 155.92723404802382, 155.9271143109072 ], "source_cases": [ "ranklocal-dp8-sp1-t2va-50step-r5", "soak-ranklocal-dp8-sp1-t2va-50step-r15" ], "sp": 1, "steps_requested": 50, "sustained_videos_per_hour": 183.78017550348957, "task": "t2va", "telemetry_coverage_pct": 100.0, "topology": "dp8-sp1", "warmup_wave_s_values": [ 166.2483102818951, 161.5930544710718 ], "wave_cv_pct": 0.1527385373769954, "wave_max_s": 157.2165887071751, "wave_min_s": 156.29264628794044, "wave_p50_s": 156.74228433542885, "wave_p95_s": 157.0282817220781, "wave_samples": 20, "wave_values_s": [ 156.73276200098917, 157.2165887071751, 156.67835334688425, 156.92485627299175, 156.96818466996774, 156.7593610167969, 156.61332127009518, 156.95707907597534, 156.51728480495512, 156.54197374801151, 156.85744950803928, 156.8194091389887, 156.5648122720886, 157.01837082812563, 156.34758708183654, 156.44614280201495, 156.29264628794044, 156.38694050116464, 156.75180666986853, 156.7842513031792 ], "width": 1344 } ], "expected_failed_case": { "case": "allgather-dp4-sp2-ref2va-50step-failed-subgroup-src0", "error_line_count": 702, "expected_failure": true, "group_rank_error": true, "oom": false }, "formal_cases": [ { "audit": { "cleanup_status": "delayed-reap", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 5, "shape_ok": true, "status": "passed" }, "case": "ranklocal-dp1-sp8-t2va-50step-r5", "concurrency": 1, "denoise_updates_executed": 49, "dp": 1, "energy_wh_per_video": 94.06434989906225, "engine_init_s": 348.31003612000495, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 26.263671875, "max_temperature_c": 60.0, "mean_gpu_util_pct": 95.86309523809524, "mean_node_power_w": 4290.892564082633, "mean_sm_clock_mhz": 2032.0, "mean_wave_s": 78.91867499351501, "measured_peak_memory_gib": 26.263671875, "measured_videos": 5, "mode": "ranklocal", "peak_node_power_w": 6504.84, "per_gpu_lifecycle_peak_memory_gib": { "0": 22.078125, "1": 24.5546875, "2": 25.794921875, "3": 26.263671875, "4": 26.263671875, "5": 25.705078125, "6": 24.892578125, "7": 25.826171875 }, "per_gpu_measured_peak_memory_gib": { "0": 22.078125, "1": 24.5546875, "2": 25.794921875, "3": 26.263671875, "4": 26.263671875, "5": 25.705078125, "6": 24.892578125, "7": 25.826171875 }, "request_p50_s": 78.78022901318036, "request_p95_s": 78.86696606408805, "request_samples": 5, "request_values_s": [ 78.8772438399028, 78.78022901318036, 78.76366679999046, 78.7488103620708, 78.82585496082902 ], "shutdown_s": 30.068825720110908, "sp": 8, "steps_requested": 50, "sustained_videos_per_hour": 45.61657934951167, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1726, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 777, "topology": "dp1-sp8", "warmup_wave_s": 105.89164609392174, "wave_cv_pct": 0.05943924123565115, "wave_max_s": 78.99690022296272, "wave_min_s": 78.85969600989483, "wave_p50_s": 78.9212885349989, "wave_p95_s": 78.98388303234242, "wave_samples": 5, "wave_values_s": [ 78.99690022296272, 78.88367592985742, 78.9212885349989, 78.85969600989483, 78.9318142698612 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 5, "shape_ok": true, "status": "passed" }, "case": "allgather-dp1-sp8-t2va-50step-r5", "concurrency": 1, "denoise_updates_executed": 49, "dp": 1, "energy_wh_per_video": 67.74650535921454, "engine_init_s": 326.9554082560353, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 24.935546875, "max_temperature_c": 74.0, "mean_gpu_util_pct": 90.52227011494253, "mean_node_power_w": 7089.064422779068, "mean_sm_clock_mhz": 1814.6364942528735, "mean_wave_s": 34.4033074460458, "measured_peak_memory_gib": 23.52734375, "measured_videos": 5, "mode": "allgather", "peak_node_power_w": 8535.58, "per_gpu_lifecycle_peak_memory_gib": { "0": 22.205078125, "1": 23.52734375, "2": 24.935546875, "3": 23.52734375, "4": 23.52734375, "5": 24.908203125, "6": 24.935546875, "7": 24.921875 }, "per_gpu_measured_peak_memory_gib": { "0": 22.205078125, "1": 23.52734375, "2": 23.52734375, "3": 23.52734375, "4": 23.52734375, "5": 23.52734375, "6": 23.52734375, "7": 23.52734375 }, "request_p50_s": 34.112008680123836, "request_p95_s": 34.6534987162333, "request_samples": 5, "request_values_s": [ 34.43183084088378, 34.112008680123836, 34.09530321205966, 34.70891568507068, 34.05489400285296 ], "shutdown_s": 23.42470257799141, "sp": 8, "steps_requested": 50, "sustained_videos_per_hour": 104.64110189538685, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 630, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 174, "topology": "dp1-sp8", "warmup_wave_s": 68.23246542993002, "wave_cv_pct": 0.7383009954004893, "wave_max_s": 34.83860694314353, "wave_min_s": 34.16365527105518, "wave_p50_s": 34.24266396299936, "wave_p95_s": 