accelerator-utilization (AU) threshold for a valid MLPerf Storage v3.0 training result, by workload
Unet3D ≥90%, RetinaNet ≥85%observed
| Value kind | observed — Reported measurements, counts, and specifications keep the precision and scope stated by their source; an exact specification is not treated as a range. |
|---|---|
| Scope | v3.0 emulates NVIDIA B200 or AMD MI355 accelerators. The v2.0 round (August 2025, H100 emulation) used 90% for 3D U-Net and ResNet-50 and 70% for CosmoFlow; thresholds and emulated accelerators are not comparable across versions. |
| As of | 2026-09 |
| Source | MLCommons, MLPerf Storage benchmark definitions (v3.0); v3.0 results published 2026-09-01 · Benchmark definition: a result is valid only if the simulated accelerators stayed busy at least the threshold: 90% for Unet3D, 85% for RetinaNet |
| Review | checking…review by 2027-09-01 · standard cadence |
| Recorded changes | last 2026-10-12 · 3 revisions tracked |
| Claim id | accelerator-utilization-threshold-for-a-passing |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.