v8 bench result push
This commit is contained in:
@@ -1,141 +0,0 @@
|
||||
{
|
||||
"benchmark_version": "1",
|
||||
"generated_at": "2026-04-14T15:28:24.793730899Z",
|
||||
"hostname": "debian",
|
||||
"server_model": "MLT-S06",
|
||||
"benchmark_profile": "standard",
|
||||
"parallel_gpus": true,
|
||||
"ramp_step": 1,
|
||||
"ramp_total": 2,
|
||||
"ramp_run_id": "ramp-20260414-152824",
|
||||
"overall_status": "FAILED",
|
||||
"selected_gpu_indices": [
|
||||
0
|
||||
],
|
||||
"findings": [
|
||||
"0 of 1 GPU(s) passed the benchmark.",
|
||||
"GPU 0 average SM clock stayed below the requested lock target."
|
||||
],
|
||||
"normalization": {
|
||||
"status": "full",
|
||||
"gpus": [
|
||||
{
|
||||
"index": 0,
|
||||
"persistence_mode": "applied",
|
||||
"gpu_clock_lock_mhz": 1980,
|
||||
"gpu_clock_lock_status": "applied",
|
||||
"memory_clock_lock_mhz": 2619,
|
||||
"memory_clock_lock_status": "applied"
|
||||
}
|
||||
]
|
||||
},
|
||||
"host_config": {
|
||||
"cpu_model": "Intel(R) Xeon(R) Gold 6430",
|
||||
"cpu_sockets": 2,
|
||||
"cpu_cores": 64,
|
||||
"cpu_threads": 128,
|
||||
"mem_total_gib": 62.53376007080078
|
||||
},
|
||||
"gpus": [
|
||||
{
|
||||
"index": 0,
|
||||
"uuid": "GPU-9ee0af22-3dda-6f5c-1a13-35c63f324216",
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:49:00.0",
|
||||
"vbios": "96.00.74.00.01",
|
||||
"status": "UNSUPPORTED",
|
||||
"power_limit_w": 700,
|
||||
"default_power_limit_w": 700,
|
||||
"max_graphics_clock_mhz": 1980,
|
||||
"max_memory_clock_mhz": 2619,
|
||||
"locked_graphics_clock_mhz": 1980,
|
||||
"locked_memory_clock_mhz": 2619,
|
||||
"baseline": {
|
||||
"duration_sec": 0,
|
||||
"samples": 0,
|
||||
"avg_temp_c": 0,
|
||||
"p95_temp_c": 0,
|
||||
"avg_power_w": 0,
|
||||
"p95_power_w": 0,
|
||||
"avg_graphics_clock_mhz": 0,
|
||||
"p95_graphics_clock_mhz": 0,
|
||||
"avg_memory_clock_mhz": 0,
|
||||
"p95_memory_clock_mhz": 0,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"steady": {
|
||||
"duration_sec": 0,
|
||||
"samples": 0,
|
||||
"avg_temp_c": 0,
|
||||
"p95_temp_c": 0,
|
||||
"avg_power_w": 0,
|
||||
"p95_power_w": 0,
|
||||
"avg_graphics_clock_mhz": 0,
|
||||
"p95_graphics_clock_mhz": 0,
|
||||
"avg_memory_clock_mhz": 0,
|
||||
"p95_memory_clock_mhz": 0,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"cooldown": {
|
||||
"duration_sec": 0,
|
||||
"samples": 0,
|
||||
"avg_temp_c": 0,
|
||||
"p95_temp_c": 0,
|
||||
"avg_power_w": 0,
|
||||
"p95_power_w": 0,
|
||||
"avg_graphics_clock_mhz": 0,
|
||||
"p95_graphics_clock_mhz": 0,
|
||||
"avg_memory_clock_mhz": 0,
|
||||
"p95_memory_clock_mhz": 0,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"throttle_counters": {
|
||||
"sw_power_cap_us": 0,
|
||||
"sw_thermal_slowdown_us": 0,
|
||||
"sync_boost_us": 0,
|
||||
"hw_thermal_slowdown_us": 0,
|
||||
"hw_power_brake_slowdown_us": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
},
|
||||
"scores": {
|
||||
"compute_score": 0,
|
||||
"power_sustain_score": 0,
|
||||
"thermal_sustain_score": 100,
|
||||
"stability_score": 100,
|
||||
"interconnect_score": 0,
|
||||
"composite_score": 0
|
||||
},
|
||||
"degradation_reasons": [
|
||||
"low_sm_clock_vs_target"
|
||||
],
|
||||
"notes": [
|
||||
"parallel warmup failed: context canceled",
|
||||
"precision plan failed: context canceled"
|
||||
]
|
||||
}
|
||||
],
|
||||
"server_power": {
|
||||
"available": false,
|
||||
"notes": [
|
||||
"IPMI power reading unavailable; server-side power characterization skipped"
|
||||
]
|
||||
}
|
||||
}
|
||||
@@ -1,385 +0,0 @@
|
||||
{
|
||||
"benchmark_version": "1",
|
||||
"generated_at": "2026-04-14T15:29:49.813251958Z",
|
||||
"hostname": "debian",
|
||||
"server_model": "MLT-S06",
|
||||
"benchmark_profile": "standard",
|
||||
"parallel_gpus": true,
|
||||
"ramp_step": 1,
|
||||
"ramp_total": 2,
|
||||
"ramp_run_id": "ramp-20260414-152949",
|
||||
"overall_status": "OK",
|
||||
"selected_gpu_indices": [
|
||||
0
|
||||
],
|
||||
"findings": [
|
||||
"All 1 GPU(s) passed the benchmark.",
|
||||
"GPU 0 average SM clock stayed below the requested lock target."
