Hyprland at 99% CPU usage #11411
Replies: 35 comments 48 replies
|
I'm not experienced with NixOS, but that seems unlikely to be the issue. I think it would be best to figure out where roughtly this issue stems from:
I will try to set up a NixOS VM tomorrow to check if this might be hardware related (which it shouldn't), but it might take some time :) |
|
anyone has a hyprland.log when this occurs, having hard time to reproduce it. so trying to figure out where in the code its going xwl selection code its going bork. |
|
I get a lot of these lines rapidly repeating: |
|
I believe I'm having the same issue on Arch Linux, although the CPU usage is less apparent for me. I checked htop and it was using 72% cpu for me, I notice the issue when my tmpfs for /run/user/1000 fills up. I then check my hyprland.log and it is something crazy like 7gb - |
|
I use clipse, and it runs as a user-service ('cause running it with exec-once does not work, somehow). |
|
Seemingly same issue here: at some point (often days and multiple suspend/resume) hyprland uses 100% cpu. Everything continues to work normally except for the cpu load (and the fans whizzing). It's spewing the Running garuda linux (arch) with: I'm on a Framework 13 Intel system. hyprctl systeminfo -c: Details``` Hyprland 0.50.1 built from branch at commit 4e242d0 e8dcc ([gha] Nix: update inputs). Date: Sat Jul 19 21:37:06 2025 Tag: v0.50.1, commits: 6291 built against: aquamarine 0.9.2 hyprlang 0.6.3 hyprutils 0.8.1 hyprcursor 0.1.12 hyprgraphics 0.1.5no flags were set System Information: GPU information: os-release: NAME="Garuda Linux" plugins: ======Config-Start====== |
|
xwm seems like it's X11 tho |
|
I can reproduce this error very reliably by copying large data inside a Xwayland program (in my case, until I exited Hyprland. From the logs this bug must be related to Xwayland. This is not observed in earlier versions. As I frequently copy images, this bug is quite annoying for me. |
|
In my case, killing the process |
|
I was experiencing the same issue for quite some time now, and I used the workaround mentioned by @pihomeserver and stopped using wl-past and wl-clip-persist for a few days. It worked so well that even the memory leak that I was experiencing with electron apps before is completely gone, thanks man lol |
|
@d-kja Glad that my "solution" have helped you. Now I still have the issue even when killing wp-paste. I can reproduce time to time the issue when I copy from xfreerdp and paste into a Brave browser window : it makes the browser crash and/or hyprland goes to 100% CPU load ... |
|
Seems like it happens to me also. It is reproducible a failed suspend triggers it (crappy nvidia driver). I need to disable nvidia driver after each kernel update (archlinux), but lately I'm unable to disable it, so at the moment I plug in my laptop all the time to prevent sleep, and hard switch it off when not in use. Major pita. |
|
Happening to me too. Not working. |
|
hyprwm/aquamarine#224 might solve it when it's ready |
|
Same for me. I think it only ever happens when I'm running a VM (virt-manager, qemu), notably with GDK_BACKEND set to x11. I'm not sure if this will be useful, but If I run strace then there are a ton of epoll_wait calls, something like Call counts from an strace sample looks like: |
|
Same issue for me, But I can reproduce it and stop it: Inside xfreerdp3 I press CTRL+Insert (not each time but almost often) and Hyprland start using 100% of one core, as soon as I press CTRL+C inside rdp it become normal! |
|
Same here for me. It seems to happen everytime I try to paste something inside pureref... |
|
It should be fixed in version 0.54. Please test it. |
|
Same issue here after copy and pasting... I've been having this bug for a long time now |
|
Just found this, got back from vacation after a week and updated today, so sometime in the last week or two of updates from master branch this started happening, hyprland is pegging one cpu core to max, makes hyprland unusable. |
|
I found a workaround for xfreerdp3 - just use sdl-freerdp3. like: sdl-freerdp3 /v:192.168.128.111 /u:mylogin /p:$RDP_PASSWORD /cert:ignore /dynamic-resolution /scale-desktop:125 -grab-keyboard /clipboard:direction-to:all /drive:media,/home/alex/Windows/Work/temp |
|
I found a bug with HDR that was causing this CPU core pinning, opening a PR (or having a contributor) open one soon, its a one liner |

















Uh oh!
There was an error while loading. Please reload this page.
Not really a crash since Hyprland continues to work but something is definitely going wrong.
For some reason, Hyprland uses 99-100% CPU on one core. I have no idea when this started happening, however it has been happening for a few weeks now. This is annoying as the laptop (Framework Laptop 13 with AMD Ryzen 5 7640U) heats up to 99C and the fans ramp up loudly.
I have my configuration in a git repo here: https://code.liv.town/liv/nixos-config/src/branch/master/modules/home/hyprland.
It seems there are no other Hyprland processes that could be causing issues:
Tree view of the processes under Hyprland:
Output for
sensors:I have no clue how to debug this, so please help me a bit here. Do you need any other information?
All reactions