You are not logged in.
Behaviour of Bug :
Both screens go black. journalctl spams "pageflip timeout error" whenever it happens. It is now happening randomly, can’t even open terminal with Ctrl + Alt + T and reboot. Though if I’m in Discord vc when it crashed, I can still continue talking to other peeps. I have a script running to play keystroke sounds, it also keeps working, though I created a cpp application to send "system(‘reboot’); " when I press Ctrl + Meta + Alt + R. That apparently doesn’t work.
It used to happen previously too, then I reinstalled arch and it’s still happening. Earlier it happened only when doing something in fullscreen (playing game in fullscreen or playing youtube/anime in fullscreen). Now it happens seemingly randomly whenever GPU is experiencing more than average usage.
Specifications : RTX 3080, AMD 5600X, 16 GB RAM, Non Identical Dual Monitors
Nvidia related packages list :
> pacman -Qe | grep nvidia
lib32-nvidia-utils 570.144-1
lib32-opencl-nvidia 570.144-1
nvidia-open 570.144-4
nvidia-settings 570.144-1
nvidia-utils 570.144-3
opencl-nvidia 570.144-3
Below are all kinds of logs that I know of :
dmesg : https://forums.developer.nvidia.com/upl … CbUipU.txt
full journalctl : https://forums.developer.nvidia.com/upl … E5ntgh.txt
wayland journalctl : https://forums.developer.nvidia.com/upl … 3a7Fk0.txt
nvidia-bug-report.log : https://forums.developer.nvidia.com/upl … Oc4M6m.log
Offline
This seems to be the issue
May 11 17:03:39 sybil kernel: NVRM: Xid (PCI:0000:06:00): 79, GPU has fallen off the busPossible causes:
- overheating
- faulty power supply
- overclocking
- in rare cases - outdated UEFI
Offline
This seems to be the issue
May 11 17:03:39 sybil kernel: NVRM: Xid (PCI:0000:06:00): 79, GPU has fallen off the busPossible causes:
- overheating
- faulty power supply
- overclocking
- in rare cases - outdated UEFI
I see. Overheating should be out of the option since it just crashed 2 times while only browser, discord, and a file browser was opened. Though I was looking into it, and for windows, I found 83 C to be a good temperature before it starts to throttle (using MSI Afterburner) but in nvidia-settings here, the slowdown temp is 95 C and hard temperature limit is 98. It also seems to be read-only, is there a way to turn it down?
- I'll order a new power supply, I also had thoughts about it.
- There is no overclocking
- I re-installed the whole of my arch like 4 days ago. So I think that should cover for outdated anything.
Are there any signs of GPU itself having faults?
Offline
Don't write off any possibilities. Damaged hardware could overheat without apparent reasons. May be add a widget to plasma panel that shows GPU temperature. Or some script that logs temperature into a file.
I re-installed the whole of my arch like 4 days ago. So I think that should cover for outdated anything.
UEFI is a motherboard firmware, something that used to be called BIOS. Reinstalling arch couldn't possibly help with it being out-to-date.
Are there any signs of GPU itself having faults?
Can't say after reading logs. But 'fallen off the bus' is a serious hardware-related error. So 'GPU itself having faults' is a possibility.
Last edited by Xephon (2025-05-12 15:02:23)
Offline
Don't write off any possibilities. Damaged hardware could overheat without apparent reasons. May be add a widget to plasma panel that shows GPU temperature. Or some script that logs temperature into a file.
I re-installed the whole of my arch like 4 days ago. So I think that should cover for outdated anything.
UEFI is a motherboard firmware, something that used to be called BIOS. Reinstalling arch couldn't possibly help with it being out-to-date.
Are there any signs of GPU itself having faults?
Can't say after reading logs. But 'fallen off the bus' is a serious hardware-related error. So 'GPU itself having faults' is a possibility.
I see. Thanks. I'll report back in a few days with more data.
Offline