Expanded Reset Support Coming For AMDGPU To Recover From More GPU Compute Hangs
AMD has introduced expanded reset support for its AMDGPU kernel driver to enhance recovery from GPU compute hangs. A series of 42 patches has been posted to enable pipe reset capabilities, which are necessary when queue resets fail to restore functionality. This development aims to improve the reliability of AMD GPUs in compute workloads on Linux systems.
- ▪The new patches enable pipe reset capabilities for AMDGPU to recover from compute hangs.
- ▪Queue reset support has existed, but pipe reset support is a new enhancement for more complex recovery scenarios.
- ▪Coordination across all components using compute queues is required for effective pipe resets.
Opening excerpt (first ~120 words) tap to expand
Expanded Reset Support Coming For AMDGPU To Recover From More GPU Compute Hangs Written by Michael Larabel in Radeon on 22 May 2026 at 05:50 AM EDT. Add A Comment A set of 42 patches were posted on Thursday for the AMDGPU kernel driver and associated AMDKFD compute driver code for enabling pipe reset capabilities for compute workloads. While there has long been queue reset support for the AMDGPU driver in helping to recover from hangs, pipe reset support is now being worked on as the next step in the recovery when simply resetting a lone queue doesn't bring the GPU back into its desired working state. AMDGPU maintainer Alex Deucher authored many of the patches in this big patch series for the pipe reset support.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Phoronix.