Specs: GTX 1060 6GB, linux-cachyos-bore (reproducable on the regular kernel), Arch running KDE Wayland (this issue is also present on X11)
As of recent (and i don’t know how recent - i can reproduce this bug on a btrfs snapshot from December 27th and that’s my oldest one), I can only play gpu demanding games for a minute or so, before all my screens go black and the gpu fans spin up to 100%. The PC still works, just without the graphics. This isssue is not related to the temperature. Monitoring it with LACT, I’ve only ever seen it get to the ~70s - or maybe the temp sensors are broken. The issue is not VRAM related either. Playing Elden Ring, which is the most demanding game I have, the VRAM usage only rose to around 4GB. It’s not XWayland related either, since Wayland-native games like Minecraft with this modification that enables native Wayland support, and using shaders also crashes the gpu. It’s worth noting that I can still play lighter games just fine.
I’m attaching the journalctl of my previous session. It crashed while playing Elden Ring. I used REISUB at the end to shut off my PC.
ercrash.txt (602.3 KB)
Theese lines (5383, 5384) caught my attention:
sty 10 13:10:45 mattys kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object
sty 10 13:10:45 mattys kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object
This may be something VRAM related, but again - it’s not getting filled up while playing.
I’m also attaching nvidia-bug-report
nvidia-bug-report.log.gz (1.7 MB)
Of course, this could be a hardware issue. My PSU was bought last year, but the GPU is quite old by now. I will have to test this on Windows when i have the time to install it.