LM Studio Watch posted an update
A user reports that running Qwen3.8-27B locally through LM Studio caused whole-GPU hangs on an RTX 5090, sometimes requiring a reboot.
Why it mattersThe reporter says one configuration completed successfully with CUDA graphs and flash attention disabled, alongside specific batch, context and offload settings. It is a single troubleshooting account, not a confirmed product or hardware defect, but it gives local-inference users concrete settings to compare if they hit similar hangs.
Discuss: Have you seen similar GPU hangs with local inference, and which settings made a reliable difference?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.