{ "latest": { "run_id": "campaign_full_20260306T000425Z", "timestamp_utc": "2026-03-06T00:04:39.891676+00:00", "mode": "nexa_vortex", "dist": "ddp", "backend": "nccl", "world_size": 8, "model_name": "distilbert-base-uncased", "dataset_name": "ag_news", "train_runtime_sec": 55.6958, "samples_per_sec": 2154.5621, "steps_per_sec": 269.3203, "eval_loss": 0.186266, "loss_final": 0.256069, "gpu_util_mean": 28.4856, "gpu_util_p95": 65.0, "h2d_time_ms_mean": 0.0779, "compute_time_ms_mean": 15.2453, "comm_time_ms_mean": 0.6829, "idle_gap_ms_mean": 1.5741, "idle_gap_ms_p95": 0.1459, "paths": { "run_dir": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z", "summary_md": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/summary.md", "summary_svg": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/summary.svg", "metrics_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/train_metrics.json", "run_config_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/run_config.json", "gpu_telemetry_csv": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/gpu_telemetry.csv" }, "published": { "summary_svg": "results/artifacts/distributed/campaign_full_20260306T000425Z__summary.svg", "metrics_json": "results/artifacts/distributed/campaign_full_20260306T000425Z__metrics.json" } }, "runs": [ { "run_id": "campaign_livecheck_20260306T003758Z", "timestamp_utc": "2026-03-06T00:38:05.692029+00:00", "mode": "nexa_vortex", "dist": "ddp", "backend": "nccl", "world_size": 1, "model_name": "distilbert-base-uncased", "dataset_name": "ag_news", "train_runtime_sec": 15.844, "samples_per_sec": 315.5767, "steps_per_sec": 39.4471, "eval_loss": 0.319241, "loss_final": 0.082397, "gpu_util_mean": 2.1875, "gpu_util_p95": 23.0, "h2d_time_ms_mean": 0.0737, "compute_time_ms_mean": 13.7176, "comm_time_ms_mean": 0.0, "idle_gap_ms_mean": 3.9947, "idle_gap_ms_p95": 0.1257, "paths": { "run_dir": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_livecheck_20260306T003758Z", "summary_md": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_livecheck_20260306T003758Z/summary.md", "summary_svg": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_livecheck_20260306T003758Z/summary.svg", "metrics_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_livecheck_20260306T003758Z/train_metrics.json", "run_config_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_livecheck_20260306T003758Z/run_config.json", "gpu_telemetry_csv": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_livecheck_20260306T003758Z/gpu_telemetry.csv" }, "published": { "summary_svg": "results/artifacts/distributed/campaign_livecheck_20260306T003758Z__summary.svg", "metrics_json": "results/artifacts/distributed/campaign_livecheck_20260306T003758Z__metrics.json" } }, { "run_id": "campaign_rerun3_20260306T003221Z", "timestamp_utc": "2026-03-06T00:32:29.654312+00:00", "mode": "nexa_vortex", "dist": "ddp", "backend": "nccl", "world_size": 1, "model_name": "distilbert-base-uncased", "dataset_name": "ag_news", "train_runtime_sec": 40.7572, "samples_per_sec": 490.7107, "steps_per_sec": 61.3388, "eval_loss": 0.201228, "loss_final": 0.328766, "gpu_util_mean": 3.5592, "gpu_util_p95": 43.0, "h2d_time_ms_mean": 0.0722, "compute_time_ms_mean": 10.9129, "comm_time_ms_mean": 0.0, "idle_gap_ms_mean": 1.086, "idle_gap_ms_p95": 0.125, "paths": { "run_dir": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_rerun3_20260306T003221Z", "summary_md": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_rerun3_20260306T003221Z/summary.md", "summary_svg": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_rerun3_20260306T003221Z/summary.svg", "metrics_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_rerun3_20260306T003221Z/train_metrics.json", "run_config_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_rerun3_20260306T003221Z/run_config.json", "gpu_telemetry_csv": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_rerun3_20260306T003221Z/gpu_telemetry.csv" }, "published": { "summary_svg": "results/artifacts/distributed/campaign_rerun3_20260306T003221Z__summary.svg", "metrics_json": "results/artifacts/distributed/campaign_rerun3_20260306T003221Z__metrics.json" } }, { "run_id": "campaign_full_20260306T000425Z", "timestamp_utc": "2026-03-06T00:04:39.891676+00:00", "mode": "nexa_vortex", "dist": "ddp", "backend": "nccl", "world_size": 