|
||||
],
|
||||
"normalization": {
|
||||
"status": "full",
|
||||
"gpus": [
|
||||
{
|
||||
"index": 0,
|
||||
"persistence_mode": "applied",
|
||||
"gpu_clock_lock_mhz": 1980,
|
||||
"gpu_clock_lock_status": "applied",
|
||||
"memory_clock_lock_mhz": 2619,
|
||||
"memory_clock_lock_status": "applied"
|
||||
}
|
||||
]
|
||||
},
|
||||
"host_config": {
|
||||
"cpu_model": "Intel(R) Xeon(R) Gold 6430",
|
||||
"cpu_sockets": 2,
|
||||
"cpu_cores": 64,
|
||||
"cpu_threads": 128,
|
||||
"mem_total_gib": 62.53376007080078
|
||||
},
|
||||
"cpu_load": {
|
||||
"avg_pct": 0.7,
|
||||
"max_pct": 1.1,
|
||||
"p95_pct": 1,
|
||||
"samples": 78,
|
||||
"status": "ok"
|
||||
},
|
||||
"cooling": {
|
||||
"available": true,
|
||||
"avg_fan_rpm": 6083.203125,
|
||||
"notes": [
|
||||
"fan duty cycle unavailable on this host; RPM-only fan telemetry was collected"
|
||||
]
|
||||
},
|
||||
"gpus": [
|
||||
{
|
||||
"index": 0,
|
||||
"uuid": "GPU-9ee0af22-3dda-6f5c-1a13-35c63f324216",
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:49:00.0",
|
||||
"vbios": "96.00.74.00.01",
|
||||
"compute_capability": "9.0",
|
||||
"backend": "cublasLt",
|
||||
"status": "OK",
|
||||
"power_limit_w": 700,
|
||||
"default_power_limit_w": 700,
|
||||
"calibrated_peak_power_w": 698.316,
|
||||
"max_graphics_clock_mhz": 1980,
|
||||
"max_memory_clock_mhz": 2619,
|
||||
"locked_graphics_clock_mhz": 1980,
|
||||
"locked_memory_clock_mhz": 2619,
|
||||
"baseline": {
|
||||
"duration_sec": 22.679193237,
|
||||
"samples": 1,
|
||||
"avg_temp_c": 66,
|
||||
"p95_temp_c": 66,
|
||||
"avg_power_w": 244.62,
|
||||
"p95_power_w": 244.62,
|
||||
"avg_graphics_clock_mhz": 1830,
|
||||
"p95_graphics_clock_mhz": 1830,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"steady": {
|
||||
"duration_sec": 664.658712143,
|
||||
"samples": 13,
|
||||
"avg_temp_c": 84.23076923076923,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 531.6661538461539,
|
||||
"p95_power_w": 612.742,
|
||||
"avg_graphics_clock_mhz": 1221.923076923077,
|
||||
"p95_graphics_clock_mhz": 1830,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 76.92307692307692,
|
||||
"avg_mem_usage_pct": 38.07692307692308,
|
||||
"clock_cv_pct": 27.28679189404879,
|
||||
"power_cv_pct": 25.428186879992428,
|
||||
"temp_cv_pct": 5.571524764445532,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"precision_steady": [
|
||||
{
|
||||
"precision": "int8",
|
||||
"steady": {
|
||||
"duration_sec": 50.799623108,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 73.5,
|
||||
"p95_temp_c": 76.65,
|
||||
"avg_power_w": 641.45,
|
||||
"p95_power_w": 692.912,
|
||||
"avg_graphics_clock_mhz": 1245,
|
||||
"p95_graphics_clock_mhz": 1299,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 80.5,
|
||||
"clock_cv_pct": 4.819277108433735,
|
||||
"power_cv_pct": 8.91417881362538,
|
||||
"temp_cv_pct": 4.761904761904762,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 1104.4978767691775,
|
||||
"weighted_teraops_per_sec": 276.1244691922944,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp8",
|
||||
"steady": {
|
||||
"duration_sec": 119.043156359,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 84.33333333333333,
|
||||
"p95_temp_c": 86.8,
|
||||
"avg_power_w": 681.4366666666666,
|
||||
"p95_power_w": 699.73,
|
||||
"avg_graphics_clock_mhz": 1145,
|
||||
"p95_graphics_clock_mhz": 1170,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 52,
|
||||
"clock_cv_pct": 3.087802537932522,
|
||||
"power_cv_pct": 3.7139603513146184,
|
||||
"temp_cv_pct": 2.957831926303511,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 1096.7958204099243,
|
||||
"weighted_teraops_per_sec": 274.1989551024811,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp16",
|
||||
"steady": {
|
||||
"duration_sec": 179.687679749,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 612.785,
|
||||
"p95_power_w": 613.1315,
|
||||
"avg_graphics_clock_mhz": 1087.5,
|
||||
"p95_graphics_clock_mhz": 1107.75,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 40.5,
|
||||
"clock_cv_pct": 2.0689655172413794,
|
||||
"power_cv_pct": 0.06282790864658745,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 702.1073356117333,
|
||||
"weighted_teraops_per_sec": 351.05366780586667,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp32",
|
||||
"steady": {
|
||||
"duration_sec": 232.355201725,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 603.5899999999999,
|
||||
"p95_power_w": 609.4129999999999,
|
||||
"avg_graphics_clock_mhz": 1102.5,
|
||||
"p95_graphics_clock_mhz": 1109.25,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 40,
|
||||
"clock_cv_pct": 0.6802721088435374,
|
||||
"power_cv_pct": 1.071919680577871,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 297.2914085265408,
|
||||
