Files
hamlab/internal/metrics
Mitchell-JohnsonandClaude Opus 5.5 574d9855ce GPU: a slow read is late, not stalled; a stall clears when the read returns
nvidia-smi on a busy card can take past the 3 s deadline, which latched the card
as stalled until restart. A late read now carries the last readings and is used
when it lands; no second read starts while one is out; only one out for 15 s
(past both nvidia-smi timeouts) marks the card stalled, and its return clears it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 18:42:13 -08:00
..