https://bugs.kde.org/show_bug.cgi?id=502224
--- Comment #6 from skrubpr...@gmail.com <skrubpr...@gmail.com> --- (In reply to Zamundaaa from comment #4) > That sequence of events sound a lot like a GPU reset, and the image and "X > connection to :0 broken (explicit kill or server shutdown)" matches too. > > You can check with > > sudo dmesg > after it happens (or journalctl -k --boot -1 a reboot afterwards) to make > sure, amdgpu prints about starting and completing the reset. If you can see > artifacts afterwards, it probably doesn't reset successfully though. > > If you're using shaders in Minecraft, try disabling them... they don't > always use the most well behaved OpenGL code (to put it very nicely). Yeah seems that shaders can cause problems. I was using complementary on CachyOS with kde and it was just not working out. Then I switched to bsl and had no issues until over time I had the same timeouts, and then I realized that the vram oc I had was creating instability. Now I am back on nobara without that dreaded overclock and have no issues for now. But I really do think this is more so the vram because it was also resetting the same way on other games. The reason I thought that it was minecraft at first was because it was the only game causing these timeouts, but I think I didn't give it enough time on other games to make the best conclusion. I realized I was being stubborn because on Windows I was able to push my vram to the highest amd adrenalin would let me set it (+150mhz) and have no problems, so I genuinely thought that the more conservative overclock I had on linux was fine. But I guess the behavior is different on here and I just thought I could have the same clocks and be fine. What I did now was change my power profile + reset vram clocks and that is somehow giving the same performance, even though the vram clock is at stock speeds. So far I have no issues but I am gonna give it more time to confirm this -- You are receiving this mail because: You are watching all bug changes.
