Oryx Pro 10 Laptop - Repeated GPU Has Fallen Of The Bus with External Monitor

I’m currently running PopOS with latest 535 driver. I’ve tried 470, 525, 535. All give the same problem. Eventually after some random amount of time, the built-in and external monitors will completely freeze. The system is accessible via ssh, but no key combo allows dropping to a shell to reboot.

This is the message in syslog
an 25 09:42:17 blinky kernel: [75164.301780] NVRM: GPU at PCI:0000:01:00: GPU-c637a655-4250-786a-3d8d-bde6ad41ff4d
Jan 25 09:42:17 blinky kernel: [75164.301787] NVRM: Xid (PCI:0000:01:00): 79, pid=‘’, name=, GPU has fallen off the bus.
Jan 25 09:42:17 blinky kernel: [75164.301792] NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
Jan 25 09:42:17 blinky kernel: [75164.301802] NVRM: A GPU crash dump has been created. If possible, please run
Jan 25 09:42:17 blinky kernel: [75164.301802] NVRM: nvidia-bug-report.sh as root to collect this data before
Jan 25 09:42:17 blinky kernel: [75164.301802] NVRM: the NVIDIA kernel module is unloaded.
Jan 25 09:42:18 blinky kernel: [75165.844531] NVRM: Error in service of callback
Jan 25 09:42:27 blinky kernel: [75174.544581] Asynchronous wait on fence NVIDIA:nvidia.prime:8fcaf timed out (hint:intel_atomic_commit_ready [i915])

nvidia-bug-report.log.gz (368.7 KB)

Bug report from the most recent crash

To add a bit more detail. At the time of these crashes, the GPU is not under load. It typically happens when navigating with Google Chrome or typing in VSCode on the external monitor.

Sometimes it works for minutes, sometimes for hours. I’d guess the gpu is broken, you should check warranty status and contact System76.