34.77942885910161, "wave_samples": 5, "wave_values_s": [ 34.542716522933915, 34.24266396299936, 34.22889453009702, 34.83860694314353, 34.16365527105518 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 20, "shape_ok": true, "status": "passed" }, "case": "ranklocal-dp4-sp2-t2va-50step-r5", "concurrency": 4, "denoise_updates_executed": 49, "dp": 4, "energy_wh_per_video": 51.865238338504795, "engine_init_s": 357.1848510671407, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 23.912109375, "max_temperature_c": 77.0, "mean_gpu_util_pct": 94.22685669456067, "mean_node_power_w": 7739.723069663594, "mean_sm_clock_mhz": 1640.114670502092, "mean_wave_s": 96.49688759343698, "measured_peak_memory_gib": 23.912109375, "measured_videos": 20, "mode": "ranklocal", "peak_node_power_w": 8678.800000000001, "per_gpu_lifecycle_peak_memory_gib": { "0": 22.591796875, "1": 21.16015625, "2": 19.830078125, "3": 21.16015625, "4": 21.25390625, "5": 23.912109375, "6": 22.076171875, "7": 21.16015625 }, "per_gpu_measured_peak_memory_gib": { "0": 19.822265625, "1": 21.16015625, "2": 19.822265625, "3": 21.16015625, "4": 21.25390625, "5": 23.912109375, "6": 19.822265625, "7": 21.16015625 }, "request_p50_s": 96.2040189160034, "request_p95_s": 96.45805224406067, "request_samples": 20, "request_values_s": [ 96.45829877303913, 96.45803926885128, 96.45802065893076, 96.45793430996127, 96.2355456170626, 96.2364099000115, 96.236134394072, 96.235898520099, 95.86966452584602, 95.6738558060024, 95.86923715402372, 95.86926569300704, 95.92915286705829, 95.7276580107864, 95.92872075596824, 95.92873368482105, 96.20430390001275, 96.20407034503296, 96.20396748697385, 96.20380566897802 ], "shutdown_s": 30.071337383938953, "sp": 2, "steps_requested": 50, "sustained_videos_per_hour": 149.22761095332345, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1985, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 956, "topology": "dp4-sp2", "warmup_wave_s": 136.4233135059476, "wave_cv_pct": 0.2556069403240488, "wave_max_s": 96.80010823300108, "wave_min_s": 96.18466887599789, "wave_p50_s": 96.53021865617484, "wave_p95_s": 96.78464542301373, "wave_samples": 5, "wave_values_s": [ 96.80010823300108, 96.72279418306425, 96.18466887599789, 96.24664801894687, 96.53021865617484 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 20, "shape_ok": true, "status": "passed" }, "case": "allgather-dp4-sp2-t2va-50step-r5", "concurrency": 4, "denoise_updates_executed": 49, "dp": 4, "energy_wh_per_video": 51.66884957844369, "engine_init_s": 311.8539885180071, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 24.974609375, "max_temperature_c": 78.0, "mean_gpu_util_pct": 94.11017628205128, "mean_node_power_w": 7881.261385587856, "mean_sm_clock_mhz": 1611.5464743589744, "mean_wave_s": 94.40509399683215, "measured_peak_memory_gib": 24.974609375, "measured_videos": 20, "mode": "allgather", "peak_node_power_w": 8714.59, "per_gpu_lifecycle_peak_memory_gib": { "0": 23.45703125, "1": 23.68359375, "2": 23.33203125, "3": 24.974609375, "4": 23.11328125, "5": 24.880859375, "6": 23.35546875, "7": 24.734375 }, "per_gpu_measured_peak_memory_gib": { "0": 23.025390625, "1": 23.68359375, "2": 22.744140625, "3": 24.974609375, "4": 22.837890625, "5": 24.880859375, "6": 22.650390625, "7": 24.734375 }, "request_p50_s": 94.0254583255155, "request_p95_s": 94.27987037707352, "request_samples": 20, "request_values_s": [ 94.17101976904087, 94.17073291400447, 94.17059959704056, 94.17050093901344, 94.28023565607145, 94.27985115186311, 94.06823320197873, 94.27973123500124, 93.83675427408889, 93.83531144005246, 93.8363644410856, 93.8363570610527, 93.87086566514336, 93.87051565083675, 93.87055804999545, 93.870468033012, 94.0576149970293, 93.9937849259004, 93.99374685715884, 94.0571317251306 ], "shutdown_s": 29.547646628925577, "sp": 2, "steps_requested": 50, "sustained_videos_per_hour": 152.5341418597942, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1821, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 936, "topology": "dp4-sp2", "warmup_wave_s": 111.8642603319604, "wave_cv_pct": 0.23564622548244135, "wave_max_s": 94.71106801903807, "wave_min_s": 94.15086788102053, "wave_p50_s": 94.37644968321547, "wave_p95_s": 94.68954103840515, "wave_samples": 5, "wave_values_s": [ 94.60343311587349, 94.71106801903807, 94.15086788102053, 94.18365128501318, 94.37644968321547 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 40, "shape_ok": true, "status": "passed" }, "case": "ranklocal-dp8-sp1-t2va-50step-r5", "concurrency": 8, "denoise_updates_executed": 49, "dp": 8, "energy_wh_per_video": 43.96514660893554, "engine_init_s": 420.8009794289246, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 