The GT635, despite having a ‘6’ might still be a Fermi device, so the correct driver is 390.
I’m getting fairly confused.
After some googling, apparently GT635 is Fermi. And Fermi doesn’t support Vulkan.
But GT635 is listed as supporting Vulkan, and I also can run applications that requires Vulkan on both on Linux and Windows.
So what does this mean?
Please run nvidia-bug-report.sh as root and attach the resulting nvidia-bug-report.log.gz file to your post.
Here’s the log:
nvidia-bug-report.log.gz (1.0 MB)
Ok, you have a Kepler based 635 so 470 is the correct driver. The problem with low-end GT6/7/8xx cards are that nvidia did a lot of relabeling so you never knew what exactly you were getting.
Since the driver is working fine, which DE/WM are you using?
That’s a very simple 2D WM without vsync, you might try if creating following config helps:
Section "OutputClass" Identifier "nvidia" MatchDriver "nvidia-drm" Driver "nvidia" Option "ForceCompositionPipeline" "true" EndSection
I’ve tried adding that config and rebooting, but it doesn’t work. I think I’ve also tried enabling that option in the past with the
nvidia-settings tool, but it didn’t help.
Note that I already have vsync enabled via picom: GitHub - yshui/picom: A lightweight compositor for X11
It fixes tearing issues that would normally occur in videos like this: Screen Tearing Test - YouTube
But it doesn’t seem to affect the artifacts I’m currently experiencing.
Looks like a known issue:
I’ve tried traversing some of the issues in the repository. But I can’t quite tell if those are the same issues I’m experiencing or not. I wish they would attach a video.
It seems that they’re experiencing slow downs and freezes when resizing windows, but in my experience resizing windows feels smooth. It’s just that I’m seeing trailing artifacts when the window redraws. Also, windows with videos playing tend to flicker when resized.
I’ve also tried dwm, and the effects were much less severe, but it’s still there. The GitHub issue linked to a thread where someone mentions that AwesomeWM doesn’t suffer from these issues, so I might test that one out next.