8, "model_name": "distilbert-base-uncased", "dataset_name": "ag_news", "train_runtime_sec": 55.6958, "samples_per_sec": 2154.5621, "steps_per_sec": 269.3203, "eval_loss": 0.186266, "loss_final": 0.256069, "gpu_util_mean": 28.4856, "gpu_util_p95": 65.0, "h2d_time_ms_mean": 0.0779, "compute_time_ms_mean": 15.2453, "comm_time_ms_mean": 0.6829, "idle_gap_ms_mean": 1.5741, "idle_gap_ms_p95": 0.1459, "paths": { "run_dir": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z", "summary_md": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/summary.md", "summary_svg": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/summary.svg", "metrics_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/train_metrics.json", "run_config_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/run_config.json", "gpu_telemetry_csv": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_full_20260306T000425Z/gpu_telemetry.csv" }, "published": { "summary_svg": "results/artifacts/distributed/campaign_full_20260306T000425Z__summary.svg", "metrics_json": "results/artifacts/distributed/campaign_full_20260306T000425Z__metrics.json" } }, { "run_id": "campaign_stable_20260306T000153Z", "timestamp_utc": "2026-03-06T00:02:08.333480+00:00", "mode": "nexa_vortex", "dist": "ddp", "backend": "nccl", "world_size": 8, "model_name": "distilbert-base-uncased", "dataset_name": "ag_news", "train_runtime_sec": 34.5195, "samples_per_sec": 1738.1486, "steps_per_sec": 217.3845, "eval_loss": 0.238709, "loss_final": 0.023407, "gpu_util_mean": 25.5588, "gpu_util_p95": 66.0, "h2d_time_ms_mean": 0.0761, "compute_time_ms_mean": 17.3532, "comm_time_ms_mean": 0.9353, "idle_gap_ms_mean": 3.1611, "idle_gap_ms_p95": 0.1239, "paths": { "run_dir": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_stable_20260306T000153Z", "summary_md": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_stable_20260306T000153Z/summary.md", "summary_svg": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_stable_20260306T000153Z/summary.svg", "metrics_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_stable_20260306T000153Z/train_metrics.json", "run_config_json": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_stable_20260306T000153Z/run_config.json", "gpu_telemetry_csv": "benchmark/remote_results/runpod_h100_8x/campaign_v5/campaign_stable_20260306T000153Z/gpu_telemetry.csv" }, "published": { "summary_svg": "results/artifacts/distributed/campaign_stable_20260306T000153Z__summary.svg", "metrics_json": "results/artifacts/distributed/campaign_stable_20260306T000153Z__metrics.json" } }, { "run_id": "campaign_live_20260305T233837Z", "timestamp_utc": "2026-03-05T23:38:51.650069+00:00", "mode": "nexa_vortex", "dist": "ddp", "backend": "nccl", "world_size": 8, "model_name": "distilbert-base-uncased", "dataset_name": "ag_news", "train_runtime_sec": 52.3724, "samples_per_sec": 2291.2819, "steps_per_sec": 286.4102, "eval_loss": 0.186537, "loss_final": 0.255463, "gpu_util_mean": 28.42, "gpu_util_p95": 63.0, "h2d_time_ms_mean": 0.0855, "compute_time_ms_mean": 14.6285, "comm_time_ms_mean": 0.6477, "idle_gap_ms_mean": 0.4476, "idle_gap_ms_p95": 0.4215, "paths": { "run_dir": "benchmark/remote_results/runpod_h100_8x/campaign_v4/campaign_live_20260305T233837Z", "summary_md": "benchmark/remote_results/runpod_h100_8x/campaign_v4/campaign_live_20260305T233837Z/summary.md", "summary_svg": "benchmark/remote_results/runpod_h100_8x/campaign_v4/campaign_live_20260305T233837Z/summary.svg", "metrics_json": "benchmark/remote_results/runpod_h100_8x/campaign_v4/campaign_live_20260305T233837Z/train_metrics.json", "run_config_json": "benchmark/remote_results/runpod_h100_8x/campaign_v4/campaign_live_20260305T233837Z/run_config.json", "gpu_telemetry_csv": "benchmark/remote_results/runpod_h100_8x/campaign_v4/campaign_live_20260305T233837Z/gpu_telemetry.csv" }, "published": { "summary_svg": "results/artifacts/distributed/campaign_live_20260305T233837Z__summary.svg", "metrics_json": "results/artifacts/distributed/campaign_live_20260305T233837Z__metrics.json" } } ], "notes": "Latest successful distributed runs from RunPod H100 campaign artifacts.", "visuals": { "throughput_svg": "results/artifacts/images/latest_distributed_throughput.svg", "pipeline_svg": "results/artifacts/images/latest_distributed_pipeline.svg" } }