20.02734375, "max_temperature_c": 73.0, "mean_gpu_util_pct": 94.38950143815916, "mean_node_power_w": 8069.86866347063, "mean_sm_clock_mhz": 1522.1229626078618, "mean_wave_s": 156.9041489996016, "measured_peak_memory_gib": 20.02734375, "measured_videos": 40, "mode": "ranklocal", "peak_node_power_w": 8780.83, "per_gpu_lifecycle_peak_memory_gib": { "0": 18.078125, "1": 17.19140625, "2": 17.193359375, "3": 19.984375, "4": 20.02734375, "5": 18.951171875, "6": 17.193359375, "7": 17.19140625 }, "per_gpu_measured_peak_memory_gib": { "0": 18.078125, "1": 17.19140625, "2": 17.193359375, "3": 19.984375, "4": 20.02734375, "5": 18.951171875, "6": 17.193359375, "7": 17.19140625 }, "request_p50_s": 156.05626885558013, "request_p95_s": 156.30276232995092, "request_samples": 40, "request_values_s": [ 156.09283326216973, 156.09258772502653, 156.0923035289161, 156.09241906786337, 156.0924201279413, 156.0921921711415, 156.09218069096096, 155.62114030402154, 156.3031671489589, 156.3027737857774, 156.3027617270127, 156.30260170786642, 156.3024390710052, 156.3026372389868, 156.3023941020947, 156.30238190293312, 156.0568288420327, 156.05649811797775, 155.85895555000752, 156.05631653987803, 156.05636846879497, 155.85741406399757, 156.05628082016483, 156.05625689099543, 155.96664117788896, 155.96618592413142, 155.96604107692838, 155.9659096391406, 155.50678721209988, 155.96590561117046, 155.9659023489803, 155.96573955309577, 156.05211325804703, 156.05160844698548, 156.0514844378922, 155.56675495905802, 156.05146950902417, 156.05151317897253, 156.0514317189809, 156.0512709619943 ], "shutdown_s": 30.012816147878766, "sp": 1, "steps_requested": 50, "sustained_videos_per_hour": 183.55155159136757, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1745, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 1043, "topology": "dp8-sp1", "warmup_wave_s": 166.2483102818951, "wave_cv_pct": 0.12176376929582645, "wave_max_s": 157.2165887071751, "wave_min_s": 156.67835334688425, "wave_p50_s": 156.92485627299175, "wave_p95_s": 157.16690789973364, "wave_samples": 5, "wave_values_s": [ 156.73276200098917, 157.2165887071751, 156.67835334688425, 156.92485627299175, 156.96818466996774 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 40, "shape_ok": true, "status": "passed" }, "case": "allgather-dp8-sp1-t2va-50step-r5", "concurrency": 8, "denoise_updates_executed": 49, "dp": 8, "energy_wh_per_video": 45.08265012415831, "engine_init_s": 357.66571644716896, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 94.03125, "max_temperature_c": 77.0, "mean_gpu_util_pct": 94.89295708955224, "mean_node_power_w": 7982.5736833893125, "mean_sm_clock_mhz": 1492.9177938432836, "mean_wave_s": 163.58930882075802, "measured_peak_memory_gib": 94.03125, "measured_videos": 40, "mode": "allgather", "peak_node_power_w": 8675.76, "per_gpu_lifecycle_peak_memory_gib": { "0": 94.03125, "1": 19.470703125, "2": 20.123046875, "3": 18.26953125, "4": 19.658203125, "5": 19.408203125, "6": 19.701171875, "7": 19.345703125 }, "per_gpu_measured_peak_memory_gib": { "0": 94.03125, "1": 19.470703125, "2": 20.123046875, "3": 18.26953125, "4": 19.658203125, "5": 19.408203125, "6": 19.701171875, "7": 19.345703125 }, "request_p50_s": 162.1479018710088, "request_p95_s": 164.9133004831732, "request_samples": 40, "request_values_s": [ 164.91384473186918, 164.384030388901, 164.91350173717365, 164.9132898908574, 164.91318743187003, 164.91324513102882, 164.91303581511602, 164.91310858400539, 163.15959563897923, 163.1591684781015, 163.15901148109697, 163.159060689155, 163.1588529041037, 163.15888896305114, 163.1590021601878, 163.15889189112931, 161.85883307200857, 161.85860590799712, 161.85850155097432, 161.34619296598248, 161.85817777900957, 161.85821325797588, 161.8582326869946, 161.85807402199134, 162.1484319230076, 162.1481835891027, 162.14796662400477, 162.14785565715283, 161.69018683885224, 162.14794808486477, 162.14784130686894, 162.14772323984653, 161.89923152397387, 161.89888127194718, 161.89891349105164, 161.8987971530296, 161.89859708794393, 161.8987076850608, 161.89856611797586, 161.8983524919022 ], "shutdown_s": 30.100176807958633, "sp": 1, "steps_requested": 50, "sustained_videos_per_hour": 176.05062462581625, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1687, "telemetry_coverage_pct": 99.4269144000194, "telemetry_measured_blocks": 1072, "topology": "dp8-sp1", "warmup_wave_s": 184.2205602650065, "wave_cv_pct": 0.7680935656001869, "wave_max_s": 165.89930333010852, "wave_min_s": 162.48488878598437, "wave_p50_s": 162.8354269778356, "wave_p95_s": 165.50772324628196, "wave_samples": 5, "wave_values_s": [ 165.89930333010852, 163.94140291097574, 162.48488878598437, 162.78552209888585, 