"weighted_teraops_per_sec": 297.2914085265408,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp64",
|
||||
"steady": {
|
||||
"duration_sec": 282.807869037,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 617.8499999999999,
|
||||
"p95_power_w": 620.3879999999999,
|
||||
"avg_graphics_clock_mhz": 1117.5,
|
||||
"p95_graphics_clock_mhz": 1124.25,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 47,
|
||||
"clock_cv_pct": 0.6711409395973155,
|
||||
"power_cv_pct": 0.4564214615197853,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp4",
|
||||
"steady": {
|
||||
"duration_sec": 355.407457812,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 86.33333333333333,
|
||||
"p95_temp_c": 86.9,
|
||||
"avg_power_w": 598.8033333333333,
|
||||
"p95_power_w": 604.278,
|
||||
"avg_graphics_clock_mhz": 1020,
|
||||
"p95_graphics_clock_mhz": 1033.5,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 47,
|
||||
"clock_cv_pct": 1.2007302660701853,
|
||||
"power_cv_pct": 0.7587124023422941,
|
||||
"temp_cv_pct": 0.5460284024606545,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
}
|
||||
],
|
||||
"cooldown": {
|
||||
"duration_sec": 0,
|
||||
"samples": 0,
|
||||
"avg_temp_c": 0,
|
||||
"p95_temp_c": 0,
|
||||
"avg_power_w": 0,
|
||||
"p95_power_w": 0,
|
||||
"avg_graphics_clock_mhz": 0,
|
||||
"p95_graphics_clock_mhz": 0,
|
||||
"avg_memory_clock_mhz": 0,
|
||||
"p95_memory_clock_mhz": 0,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"throttle_counters": {
|
||||
"sw_power_cap_us": 0,
|
||||
"sw_thermal_slowdown_us": 0,
|
||||
"sync_boost_us": 0,
|
||||
"hw_thermal_slowdown_us": 0,
|
||||
"hw_power_brake_slowdown_us": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
},
|
||||
"precision_results": [
|
||||
{
|
||||
"name": "fp16_tensor",
|
||||
"category": "fp16_bf16",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 9856,
|
||||
"n": 9856,
|
||||
"k": 9856,
|
||||
"iterations": 28032,
|
||||
"teraops_per_sec": 178.92248028971008,
|
||||
"weight": 0.5,
|
||||
"weighted_teraops_per_sec": 89.46124014485504
|
||||
},
|
||||
{
|
||||
"name": "fp32_tf32",
|
||||
"category": "fp32_tf32",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 6912,
|
||||
"n": 6912,
|
||||
"k": 6912,
|
||||
"iterations": 28032,
|
||||
"teraops_per_sec": 61.71262413963264,
|
||||
"weight": 1,
|
||||
"weighted_teraops_per_sec": 61.71262413963264
|
||||
},
|
||||
{
|
||||
"name": "fp64",
|
||||
"category": "fp64",
|
||||
"supported": false,
|
||||
"weight": 2,
|
||||
"notes": "unsupported"
|
||||
},
|
||||
{
|
||||
"name": "fp8_e4m3",
|
||||
"category": "fp8",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 11392,
|
||||
"n": 11392,
|
||||
"k": 11392,
|
||||
"iterations": 28032,
|
||||
"teraops_per_sec": 276.28846547206143,
|
||||
"weight": 0.25,
|
||||
"weighted_teraops_per_sec": 69.07211636801536
|
||||
},
|
||||
{
|
||||
"name": "fp8_e5m2",
|
||||
"category": "fp8",
|
||||
"supported": false,
|
||||
"weight": 0.25,
|
||||
"notes": "unsupported"
|
||||
},
|
||||
{
|
||||
"name": "int8_tensor",
|
||||
"category": "int8",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 8832,
|
||||
"n": 8832,
|
||||
"k": 8832,
|
||||
"iterations": 28032,
|
||||
"teraops_per_sec": 128.74785629405184,
|
||||
"weight": 0.25,
|
||||
"weighted_teraops_per_sec": 32.18696407351296
|
||||
}
|
||||
],
|
||||
"scores": {
|
||||
"compute_score": 1274.3983840449878,
|
||||
"synthetic_score": 1198.668500627183,
|
||||
"mixed_score": 252.432944726016,
|
||||
"mixed_efficiency": 0.21059445926370363,
|
||||
"power_sustain_score": 99.75942857142857,
|
||||
"thermal_sustain_score": 100,
|
||||
"stability_score": 86.6552008085787,
|
||||
"interconnect_score": 0,
|
||||
"composite_score": 1231.422031972879
|
||||
},
|
||||
"degradation_reasons": [
|
||||
"low_sm_clock_vs_target"
|
||||
]
|
||||
}
|
||||
],
|
||||
"server_power": {
|
||||
"available": false,
|
||||
"notes": [
|
||||
"IPMI power reading unavailable; server-side power characterization skipped"
|
||||
]
|
||||
}
|
||||
}
|
||||
@@ -1,735 +0,0 @@
|
||||
{
|
||||
"benchmark_version": "1",
|
||||
"generated_at": "2026-04-14T15:46:57.112694192Z",
|
||||
"hostname": "debian",
|
||||
"server_model": "MLT-S06",
|
||||
"benchmark_profile": "standard",
|
||||
"parallel_gpus": true,
|
||||
"ramp_step": 2,
|
||||
"ramp_total": 2,
|
||||
"ramp_run_id": "ramp-20260414-152949",
|
||||
"overall_status": "OK",
|
||||
"selected_gpu_indices": [
|
||||
0,
|
||||
1
|
||||
],
|
||||
"findings": [
|
||||
"All 2 GPU(s) passed the benchmark.",
|
||||
"GPU 0 average SM clock stayed below the requested lock target.",
|
||||
"GPU 0 showed unstable clocks/power over the benchmark window.",
|
||||
"GPU 1 average SM clock stayed below the requested lock target.",
|
||||
"GPU 1 showed unstable clocks/power over the benchmark window.",
|
||||
"Multi-GPU all_reduce max bus bandwidth: 12.1 GB/s."
