4 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
Open Discovery	1692ef3f58	fix(deep_cluster): resolve bus ids and UUIDs in NUMA lookup ... which_numa passed the CUDA_VISIBLE_DEVICES entry straight to int(), so a device pinned by PCI bus id or GPU UUID - both accepted forms - crashed with ValueError during NUMA binding at engine startup, before the existing "no such GPU found" path could produce an actionable error. Resolve nvidia identifiers against procfs (ordinal, bus id, full UUID, or unambiguous UUID prefix) and route unresolvable ones through the documented IndexError -> AssertionError path. Covered by CPU-only tests over a fake procfs/sysfs tree.	2026-10-20 19:50:19 +05:30
hkeee21	774ce244b8	update Q&A EN doc	2026-10-20 19:03:13 +08:00
Jiashi Li	8b16149741	update Q&A documentation	2026-10-20 17:06:57 +08:00
Jiashi Li	d585e6542c	initial commit	2026-10-20 12:09:42 +08:00