162.8354269778356 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 5, "shape_ok": true, "status": "passed" }, "case": "allgather-dp1-sp8-fl2va-50step-r5", "concurrency": 1, "denoise_updates_executed": 49, "dp": 1, "energy_wh_per_video": 73.86722819523075, "engine_init_s": 313.0294175001327, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 34.673828125, "max_temperature_c": 77.0, "mean_gpu_util_pct": 91.28769841269842, "mean_node_power_w": 7015.01562654426, "mean_sm_clock_mhz": 1863.6025132275133, "mean_wave_s": 37.9075237070676, "measured_peak_memory_gib": 23.59375, "measured_videos": 5, "mode": "allgather", "peak_node_power_w": 8548.64, "per_gpu_lifecycle_peak_memory_gib": { "0": 34.673828125, "1": 23.59375, "2": 23.59375, "3": 24.97265625, "4": 24.97265625, "5": 23.59375, "6": 24.865234375, "7": 25.0078125 }, "per_gpu_measured_peak_memory_gib": { "0": 22.291015625, "1": 23.59375, "2": 23.59375, "3": 23.59375, "4": 23.59375, "5": 23.59375, "6": 23.59375, "7": 23.59375 }, "request_p50_s": 37.825976692838594, "request_p95_s": 38.232793235499415, "request_samples": 5, "request_values_s": [ 37.21242911298759, 37.892739449162036, 37.825976692838594, 37.81951795099303, 38.317806682083756 ], "shutdown_s": 20.448273312067613, "sp": 8, "steps_requested": 50, "sustained_videos_per_hour": 94.96795485295192, "task": "fl2va", "telemetry_bad_lines": 0, "telemetry_blocks": 640, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 189, "topology": "dp1-sp8", "warmup_wave_s": 82.358262272086, "wave_cv_pct": 0.93753889559733, "wave_max_s": 38.40932705416344, "wave_min_s": 37.296760766999796, "wave_p50_s": 37.924259808147326, "wave_p95_s": 38.324772434355694, "wave_samples": 5, "wave_values_s": [ 37.296760766999796, 37.98655395512469, 37.924259808147326, 37.920716950902715, 38.40932705416344 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 20, "shape_ok": true, "status": "passed" }, "case": "allgather-dp4-sp2-fl2va-50step-r5", "concurrency": 4, "denoise_updates_executed": 49, "dp": 4, "energy_wh_per_video": 55.98431490148273, "engine_init_s": 309.08109516091645, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 25.15625, "max_temperature_c": 78.0, "mean_gpu_util_pct": 94.0524873524452, "mean_node_power_w": 7832.813728932739, "mean_sm_clock_mhz": 1621.3094435075886, "mean_wave_s": 102.92264100080357, "measured_peak_memory_gib": 25.15625, "measured_videos": 20, "mode": "allgather", "peak_node_power_w": 8684.05, "per_gpu_lifecycle_peak_memory_gib": { "0": 22.3515625, "1": 25.15625, "2": 23.78515625, "3": 22.29296875, "4": 23.708984375, "5": 24.244140625, "6": 23.783203125, "7": 24.99609375 }, "per_gpu_measured_peak_memory_gib": { "0": 22.3515625, "1": 25.15625, "2": 23.78515625, "3": 22.29296875, "4": 23.708984375, "5": 24.244140625, "6": 23.783203125, "7": 24.99609375 }, "request_p50_s": 102.65562396950554, "request_p95_s": 102.86027336129919, "request_samples": 20, "request_values_s": [ 102.85845161089674, 102.86037809611298, 102.86026370897889, 102.86026784894057, 102.52671528491192, 102.52636918099597, 102.33353712689131, 102.52630649297498, 102.3488880789373, 102.14430905412883, 102.3484313681256, 102.34839994902723, 102.65591403888538, 102.65564756398089, 102.65560037503019, 102.65428707702085, 102.83242110093124, 102.83194842096418, 102.83159901201725, 102.83143789088354 ], "shutdown_s": 25.446903855074197, "sp": 2, "steps_requested": 50, "sustained_videos_per_hour": 139.91090648254522, "task": "fl2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1154, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 593, "topology": "dp4-sp2", "warmup_wave_s": 150.79273109184578, "wave_cv_pct": 0.16966649153229205, "wave_max_s": 103.14217631611973, "wave_min_s": 102.69971036794595, "wave_p50_s": 102.90282876207493, "wave_p95_s": 103.13342724968679, "wave_samples": 5, "wave_values_s": [ 103.14217631611973, 102.77005857392214, 102.69971036794595, 102.90282876207493, 103.09843098395504 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 40, "shape_ok": true, "status": "passed" }, "case": "ranklocal-dp8-sp1-fl2va-50step-r5", "concurrency": 8, "denoise_updates_executed": 49, "dp": 8, "energy_wh_per_video": 46.882398891267236, "engine_init_s": 467.04236703785136, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 93.83984375, "max_temperature_c": 76.0, "mean_gpu_util_pct": 92.54583712465877, "mean_node_power_w": 7857.905438366206, "mean_sm_clock_mhz": 1554.9924931756143, "mean_wave_s": 173.57223660545424, "measured_peak_memory_gib": 31.140625, "measured_videos": 40, "mode": "ranklocal", "peak_node_power_w": 8703.85, "per_gpu_lifecycle_peak_memory_gib": { "0": 92.412109375, "1": 93.83984375, "2": 