|
||||
],
|
||||
"normalization": {
|
||||
"status": "full",
|
||||
"gpus": [
|
||||
{
|
||||
"index": 0,
|
||||
"persistence_mode": "applied",
|
||||
"gpu_clock_lock_mhz": 1980,
|
||||
"gpu_clock_lock_status": "applied",
|
||||
"memory_clock_lock_mhz": 2619,
|
||||
"memory_clock_lock_status": "applied"
|
||||
},
|
||||
{
|
||||
"index": 1,
|
||||
"persistence_mode": "applied",
|
||||
"gpu_clock_lock_mhz": 1980,
|
||||
"gpu_clock_lock_status": "applied",
|
||||
"memory_clock_lock_mhz": 2619,
|
||||
"memory_clock_lock_status": "applied"
|
||||
}
|
||||
]
|
||||
},
|
||||
"host_config": {
|
||||
"cpu_model": "Intel(R) Xeon(R) Gold 6430",
|
||||
"cpu_sockets": 2,
|
||||
"cpu_cores": 64,
|
||||
"cpu_threads": 128,
|
||||
"mem_total_gib": 62.53376007080078
|
||||
},
|
||||
"cpu_load": {
|
||||
"avg_pct": 1.5,
|
||||
"max_pct": 1.9,
|
||||
"p95_pct": 1.8,
|
||||
"samples": 73,
|
||||
"status": "ok"
|
||||
},
|
||||
"cooling": {
|
||||
"available": true,
|
||||
"avg_fan_rpm": 6134.375,
|
||||
"notes": [
|
||||
"fan duty cycle unavailable on this host; RPM-only fan telemetry was collected"
|
||||
]
|
||||
},
|
||||
"gpus": [
|
||||
{
|
||||
"index": 0,
|
||||
"uuid": "GPU-9ee0af22-3dda-6f5c-1a13-35c63f324216",
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:49:00.0",
|
||||
"vbios": "96.00.74.00.01",
|
||||
"compute_capability": "9.0",
|
||||
"backend": "cublasLt",
|
||||
"status": "OK",
|
||||
"power_limit_w": 700,
|
||||
"default_power_limit_w": 700,
|
||||
"max_graphics_clock_mhz": 1980,
|
||||
"max_memory_clock_mhz": 2619,
|
||||
"locked_graphics_clock_mhz": 1980,
|
||||
"locked_memory_clock_mhz": 2619,
|
||||
"baseline": {
|
||||
"duration_sec": 20.631671609,
|
||||
"samples": 1,
|
||||
"avg_temp_c": 68,
|
||||
"p95_temp_c": 68,
|
||||
"avg_power_w": 252.73,
|
||||
"p95_power_w": 252.73,
|
||||
"avg_graphics_clock_mhz": 1830,
|
||||
"p95_graphics_clock_mhz": 1830,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"steady": {
|
||||
"duration_sec": 603.015967594,
|
||||
"samples": 12,
|
||||
"avg_temp_c": 86.33333333333333,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 586.6108333333333,
|
||||
"p95_power_w": 596.883,
|
||||
"avg_graphics_clock_mhz": 1023.75,
|
||||
"p95_graphics_clock_mhz": 1050,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 47.75,
|
||||
"clock_cv_pct": 1.8069167993832407,
|
||||
"power_cv_pct": 1.1372750232456648,
|
||||
"temp_cv_pct": 0.5460284024606545,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"precision_steady": [
|
||||
{
|
||||
"precision": "int8",
|
||||
"steady": {
|
||||
"duration_sec": 58.789805111,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 78.5,
|
||||
"p95_temp_c": 81.65,
|
||||
"avg_power_w": 612.745,
|
||||
"p95_power_w": 689.4024999999999,
|
||||
"avg_graphics_clock_mhz": 1222.5,
|
||||
"p95_graphics_clock_mhz": 1269.75,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 79.5,
|
||||
"clock_cv_pct": 4.294478527607362,
|
||||
"power_cv_pct": 13.90056222408995,
|
||||
"temp_cv_pct": 4.45859872611465,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 1095.3120993509376,
|
||||
"weighted_teraops_per_sec": 273.8280248377344,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp8",
|
||||
"steady": {
|
||||
"duration_sec": 107.153140143,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 86.5,
|
||||
"p95_temp_c": 86.95,
|
||||
"avg_power_w": 655.985,
|
||||
"p95_power_w": 694.4375,
|
||||
"avg_graphics_clock_mhz": 1102.5,
|
||||
"p95_graphics_clock_mhz": 1136.25,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 56,
|
||||
"clock_cv_pct": 3.4013605442176873,
|
||||
"power_cv_pct": 6.51310624480743,
|
||||
"temp_cv_pct": 0.5780346820809248,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 1093.6418424935764,
|
||||
"weighted_teraops_per_sec": 273.4104606233941,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp16",
|
||||
"steady": {
|
||||
"duration_sec": 162.145217446,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 593.8366666666667,
|
||||
"p95_power_w": 596.7439999999999,
|
||||
"avg_graphics_clock_mhz": 1065,
|
||||
"p95_graphics_clock_mhz": 1090.5,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 36.666666666666664,
|
||||
"clock_cv_pct": 1.9918500878494296,
|
||||
"power_cv_pct": 0.6401906405475871,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 699.5542180276906,
|
||||
