I’ve been having consistent problems with some games, where their window will unpredictably freeze, then show me an image from a few seconds before it stopped responding. Everything else works fine, though my computer considers the Steam window to be a part of the game as well; as in, the custom cursor from the game appears over the Steam window and I cannot interact with it in any way.
I suspected it was a problem with using the Integrated GPU, so I did some searching, installed envycontrol and ran sudo envycontrol -s nvidia --force-comp, which didn’t work, then found the prime-run command, which, when added as a launch option to the game in my Steam library, allowed my game to last longer, but it still eventually failed.
I managed to catch this screenshot, however, showing the PC splitting the work between both GPUs before deciding at some point to dump everything on GPU1 (which, although I believe it is my integrated GPU, I do not have a way to confirm this) (the freeze happens roughly when GPU1 reaches 100% usage):
OpenGL renderer string: NVIDIA GeForce RTX 3070 Laptop GPU/PCIe/SSE2
X Error of failed request: BadWindow (invalid Window parameter)
Major opcode of failed request: 146 ()
Minor opcode of failed request: 5
Resource id in failed request: 0x6000003
Serial number of failed request: 55
Current serial number in output stream: 56
My guess is this could be related to (X)Wayland somehow. You could try this by installing kwin-x11 and log into an X11 session, or by installing icewm.
I installed kwin-x11, but I’m not sure how to use it, and I can’t find any guides online.
I tried running the --desktopfile command and got:
kwin: unable to claim manager selection, another wm running? (try using --replace)
QThreadStorage: entry 1 destroyed before end of thread 0x55ee117dd7d0
QThreadStorage: entry 0 destroyed before end of thread 0x55ee117dd7d0
So I ran kwin_x11 --replace and received:
qt.qpa.services: Failed to register with host portal QDBusError(“org.freedesktop.portal.Error.Failed”, “Could not register app ID: App info not found for ‘org.kde.kwin’”)
kwin: another window manager is running (try using --replace)
QThreadStorage: entry 7 destroyed before end of thread 0x55a7354277d0
QThreadStorage: entry 1 destroyed before end of thread 0x55a7354277d0
QThreadStorage: entry 0 destroyed before end of thread 0x55a7354277d0
I found this post and followed its advice, but all that accomplished was hiding the first two lines of output. It still recommends running --replace to the --replace command…
When you say log out, do you mean of my account on my computer? I didn’t receive a choice for anything after doing so.
I set my auto-login in System Settings > Login Screen SDDM > Behavior to use a Plasma(X11) session, then logged out and logged back in, but the command echo "$XDG_SESSION_TYPE" still printed
wayland
Okay. I reset my computer and that seems to have done it. The session is now x11, and when I run glxinfo | grep "OpenGL renderer", it prints out my NVIDIA GPU.
I’m going to test if this actually fixes the issue or not, but my thinking would be the problem is different from what I assumed if it still occurs.
Edit: It seems to work now. Surprisingly, GPU1 is the NVIDIA one, so I’m not sure what was going on previously, but hopefully I do not have to worry about that.
Okay, I was mistaken. It might be an issue related to overheating, as it didn’t start to happen until my computer became noticeably hot (which is consistent with the circumstances that caused me to make the post in the first place), but according to inxi -zv8, my GPU is only 66 degrees C, and my fans are still working fine, so… I don’t know.
Something that was different is that my game closed seconds after freezing, rather than staying on screen until I forced it to stop like I had to before.
Edit: Okay, X11 is just worse. Second freeze made my entire laptop unusable even after the game closed, forcing me to manually reset it. I’m going to uninstall and switch back to wayland.
X11 mostly gives better gaming performance, specially on nvidia GPU’s. Probably your hardware utilise better on X11 as on Wayland and when your hardware is overheating the crash cascade then even stronger.
Are you aware that we also have the magic system hotkeys? As long you get have a Kernel Crash, you can revive your system, or at least you can initiate a clean shutdown.
For investigating closer around your problem, you might want to temporary disable your i-gpu in your Bios to find the culprit… its not always a easy task to find a solution.
You can also try out different Kernels 6.18 LTS for example.
Did you created this system info, while your game was running?
I ask this because for idle CPU temps, 66 degree is really hot.
I was not playing anything at the time of posting my system info.
Although, that’s also the result of spending most of the day on my laptop. Grabbing the temp rn, the resting temperature seems to be about 42 degrees C.
I don’t know? I install updates as they arrive. I was under the impression I had the latest versions of everything, and I only moved to Manjaro a year ago.
Found Manjaro Settings Manager; I didn’t know about this until now. Thanks for the advice.
Can’t find a way to install the .exe updates from HP, though.
Okay, I was wrong. The idle CPU temperature does seem to be around 60 degrees C, with just LibreOffice Writer and Firefox running. I can’t do anything with my fans, though, because I have “no pwm-capable sensor modules installed.”
I’m going to attempt to use Hiren’s Boot CD as soon as I can, but that’ll be a minimum of two weeks from now. Any suggestions until then?
Open your Laptop, clean the dust and depends how much dust can be removed or not.
I would thinking about to repaste your Heatpaste then, if you can’t adjust the fanspeed in your Bios or with a Software Tool.
Idk how many reminder’s you need to change your Kernel
I was just about to suggest this. I have seen a similar report elsewhere (unfortunately I can’t find it now) but when the card was overheating, it would fail causing the kernel to drop the module or perhaps the card just failed to communicate any longer; I can’t quite remember the specifics.
Definitely add some heat monitoring (preferably with logging), remove any overclocks if you have any.
Also you don’t seem to have mentioned or provided any info from your syslogs? That should at least narrow down the issue.
Edit: Although, I had powered off my computer 2 hours ago and it was about room temperature when I turned it back on (I was moving it around). So, it’s probably not related to length of time I’ve been using it?
Edit 2: I shut down the Steam app and a bit later the Cores all dropped by 40 degrees C. Not sure how to interpret that. I had only been playing Rocket League with Discord open when the 90’s were recorded.