20.09765625, "3": 17.25, "4": 20.08984375, "5": 17.25, "6": 19.07421875, "7": 20.056640625 }, "per_gpu_measured_peak_memory_gib": { "0": 31.140625, "1": 31.140625, "2": 17.25, "3": 17.25, "4": 20.08984375, "5": 17.25, "6": 19.07421875, "7": 17.25 }, "request_p50_s": 169.40395385259762, "request_p95_s": 178.00542293251493, "request_samples": 40, "request_values_s": [ 178.00583215500228, 178.0054188100621, 178.00550125911832, 178.00518619385548, 178.00526000279933, 178.00530409114435, 178.00519044301473, 178.0050191259943, 177.56744940788485, 177.4301577450242, 177.56690162490122, 177.56679046805948, 177.566852445947, 177.56662086001597, 177.5666677588597, 177.42968496284448, 169.15449684602208, 169.15405780193396, 169.15410422091372, 169.15395428310148, 168.663054686971, 169.1538789048791, 169.15385962603614, 169.15388402505778, 168.93832759885117, 168.93794367392547, 168.4613920710981, 168.93765432899818, 168.93735939403996, 168.9373201439157, 168.93706271820702, 168.93693099007942, 169.40468087093905, 169.40433028689586, 169.40428956691176, 169.40400456101634, 169.40388387208804, 169.4039031441789, 169.4038754128851, 169.4038769339677 ], "shutdown_s": 30.027144169900566, "sp": 1, "steps_requested": 50, "sustained_videos_per_hour": 165.92515348791102, "task": "fl2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1813, "telemetry_coverage_pct": 98.99543290631041, "telemetry_measured_blocks": 1099, "topology": "dp8-sp1", "warmup_wave_s": 185.89386785891838, "wave_cv_pct": 2.4783279836531307, "wave_max_s": 178.94095333013684, "wave_min_s": 169.88679977110587, "wave_p50_s": 170.21746728708968, "wave_p95_s": 178.8999878913164, "wave_samples": 5, "wave_values_s": [ 178.94095333013684, 178.73612613603473, 170.21746728708968, 169.88679977110587, 170.07983650290407 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 5, "shape_ok": true, "status": "passed" }, "case": "allgather-dp1-sp8-ref2va-50step-r5", "concurrency": 1, "denoise_updates_executed": 49, "dp": 1, "energy_wh_per_video": 103.66866568038772, "engine_init_s": 333.59189733304083, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 102.18359375, "max_temperature_c": 77.0, "mean_gpu_util_pct": 92.02881205673759, "mean_node_power_w": 6701.995211280079, "mean_sm_clock_mhz": 1754.1183510638298, "mean_wave_s": 55.685962193552406, "measured_peak_memory_gib": 97.20703125, "measured_videos": 5, "mode": "allgather", "peak_node_power_w": 8708.87, "per_gpu_lifecycle_peak_memory_gib": { "0": 102.18359375, "1": 102.080078125, "2": 26.697265625, "3": 26.634765625, "4": 23.8359375, "5": 25.267578125, "6": 26.60546875, "7": 26.60546875 }, "per_gpu_measured_peak_memory_gib": { "0": 97.20703125, "1": 93.009765625, "2": 26.697265625, "3": 26.634765625, "4": 23.8359375, "5": 25.267578125, "6": 26.60546875, "7": 26.60546875 }, "request_p50_s": 55.308971765916795, "request_p95_s": 56.18421814506873, "request_samples": 5, "request_values_s": [ 55.90615408914164, 56.25373415905051, 55.06021963711828, 55.29280473408289, 55.308971765916795 ], "shutdown_s": 21.485767379868776, "sp": 8, "steps_requested": 50, "sustained_videos_per_hour": 64.64824990339892, "task": "ref2va", "telemetry_bad_lines": 0, "telemetry_blocks": 737, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 282, "topology": "dp1-sp8", "warmup_wave_s": 96.65881789801642, "wave_cv_pct": 0.7630638539759705, "wave_max_s": 56.36234005889855, "wave_min_s": 55.214124653022736, "wave_p50_s": 55.44087025686167, "wave_p95_s": 56.288000704720616, "wave_samples": 5, "wave_values_s": [ 55.99064328800887, 56.36234005889855, 55.214124653022736, 55.44087025686167, 55.42183271097019 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 20, "shape_ok": true, "status": "passed" }, "case": "allgather-dp4-sp2-ref2va-50step-r5", "concurrency": 4, "denoise_updates_executed": 49, "dp": 4, "energy_wh_per_video": 85.52252901386638, "engine_init_s": 321.64542035292834, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 98.294921875, "max_temperature_c": 80.0, "mean_gpu_util_pct": 92.19732620320856, "mean_node_power_w": 7314.312172206169, "mean_sm_clock_mhz": 1621.2655080213904, "mean_wave_s": 168.89986491180025, "measured_peak_memory_gib": 98.294921875, "measured_videos": 20, "mode": "allgather", "peak_node_power_w": 8761.34, "per_gpu_lifecycle_peak_memory_gib": { "0": 98.294921875, "1": 22.517578125, "2": 23.796875, "3": 23.94921875, "4": 23.69140625, "5": 25.287109375, "6": 21.216796875, "7": 25.318359375 }, "per_gpu_measured_peak_memory_gib": { "0": 98.294921875, "1": 22.517578125, "2": 23.796875, "3": 23.94921875, "4": 23.69140625, "5": 25.287109375, "6": 21.20703125, "7": 25.318359375 }, "request_p50_s": 166.6427594795823, "request_p95_s": 