"weighted_teraops_per_sec": 349.7771090138453,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp32",
|
||||
"steady": {
|
||||
"duration_sec": 221.67717378,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 595.0466666666666,
|
||||
"p95_power_w": 600.4,
|
||||
"avg_graphics_clock_mhz": 1100,
|
||||
"p95_graphics_clock_mhz": 1108.5,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 33.666666666666664,
|
||||
"clock_cv_pct": 0.6428243465332251,
|
||||
"power_cv_pct": 1.0242875350150111,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 296.763047018496,
|
||||
"weighted_teraops_per_sec": 296.763047018496,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp64",
|
||||
"steady": {
|
||||
"duration_sec": 283.877134755,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 599.5233333333334,
|
||||
"p95_power_w": 615.675,
|
||||
"avg_graphics_clock_mhz": 1090,
|
||||
"p95_graphics_clock_mhz": 1095,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 46.333333333333336,
|
||||
"clock_cv_pct": 0.6487218176023373,
|
||||
"power_cv_pct": 2.322932483924268,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp4",
|
||||
"steady": {
|
||||
"duration_sec": 353.430169396,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 594.26,
|
||||
"p95_power_w": 602.7589999999999,
|
||||
"avg_graphics_clock_mhz": 1020,
|
||||
"p95_graphics_clock_mhz": 1047,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 48.666666666666664,
|
||||
"clock_cv_pct": 2.4014605321403706,
|
||||
"power_cv_pct": 1.1839597038195573,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
}
|
||||
],
|
||||
"cooldown": {
|
||||
"duration_sec": 0,
|
||||
"samples": 0,
|
||||
"avg_temp_c": 0,
|
||||
"p95_temp_c": 0,
|
||||
"avg_power_w": 0,
|
||||
"p95_power_w": 0,
|
||||
"avg_graphics_clock_mhz": 0,
|
||||
"p95_graphics_clock_mhz": 0,
|
||||
"avg_memory_clock_mhz": 0,
|
||||
"p95_memory_clock_mhz": 0,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"throttle_counters": {
|
||||
"sw_power_cap_us": 100013,
|
||||
"sw_thermal_slowdown_us": 100013,
|
||||
"sync_boost_us": 0,
|
||||
"hw_thermal_slowdown_us": 0,
|
||||
"hw_power_brake_slowdown_us": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
},
|
||||
"precision_results": [
|
||||
{
|
||||
"name": "fp16_tensor",
|
||||
"category": "fp16_bf16",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 9856,
|
||||
"n": 9856,
|
||||
"k": 9856,
|
||||
"iterations": 28000,
|
||||
"teraops_per_sec": 178.71823088298666,
|
||||
"weight": 0.5,
|
||||
"weighted_teraops_per_sec": 89.35911544149333
|
||||
},
|
||||
{
|
||||
"name": "fp32_tf32",
|
||||
"category": "fp32_tf32",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 6912,
|
||||
"n": 6912,
|
||||
"k": 6912,
|
||||
"iterations": 28000,
|
||||
"teraops_per_sec": 61.64217593856,
|
||||
"weight": 1,
|
||||
"weighted_teraops_per_sec": 61.64217593856
|
||||
},
|
||||
{
|
||||
"name": "fp64",
|
||||
"category": "fp64",
|
||||
"supported": false,
|
||||
"weight": 2,
|
||||
"notes": "unsupported"
|
||||
},
|
||||
{
|
||||
"name": "fp8_e4m3",
|
||||
"category": "fp8",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 11392,
|
||||
"n": 11392,
|
||||
"k": 11392,
|
||||
"iterations": 28000,
|
||||
"teraops_per_sec": 275.97306768042665,
|
||||
"weight": 0.25,
|
||||
"weighted_teraops_per_sec": 68.99326692010666
|
||||
},
|
||||
{
|
||||
"name": "fp8_e5m2",
|
||||
"category": "fp8",
|
||||
"supported": false,
|
||||
"weight": 0.25,
|
||||
"notes": "unsupported"
|
||||
},
|
||||
{
|
||||
"name": "int8_tensor",
|
||||
"category": "int8",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 8832,
|
||||
"n": 8832,
|
||||
"k": 8832,
|
||||
"iterations": 28000,
|
||||
"teraops_per_sec": 128.60088385536,
|
||||
"weight": 0.25,
|
||||
"weighted_teraops_per_sec": 32.15022096384
|
||||
}
|
||||
],
|
||||
"scores": {
|
||||
"compute_score": 1269.4220752726699,
|
||||
"synthetic_score": 1193.7786414934699,
|
||||
"mixed_score": 252.14477926399996,
|
||||
"mixed_efficiency": 0.21121568982718245,
|
||||
"power_sustain_score": 0,
|
||||
"thermal_sustain_score": 99.98341453537307,
|
||||
"stability_score": 82.5511898186318,
|
||||
"interconnect_score": 12.1,
|
||||
"composite_score": 1150.5235745430427
|
||||
},
|
||||
"degradation_reasons": [
|
||||
"low_sm_clock_vs_target",
|
||||
"variance_too_high"
|
||||
]
|
||||
},
|
||||
{
|
||||
"index": 1,
|
