Frequent full system freeze while gaming on AMD

Issues & Assistance
,
1

Hello! I’m a fairly new Linux user, and I’ve been suffering from this problem for some weeks now, and no topic or article I searched seemed to help me out.

It happens especially when I’m playing Warframe, the screen turns a solid color, the last second of audio keeps repeating for a short time before stopping completely, and absolutely nothing works; keyboard LEDs don’t switch when I press their keys, no TTY, nothing, forcing me to hard reset through the power button.

I’ve tried multiple boot settings, Kernel versions and even different distros, still happened on Nobara and Mint (so, it is not a Wayland-exclusive problem).

journalctl shows no signs of significant errors, except for one time on Nobara, which I managed to enter TTY and reboot through the terminal, which gave me a kernel: amdgpu 0000:01:00.0: ring gfx timeout, but soft recovered .

My GPU and PSU are recently bought, my other components are less than five years old, so hardware issues seem unlikely to me. Temperatures are way below the danger zone, never really going over 75°C. I think it is important to note that I switched from Nvidia to AMD, and I switched from Mint to Cachy shortly (hours) before my Nvidia GPU died, and I never came across this issue using Nvidia on Mint, which I used for about a month.

The games which I got this issue were Warframe and Haste (the latter was minimized while I was watching a video).

Here’s my latest log since the most recent crash: 726d6f1

I’ll edit the post if I remember any important details.
This issue is already driving me insane, so, any help at all is greatly appreciated!

2

there is an error on parameter boot kernel

    0.032975] Unknown kernel command line parameters "splash pcie=pm=off", will be passed to user space.

something wrong with Rtkit and wireplumber that use it

jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: could not set nice-level to -11: Permission denied
jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: RTKit error: org.freedesktop.DBus.Error.ServiceUnknown
jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: RTKit does not give us MaxRealtimePriority, using 1
jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: RTKit error: org.freedesktop.DBus.Error.ServiceUnknown
jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: RTKit does not give us MinNiceLevel, using 0
jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: RTKit error: org.freedesktop.DBus.Error.ServiceUnknown
jun 28 02:19:51 <hostname-redacted> pipewire[1023]: mod.rt: RTKit does not give us RTTimeUSecMax, using -1
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: could not set nice-level to -11: Permission denied
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: RTKit error: org.freedesktop.DBus.Error.ServiceUnknown
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: RTKit does not give us MaxRealtimePriority, using 1
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: RTKit error: org.freedesktop.DBus.Error.ServiceUnknown
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: RTKit does not give us MinNiceLevel, using 0
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: RTKit error: org.freedesktop.DBus.Error.ServiceUnknown
jun 28 02:19:51 <hostname-redacted> wireplumber[1024]: mod.rt: RTKit does not give us RTTimeUSecMax, using -1

and many errors from kde & QDBusError

jun 28 01:54:03 <hostname-redacted> ksmserver[1105]: Failed to register with host portal QDBusError
...
3

I corrected the typo on the boot config, I was trying that parameter based on this thread, will have to check if anything changes later as I can’t do any testing right now. Thank you for pointing that out!

Edit: No luck with those settings, just happened again.

Edit 2: It seems disabling VRR/Freesync on my monitor, and keeping the boot parameters amdgpu.gpu_recovery=1 amdgpu.gfx_off=0 amdgpu.runpm=0 has stabilized my system for now, I will use it like this for some time before marking as a solution.

Edit 3: Crashed again after some hours. I’ll be moving to Windows to check if this problem persists there.

4

It crashed on Windows very quickly, no error logs at all.

I’m a bit lost now. Could this be a faulty gpu out of the box?

Edit: It seems my GPU came using way too much power, I decided to lower its TDP on LACT (105W → 70W), and it seems to have stabilized. Will give it some more test time before deciding. Dropping it to around 70W to 80W seems to have even improved the performance and dropped the temps by 10 to 15°C, so, I’m hopeful on this being the solution.

5

Lowering the power usage of the GPU did make it much more stable, but it’s not a perfect solution, the next crash only came after 5 days. I guess my GPU is faulty or something, since my PSU is very much capable of supporting it. I don’t think if I should mark this as a solution, but I can live with these more infrequent crashes until I can get a better GPU.