176.21028091231128, "request_samples": 20, "request_values_s": [ 166.68245272198692, 166.6819431600161, 166.46492357202806, 166.6817997021135, 166.35471620899625, 166.35427712695673, 166.35415015812032, 166.12434804695658, 166.4840814229101, 166.48367033083923, 166.40860714996234, 166.40858205105178, 176.21064554410987, 176.210261721164, 176.20871036802419, 176.20977124897763, 166.8371843120549, 166.8367148800753, 166.60371925705113, 166.8366431510076 ], "shutdown_s": 30.09092747210525, "sp": 2, "steps_requested": 50, "sustained_videos_per_hour": 85.25761703551214, "task": "ref2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1482, "telemetry_coverage_pct": 99.68737398718007, "telemetry_measured_blocks": 935, "topology": "dp4-sp2", "warmup_wave_s": 184.97215842991136, "wave_cv_pct": 2.306119678806293, "wave_max_s": 176.68391854013316, "wave_min_s": 166.68619965412654, "wave_p50_s": 167.01304856687784, "wave_p95_s": 174.778839964699, "wave_samples": 5, "wave_values_s": [ 167.01304856687784, 166.68619965412654, 166.95763213490136, 176.68391854013316, 167.15852566296235 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 40, "shape_ok": true, "status": "passed" }, "case": "ranklocal-dp8-sp1-ref2va-50step-r5", "concurrency": 8, "denoise_updates_executed": 49, "dp": 8, "energy_wh_per_video": 74.06088714952557, "engine_init_s": 451.6622830799315, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 21.0, "max_temperature_c": 81.0, "mean_gpu_util_pct": 86.5777282549973, "mean_node_power_w": 7410.90669444745, "mean_sm_clock_mhz": 1543.9590761750405, "mean_wave_s": 287.8127244349569, "measured_peak_memory_gib": 19.359375, "measured_videos": 40, "mode": "ranklocal", "peak_node_power_w": 8803.06, "per_gpu_lifecycle_peak_memory_gib": { "0": 21.0, "1": 17.984375, "2": 17.984375, "3": 17.984375, "4": 18.849609375, "5": 17.984375, "6": 19.359375, "7": 17.984375 }, "per_gpu_measured_peak_memory_gib": { "0": 17.984375, "1": 17.984375, "2": 17.984375, "3": 17.984375, "4": 18.849609375, "5": 17.984375, "6": 19.359375, "7": 17.984375 }, "request_p50_s": 287.01811501255725, "request_p95_s": 288.7633355701808, "request_samples": 40, "request_values_s": [ 288.76422577816993, 288.235728032887, 288.7636806508526, 288.76331740804017, 288.76327256904915, 288.76308715203777, 288.7628085180186, 288.7628573179245, 288.62950287899, 288.6290224590339, 288.62902920902707, 288.62892909999937, 288.6289811388124, 288.62889484106563, 288.35971563798375, 288.35970075801015, 287.0192966419272, 287.0189380599186, 287.0184978879988, 287.017018977087, 287.0182366850786, 287.0179933400359, 287.0179214309901, 287.01757304696366, 285.601361281937, 286.0970417561475, 286.0968658400234, 286.09687037882395, 286.09662870387547, 286.0967230910901, 286.0965663539246, 286.0965491461102, 284.48506064503454, 284.9400555999018, 284.93986854399554, 284.93994152196683, 284.9399063629098, 284.9397151370067, 284.93968722689897, 284.939661267912 ], "shutdown_s": 24.380246096057817, "sp": 1, "steps_requested": 50, "sustained_videos_per_hour": 100.06506854949195, "task": "ref2va", "telemetry_bad_lines": 0, "telemetry_blocks": 2719, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 1851, "topology": "dp8-sp1", "warmup_wave_s": 294.71439594309777, "wave_cv_pct": 0.5124197523346558, "wave_max_s": 289.68081891187467, "wave_min_s": 285.8072912108619, "wave_p50_s": 287.6175738100428, "wave_p95_s": 289.5961993737146, "wave_samples": 5, "wave_values_s": [ 289.68081891187467, 289.25772122107446, 287.6175738100428, 286.7002170209307, 285.8072912108619 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 15, "shape_ok": true, "status": "passed" }, "case": "soak-allgather-dp1-sp8-t2va-50step-r15", "concurrency": 1, "denoise_updates_executed": 49, "dp": 1, "energy_wh_per_video": 68.19479741395993, "engine_init_s": 322.72203186899424, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 26.365234375, "max_temperature_c": 74.0, "mean_gpu_util_pct": 90.28466981132075, "mean_node_power_w": 7063.123874381645, "mean_sm_clock_mhz": 1828.3627358490567, "mean_wave_s": 34.75814999191401, "measured_peak_memory_gib": 26.365234375, "measured_videos": 15, "mode": "allgather", "peak_node_power_w": 8514.55, "per_gpu_lifecycle_peak_memory_gib": { "0": 23.98046875, "1": 23.529296875, "2": 23.984375, "3": 23.529296875, "4": 23.859375, "5": 24.046875, "6": 26.361328125, "7": 26.365234375 }, "per_gpu_measured_peak_memory_gib": { "0": 23.98046875, "1": 23.529296875, "2": 23.984375, "3": 23.529296875, "4": 23.859375, "5": 24.046875, "6": 26.361328125, "7": 26.365234375 }, "request_p50_s": 34.52474875305779, "request_p95_s": 35.38997254313435, "request_samples": 15, "request_values_s": [ 