||||
"uuid": "GPU-0182c11c-0c2c-aafd-0ada-113b64147ee6",
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:C8:00.0",
|
||||
"vbios": "96.00.74.00.01",
|
||||
"compute_capability": "9.0",
|
||||
"backend": "cublasLt",
|
||||
"status": "OK",
|
||||
"power_limit_w": 700,
|
||||
"default_power_limit_w": 700,
|
||||
"max_graphics_clock_mhz": 1980,
|
||||
"max_memory_clock_mhz": 2619,
|
||||
"locked_graphics_clock_mhz": 1980,
|
||||
"locked_memory_clock_mhz": 2619,
|
||||
"baseline": {
|
||||
"duration_sec": 20.631671609,
|
||||
"samples": 1,
|
||||
"avg_temp_c": 69,
|
||||
"p95_temp_c": 69,
|
||||
"avg_power_w": 270.57,
|
||||
"p95_power_w": 270.57,
|
||||
"avg_graphics_clock_mhz": 1830,
|
||||
"p95_graphics_clock_mhz": 1830,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"steady": {
|
||||
"duration_sec": 603.015967594,
|
||||
"samples": 12,
|
||||
"avg_temp_c": 86.16666666666667,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 552.3258333333333,
|
||||
"p95_power_w": 562.0735,
|
||||
"avg_graphics_clock_mhz": 917.5,
|
||||
"p95_graphics_clock_mhz": 960,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 41.666666666666664,
|
||||
"clock_cv_pct": 4.3168881519768005,
|
||||
"power_cv_pct": 1.1828641347909323,
|
||||
"temp_cv_pct": 0.43250831286262864,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"precision_steady": [
|
||||
{
|
||||
"precision": "int8",
|
||||
"steady": {
|
||||
"duration_sec": 58.789805111,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 81.5,
|
||||
"p95_temp_c": 84.65,
|
||||
"avg_power_w": 618.75,
|
||||
"p95_power_w": 691.767,
|
||||
"avg_graphics_clock_mhz": 1185,
|
||||
"p95_graphics_clock_mhz": 1239,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 79,
|
||||
"clock_cv_pct": 5.063291139240507,
|
||||
"power_cv_pct": 13.111919191919192,
|
||||
"temp_cv_pct": 4.294478527607362,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 994.2685477502977,
|
||||
"weighted_teraops_per_sec": 248.56713693757442,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp8",
|
||||
"steady": {
|
||||
"duration_sec": 107.153140143,
|
||||
"samples": 2,
|
||||
"avg_temp_c": 86.5,
|
||||
"p95_temp_c": 86.95,
|
||||
"avg_power_w": 595.51,
|
||||
"p95_power_w": 611.8629999999999,
|
||||
"avg_graphics_clock_mhz": 960,
|
||||
"p95_graphics_clock_mhz": 973.5,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 53,
|
||||
"clock_cv_pct": 1.5625,
|
||||
"power_cv_pct": 3.0511662272673776,
|
||||
"temp_cv_pct": 0.5780346820809248,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 1073.9294805164031,
|
||||
"weighted_teraops_per_sec": 268.4823701291008,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp16",
|
||||
"steady": {
|
||||
"duration_sec": 162.145217446,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87.9,
|
||||
"avg_power_w": 554.0466666666666,
|
||||
"p95_power_w": 604.477,
|
||||
"avg_graphics_clock_mhz": 945,
|
||||
"p95_graphics_clock_mhz": 999,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 34.666666666666664,
|
||||
"clock_cv_pct": 5.184105275731594,
|
||||
"power_cv_pct": 10.324119223592762,
|
||||
"temp_cv_pct": 0.9385018171583058,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 560.1539979389611,
|
||||
"weighted_teraops_per_sec": 280.07699896948054,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp32",
|
||||
"steady": {
|
||||
"duration_sec": 221.67717378,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 86,
|
||||
"p95_temp_c": 86,
|
||||
"avg_power_w": 565.8233333333334,
|
||||
"p95_power_w": 594,
|
||||
"avg_graphics_clock_mhz": 1075,
|
||||
"p95_graphics_clock_mhz": 1092,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 35,
|
||||
"clock_cv_pct": 1.315547499881949,
|
||||
"power_cv_pct": 4.1295451621968935,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"teraops_per_sec": 293.76899847290883,
|
||||
"weighted_teraops_per_sec": 293.76899847290883,
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp64",
|
||||
"steady": {
|
||||
"duration_sec": 283.877134755,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 565.93,
|
||||
"p95_power_w": 581.971,
|
||||
"avg_graphics_clock_mhz": 1055,
|
||||
"p95_graphics_clock_mhz": 1077,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 42,
|
||||
"clock_cv_pct": 1.7732973397032898,
|
||||
"power_cv_pct": 2.3474912323387347,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
