RTX 5070 Ti dies mid 6-hour CUDA job: Blackwell PCIe 5.0 hang issue officially acknowledged
Reasonable-Height704 · reddit · 2026-10-09
A user with years of stable heavy compute on 3090/4090/2080Ti reports that his new RTX 5070 Ti crashed near the end of a 6-hour CUDA job, with the journal logging NVRM: krcWatchdogIMPL: RC watchdog: GPU is probably locked! Notify Timeout Seconds: 7. The code was fine and the crash couldn't be reliably reproduced.
Digging further, this turns out to be a known Blackwell issue (RTX 5070/5080/5090): gamegpu covered these cards freezing due to an unidentified PCIe Gen 5 fault, and NVIDIA only officially acknowledged it (XID 79 / GSP timeout) in recent months. The author is surprised there isn't more noise about it — a warning for anyone running long local inference or compute workloads on Blackwell.
More from Infra
- PartyKit shuts down free hosted platform 2.5 years after Cloudflare acquisition — threepointone · 2026-10-09
- 3090+3060 local video gen: which model — Wan 2.2, MiniMax H3 or LTX-2? — itspacedev · 2026-10-09
- Samsung expects 780% quarterly operating profit jump on AI boom — chemist_slime · 2026-10-09
- TRL v1.15 ships fused LM head: 82% less peak VRAM, 7x longer sequences — QGallouedec · 2026-10-09
- Dev streams a 66GB unquantized 22B video model on an iGPU with only 15.6GB shared RAM — Business_Swordfish_5 · 2026-10-09
- Asia-Pacific Needs $280 Billion to Build 26.5 GW of Planned Data Centers, Johor Leads — shashib · 2026-10-09