36.0866008840967, 34.10303023084998, 34.967818598030135, 34.41919455095194, 34.593092336086556, 34.35336302802898, 34.52474875305779, 34.76744698896073, 34.4302857730072, 34.13301039696671, 34.35436272015795, 35.09141753986478, 34.064232495147735, 34.56138198100962, 35.07870242907666 ], "shutdown_s": 20.541397747118026, "sp": 8, "steps_requested": 50, "sustained_videos_per_hour": 103.57283114427807, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 1012, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 530, "topology": "dp1-sp8", "warmup_wave_s": 80.27618861896917, "wave_cv_pct": 1.4447473433661118, "wave_max_s": 36.21506123198196, "wave_min_s": 34.174797968007624, "wave_p50_s": 34.6710405908525, "wave_p95_s": 35.505773949809374, "wave_samples": 15, "wave_values_s": [ 36.21506123198196, 34.241454757982865, 35.114581926958635, 34.560933315893635, 34.69453239394352, 34.46163991512731, 34.6710405908525, 34.8828783580102, 34.54589975904673, 34.241072338074446, 34.501001303084195, 35.20179368602112, 34.174797968007624, 34.67267275089398, 35.19288958283141 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 0, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": false, "worker_died": false }, "fatal_error": false, "measured_output_count": 60, "shape_ok": true, "status": "passed" }, "case": "soak-allgather-dp4-sp2-t2va-50step-r15", "concurrency": 4, "denoise_updates_executed": 49, "dp": 4, "energy_wh_per_video": 51.78563707682368, "engine_init_s": 310.22449679416604, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 25.107421875, "max_temperature_c": 78.0, "mean_gpu_util_pct": 93.69374621441551, "mean_node_power_w": 7866.509107653309, "mean_sm_clock_mhz": 1619.0642035130224, "mean_wave_s": 94.93557581445202, "measured_peak_memory_gib": 25.107421875, "measured_videos": 60, "mode": "allgather", "peak_node_power_w": 8719.47, "per_gpu_lifecycle_peak_memory_gib": { "0": 23.369140625, "1": 25.095703125, "2": 23.75390625, "3": 25.1015625, "4": 23.75390625, "5": 25.107421875, "6": 23.7265625, "7": 25.08203125 }, "per_gpu_measured_peak_memory_gib": { "0": 23.369140625, "1": 25.095703125, "2": 23.75390625, "3": 25.1015625, "4": 23.75390625, "5": 25.107421875, "6": 23.7265625, "7": 25.08203125 }, "request_p50_s": 94.416650026571, "request_p95_s": 94.92976345052011, "request_samples": 60, "request_values_s": [ 94.47121634706855, 94.66622212808579, 94.66611363901757, 94.66599686001427, 94.85103903105482, 94.85062345000915, 94.8502634849865, 94.63475876813754, 94.92419367516413, 94.92395213898271, 94.72288371506147, 94.92379026301205, 94.93037632503547, 94.92998701287434, 94.92977987602353, 94.92976258601993, 94.3964108589571, 94.39619075111113, 94.3959930029232, 94.39601668296382, 94.35429618996568, 94.35393299395218, 94.35390495508909, 94.14929260499775, 94.35099808499217, 94.35061551816761, 94.35046598990448, 94.34920608694665, 94.43901469791308, 94.43867649207823, 94.43863055412658, 94.4385010660626, 94.18403327884153, 94.18375919293612, 94.18371233390644, 94.18352163606323, 94.33908008597791, 94.54142624605447, 94.54127552895807, 94.54125517886132, 94.27269358700141, 94.2723609409295, 94.27231317199767, 94.27217112388462, 94.59221268096007, 94.5918444271665, 94.59153180080466, 94.39337701397017, 94.23428387497552, 94.03756350209005, 94.23394960979931, 94.23391300090589, 94.43688919418491, 94.64471982698888, 94.6445708789397, 94.64454760914668, 94.29002399696037, 94.2896244241856, 94.28939079819247, 94.07687358511612 ], "shutdown_s": 26.1703458330594, "sp": 2, "steps_requested": 50, "sustained_videos_per_hour": 151.68181028515858, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 2144, "telemetry_coverage_pct": 99.85289150415059, "telemetry_measured_blocks": 1651, "topology": "dp4-sp2", "warmup_wave_s": 108.46901519596577, "wave_cv_pct": 0.25045667199344535, "wave_max_s": 95.39917947002687, "wave_min_s": 94.65760229318403, "wave_p50_s": 94.90892298403196, "wave_p95_s": 95.33061604567338, "wave_samples": 15, "wave_values_s": [ 95.1047255678568, 95.30123172095045, 95.39917947002687, 95.26087002479471, 94.71903971489519, 94.82264852989465, 94.65760229318403, 94.90892298403196, 94.6725157070905, 95.02406334900297, 94.7351494859904, 95.04445270099677, 94.70667615602724, 94.95440898905508, 94.7221505229827 ], "width": 1344 }, { "audit": { "cleanup_status": "clean", "error_line_count": 1, "error_types": { "group_rank": false, "oom": false, "operation_creation": false, "segfault": false, "shutdown_timeout": true, "worker_died": false }, "fatal_error": false, "measured_output_count": 120, "shape_ok": true, "status": "passed" }, "case": "soak-ranklocal-dp8-sp1-t2va-50step-r15", "concurrency": 8, "denoise_updates_executed": 