},
|
||||
{
|
||||
"precision": "fp4",
|
||||
"steady": {
|
||||
"duration_sec": 353.430169396,
|
||||
"samples": 3,
|
||||
"avg_temp_c": 87,
|
||||
"p95_temp_c": 87,
|
||||
"avg_power_w": 547.1333333333333,
|
||||
"p95_power_w": 548.109,
|
||||
"avg_graphics_clock_mhz": 885,
|
||||
"p95_graphics_clock_mhz": 939,
|
||||
"avg_memory_clock_mhz": 2619,
|
||||
"p95_memory_clock_mhz": 2619,
|
||||
"avg_usage_pct": 100,
|
||||
"avg_mem_usage_pct": 42.333333333333336,
|
||||
"clock_cv_pct": 5.535570040187973,
|
||||
"power_cv_pct": 0.15226269553094146,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
}
|
||||
}
|
||||
],
|
||||
"cooldown": {
|
||||
"duration_sec": 0,
|
||||
"samples": 0,
|
||||
"avg_temp_c": 0,
|
||||
"p95_temp_c": 0,
|
||||
"avg_power_w": 0,
|
||||
"p95_power_w": 0,
|
||||
"avg_graphics_clock_mhz": 0,
|
||||
"p95_graphics_clock_mhz": 0,
|
||||
"avg_memory_clock_mhz": 0,
|
||||
"p95_memory_clock_mhz": 0,
|
||||
"avg_usage_pct": 0,
|
||||
"avg_mem_usage_pct": 0,
|
||||
"clock_cv_pct": 0,
|
||||
"power_cv_pct": 0,
|
||||
"temp_cv_pct": 0,
|
||||
"clock_drift_pct": 0
|
||||
},
|
||||
"throttle_counters": {
|
||||
"sw_power_cap_us": 0,
|
||||
"sw_thermal_slowdown_us": 0,
|
||||
"sync_boost_us": 0,
|
||||
"hw_thermal_slowdown_us": 0,
|
||||
"hw_power_brake_slowdown_us": 0
|
||||
},
|
||||
"ecc": {
|
||||
"corrected": 0,
|
||||
"uncorrected": 0
|
||||
},
|
||||
"precision_results": [
|
||||
{
|
||||
"name": "fp16_tensor",
|
||||
"category": "fp16_bf16",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 9856,
|
||||
"n": 9856,
|
||||
"k": 9856,
|
||||
"iterations": 25760,
|
||||
"teraops_per_sec": 164.42077241234773,
|
||||
"weight": 0.5,
|
||||
"weighted_teraops_per_sec": 82.21038620617387
|
||||
},
|
||||
{
|
||||
"name": "fp32_tf32",
|
||||
"category": "fp32_tf32",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 6912,
|
||||
"n": 6912,
|
||||
"k": 6912,
|
||||
"iterations": 25760,
|
||||
"teraops_per_sec": 56.710801863475204,
|
||||
"weight": 1,
|
||||
"weighted_teraops_per_sec": 56.710801863475204
|
||||
},
|
||||
{
|
||||
"name": "fp64",
|
||||
"category": "fp64",
|
||||
"supported": false,
|
||||
"weight": 2,
|
||||
"notes": "unsupported"
|
||||
},
|
||||
{
|
||||
"name": "fp8_e4m3",
|
||||
"category": "fp8",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 11392,
|
||||
"n": 11392,
|
||||
"k": 11392,
|
||||
"iterations": 25760,
|
||||
"teraops_per_sec": 253.89522226599254,
|
||||
"weight": 0.25,
|
||||
"weighted_teraops_per_sec": 63.473805566498136
|
||||
},
|
||||
{
|
||||
"name": "fp8_e5m2",
|
||||
"category": "fp8",
|
||||
"supported": false,
|
||||
"weight": 0.25,
|
||||
"notes": "unsupported"
|
||||
},
|
||||
{
|
||||
"name": "int8_tensor",
|
||||
"category": "int8",
|
||||
"supported": true,
|
||||
"lanes": 16,
|
||||
"m": 8832,
|
||||
"n": 8832,
|
||||
"k": 8832,
|
||||
"iterations": 25760,
|
||||
"teraops_per_sec": 118.3128131469312,
|
||||
"weight": 0.25,
|
||||
"weighted_teraops_per_sec": 29.5782032867328
|
||||
}
|
||||
],
|
||||
"scores": {
|
||||
"compute_score": 1160.4874635859287,
|
||||
"synthetic_score": 1090.8955045090647,
|
||||
"mixed_score": 231.97319692288,
|
||||
"mixed_efficiency": 0.21264474549950116,
|
||||
"power_sustain_score": 0,
|
||||
"thermal_sustain_score": 100,
|
||||
"stability_score": 75.33829122588782,
|
||||
"interconnect_score": 12.1,
|
||||
"composite_score": 1030.9140807492215
|
||||
},
|
||||
"degradation_reasons": [
|
||||
"low_sm_clock_vs_target",
|
||||
"variance_too_high"
|
||||
]
|
||||
}
|
||||
],
|
||||
"interconnect": {
|
||||
"status": "OK",
|
||||
"attempted": true,
|
||||
"supported": true,
|
||||
"selected_gpu_indices": [
|
||||
0,
|
||||
1
|
||||
],
|
||||
"avg_algbw_gbps": 11.920000000000002,
|
||||
"max_algbw_gbps": 12.1,
|
||||
"avg_busbw_gbps": 11.920000000000002,
|
||||
"max_busbw_gbps": 12.1
|
||||
},
|
||||
"server_power": {
|
||||
"available": false,
|
||||
"notes": [
|
||||
"IPMI power reading unavailable; server-side power characterization skipped"
|
||||
]
|
||||
}
|
||||
}
|
||||
1238
perf/perf-20260416-114112/result.json
Normal file
1238
perf/perf-20260416-114112/result.json
Normal file
File diff suppressed because it is too large
Load Diff
149
power/power-20260416-105918/result.json
Normal file
149
power/power-20260416-105918/result.json
Normal file
@@ -0,0 +1,149 @@
|
||||
{
|
