RTX 6000 Blackwell crashes under load, points to firmware bug
AIFlow_ML · x · 2026-09-01
Users report that the NVIDIA RTX PRO 6000 Blackwell suffers from deterministic GSP kernel panics (Xid 120) under sustained high-load workloads, causing CUDA errors and requiring a reboot. The identical crash PC address across runs suggests a firmware code path defect rather than physical hardware degradation.
Related event: RTX PRO 6000 Blackwell Suffers Deterministic GSP Crashes Under Heavy Load(2 posts)→
More from Infra
- Building a Private AI OS on 4x RTX 2080 Ti — askincihan · 2026-09-01
- Llama Mac app adds REST API request builder for llama.cpp — ggerganov · 2026-09-01
- DRAM Share in AI Inference Servers to Drop ~40% as Flash Takes Over — AccBalanced · 2026-09-01
- MediaTek CEO Highlights Substrate Bottleneck in AI Infrastructure — AccBalanced · 2026-09-01
- llama.cpp Merges Multiple Fixes and MTP Support for Qwen4Exp — jacek2023 · 2026-09-01
- Saudi HUMAIN partners with AMD, Nvidia and Qualcomm on sovereign AI compute — HUMAIN · 2026-09-01