49, "dp": 8, "energy_wh_per_video": 43.96760399756947, "engine_init_s": 420.9326464571059, "frames": 124, "height": 768, "lifecycle_peak_memory_gib": 20.0546875, "max_temperature_c": 78.0, "mean_gpu_util_pct": 94.69609500805153, "mean_node_power_w": 8083.728230674539, "mean_sm_clock_mhz": 1519.0721819645732, "mean_wave_s": 156.6438957539387, "measured_peak_memory_gib": 20.0546875, "measured_videos": 120, "mode": "ranklocal", "peak_node_power_w": 8749.8, "per_gpu_lifecycle_peak_memory_gib": { "0": 18.828125, "1": 18.982421875, "2": 20.0, "3": 19.765625, "4": 20.0546875, "5": 19.935546875, "6": 19.796875, "7": 17.19140625 }, "per_gpu_measured_peak_memory_gib": { "0": 18.828125, "1": 18.982421875, "2": 20.0, "3": 19.765625, "4": 20.0546875, "5": 19.935546875, "6": 19.796875, "7": 17.19140625 }, "request_p50_s": 155.9272729625227, "request_p95_s": 156.2243767111795, "request_samples": 120, "request_values_s": [ 156.14487677114084, 156.14460928505287, 156.14467092510313, 155.68986208294518, 156.14449365786277, 156.14446876686998, 156.14434288116172, 156.14415347389877, 155.99486353690736, 155.99475062894635, 155.9946401210036, 155.9946259309072, 155.99446358392015, 155.9942625781987, 155.51326021691784, 155.99422487802804, 156.04613001295365, 156.04586411896162, 156.0458164010197, 156.04581121099181, 155.55696337507106, 156.04566201288253, 156.04547558678314, 156.0454567570705, 155.88891655998304, 155.8886938248761, 155.88875959487632, 155.88862597499974, 155.88852124894038, 155.88839265191928, 155.43807193008251, 155.88844056101516, 155.96787663619034, 155.96768250106834, 155.83641192200594, 155.9673461469356, 155.9674205959309, 155.96740144491196, 155.9672238889616, 155.83728947513737, 155.75283777993172, 156.2249404070899, 156.2248029599432, 156.2246645039413, 156.22477230080403, 156.22466129413806, 156.2245281659998, 156.22436873987317, 156.13527295482345, 156.13124554906972, 156.13495098007843, 156.13482284289785, 156.1348655319307, 156.13483811286278, 156.1346890348941, 156.1345515579451, 155.98085280298255, 155.46119379089214, 155.98034409293905, 155.9800725770183, 155.9797734420281, 155.97966832504608, 155.97964327596128, 155.97952164895833, 155.6094499239698, 156.10996113182046, 156.1098128540907, 156.10971948807128, 156.10958507889882, 156.1094855510164, 156.10963307786733, 156.10947553999722, 155.7665666458197, 155.7662918008864, 155.76621977286413, 155.28089245292358, 155.76613510516472, 155.76593491900712, 155.7657902110368, 155.76586777996272, 155.81967494799756, 155.41064266185276, 155.8191832168959, 155.8190922990907, 155.81780481315218, 155.41156434593722, 155.81888041296043, 155.81879498600028, 155.69791654706933, 155.69738041702658, 155.69736773706973, 155.69727411912754, 155.69727988890372, 155.69727715896443, 155.19995941780508, 155.69695729506202, 155.77147456491366, 155.77101823408157, 155.77099137404002, 155.77102703391574, 155.7709470440168, 155.7708428769838, 155.77081237686798, 155.27326258085668, 155.9037292469293, 155.9033278040588, 155.40905180596747, 155.90311726694927, 155.90291799092665, 155.9029618599452, 155.90287275216542, 155.90292202099226, 155.46773529495113, 155.92744814488105, 155.9275860011112, 155.92731187702157, 155.92731220601127, 155.92715210001916, 155.92723404802382, 155.9271143109072 ], "shutdown_s": 30.10126191098243, "sp": 1, "steps_requested": 50, "sustained_videos_per_hour": 183.85651008858954, "task": "t2va", "telemetry_bad_lines": 0, "telemetry_blocks": 3787, "telemetry_coverage_pct": 100.0, "telemetry_measured_blocks": 3105, "topology": "dp8-sp1", "warmup_wave_s": 161.5930544710718, "wave_cv_pct": 0.1388228769995783, "wave_max_s": 157.01837082812563, "wave_min_s": 156.29264628794044, "wave_p50_s": 156.61332127009518, "wave_p95_s": 156.97546660162044, "wave_samples": 15, "wave_values_s": [ 156.7593610167969, 156.61332127009518, 156.95707907597534, 156.51728480495512, 156.54197374801151, 156.85744950803928, 156.8194091389887, 156.5648122720886, 157.01837082812563, 156.34758708183654, 156.44614280201495, 156.29264628794044, 156.38694050116464, 156.75180666986853, 156.7842513031792 ], "width": 1344 } ], "generated_at_utc": "2026-08-09T01:22:36.218445+00:00", "methodology": { "energy": "trapezoidal integration of summed 8-GPU board power; no idle subtraction", "memory": "external nvidia-smi per-GPU maximum; lifecycle and measured windows reported separately", "percentile": "linear interpolation over sorted samples", "telemetry_effective_interval_s": { "median": 0.7580001354217529, "p95": 1.1340000629425049 }, "telemetry_grouping": "consecutive GPU 0..7 rows from one nvidia-smi query", "telemetry_max_interpolation_gap_s": 2.0 }, "schema_version": 1 }