||||
"benchmark_version": "2",
|
||||
"generated_at": "2026-04-16T10:59:18.318577288Z",
|
||||
"hostname": "debian",
|
||||
"server_model": "MLT-S06",
|
||||
"benchmark_profile": "standard",
|
||||
"selected_gpu_indices": [
|
||||
0,
|
||||
1,
|
||||
2,
|
||||
3
|
||||
],
|
||||
"recommended_slot_order": [
|
||||
2,
|
||||
0,
|
||||
1,
|
||||
3
|
||||
],
|
||||
"ramp_steps": [
|
||||
{
|
||||
"step_index": 1,
|
||||
"gpu_indices": [
|
||||
2
|
||||
],
|
||||
"new_gpu_index": 2,
|
||||
"new_gpu_stable_limit_w": 700,
|
||||
"total_observed_power_w": 698.892,
|
||||
"avg_observed_power_w": 698.892,
|
||||
"status": "OK"
|
||||
},
|
||||
{
|
||||
"step_index": 2,
|
||||
"gpu_indices": [
|
||||
2,
|
||||
0
|
||||
],
|
||||
"new_gpu_index": 0,
|
||||
"new_gpu_stable_limit_w": 700,
|
||||
"total_observed_power_w": 1397.43,
|
||||
"avg_observed_power_w": 698.715,
|
||||
"status": "OK"
|
||||
},
|
||||
{
|
||||
"step_index": 3,
|
||||
"gpu_indices": [
|
||||
2,
|
||||
0,
|
||||
1
|
||||
],
|
||||
"new_gpu_index": 1,
|
||||
"new_gpu_stable_limit_w": 700,
|
||||
"total_observed_power_w": 2038.9585,
|
||||
"avg_observed_power_w": 679.6528333333333,
|
||||
"status": "PARTIAL",
|
||||
"notes": [
|
||||
"GPU 2 was re-derated from 700 W to 680 W under combined thermal load.",
|
||||
"GPU 0 was re-derated from 700 W to 660 W under combined thermal load."
|
||||
]
|
||||
},
|
||||
{
|
||||
"step_index": 4,
|
||||
"gpu_indices": [
|
||||
2,
|
||||
0,
|
||||
1,
|
||||
3
|
||||
],
|
||||
"new_gpu_index": 3,
|
||||
"new_gpu_stable_limit_w": 680,
|
||||
"total_observed_power_w": 2647.3525,
|
||||
"avg_observed_power_w": 661.838125,
|
||||
"derated": true,
|
||||
"status": "PARTIAL",
|
||||
"notes": [
|
||||
"GPU 2 was re-derated from 680 W to 660 W under combined thermal load.",
|
||||
"GPU 0 was re-derated from 660 W to 650 W under combined thermal load.",
|
||||
"GPU 1 was re-derated from 700 W to 660 W under combined thermal load."
|
||||
]
|
||||
}
|
||||
],
|
||||
"overall_status": "PARTIAL",
|
||||
"platform_max_tdp_w": 2650,
|
||||
"server_power": {
|
||||
"available": true,
|
||||
"idle_w": 631.8333333333334,
|
||||
"loaded_w": 1401.1466666666668,
|
||||
"delta_w": 769.3133333333334,
|
||||
"gpu_reported_sum_w": 2650,
|
||||
"reporting_ratio": 0.29030691823899374
|
||||
},
|
||||
"findings": [
|
||||
"Recommended slot order for installation based on single-card targeted_power: 2,0,1,3.",
|
||||
"Ramp step 4 (GPU 3) required derating to 680 W under combined thermal load.",
|
||||
"GPU 2 required additional derating from 700 W (single-card) to 660 W under full-system thermal load.",
|
||||
"GPU 0 required additional derating from 700 W (single-card) to 650 W under full-system thermal load.",
|
||||
"GPU 1 required additional derating from 700 W (single-card) to 660 W under full-system thermal load.",
|
||||
"GPU 3 required additional derating from 700 W (single-card) to 680 W under full-system thermal load."
|
||||
],
|
||||
"gpus": [
|
||||
{
|
||||
"index": 2,
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:C8:00.0",
|
||||
"applied_power_limit_w": 700,
|
||||
"stable_power_limit_w": 660,
|
||||
"max_observed_power_w": 698.892,
|
||||
"max_observed_temp_c": 77.6,
|
||||
"calibration_attempts": 1,
|
||||
"derated": true,
|
||||
"status": "OK"
|
||||
},
|
||||
{
|
||||
"index": 0,
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:49:00.0",
|
||||
"applied_power_limit_w": 700,
|
||||
"stable_power_limit_w": 650,
|
||||
"max_observed_power_w": 698.664,
|
||||
"max_observed_temp_c": 75,
|
||||
"calibration_attempts": 1,
|
||||
"derated": true,
|
||||
"status": "OK"
|
||||
},
|
||||
{
|
||||
"index": 1,
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:5A:00.0",
|
||||
"applied_power_limit_w": 700,
|
||||
"stable_power_limit_w": 660,
|
||||
"max_observed_power_w": 698.6175,
|
||||
"max_observed_temp_c": 76.75,
|
||||
"calibration_attempts": 1,
|
||||
"derated": true,
|
||||
"status": "OK"
|
||||
},
|
||||
{
|
||||
"index": 3,
|
||||
"name": "NVIDIA H100 80GB HBM3",
|
||||
"bus_id": "00000000:D8:00.0",
|
||||
"applied_power_limit_w": 700,
|
||||
"stable_power_limit_w": 680,
|
||||
"max_observed_power_w": 698.205,
|
||||
"max_observed_temp_c": 77.25,
|
||||
"calibration_attempts": 1,
|
||||
"derated": true,
|
||||
"status": "OK"
|
||||
}
|
||||
]
|
||||
}
|
||||
Reference in New Issue
Block a user