failures per thousand node-days on Meta's RSC-1 cluster (11 months, ~80%+ utilization)
6.50 / 1000observed
| 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 | Job failures on Meta's A100 RSC-1 research cluster over an 11-month window at ~80%+ utilization, divided by allocated node-days (job runtime x allocated nodes). An event-per-exposure rate for one named cluster and workload mix, not unique failed-FRU counts and not a portable component failure intensity; the companion RSC-2 figure is 2.34 on the same basis. |
| As of | 2024 |
| Source | Meta, Revisiting Reliability (arXiv 2410.21680) |
| Review | checking…review by 2026-08-14 · standard cadence |
| Recorded changes | last 2026-09-16 · 2 revisions tracked |
| Claim id | failures-per-thousand-node-days-on-meta-s-rsc-1 |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.