NVIDIA / NVIDIA/open-gpu-kernel-modules

RTX 4090 (open module 610.43.03): GPU stays in P8 / 405 MHz memory clock during Wayland compositing → intermittent desktop stutter at steady FPS

Open
#1,251 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug
Dominant language
C
Stars
17.4k
Forks
1.9k
PR merge metrics
No merged PRs in 30d

Description

NVIDIA Open GPU Kernel Modules Version

610.43.03

Please confirm this issue does not happen with the proprietary driver (of the same version). This issue tracker is only for bugs specific to the open kernel driver.
  • I confirm that this does not happen with the proprietary driver package.
Operating System and Version

Fedora Linux 44 (KDE Plasma Desktop Edition)

Kernel Release

Linux rswann-desktop 7.1.4-200.fc44.x86_64

Please confirm you are running a stable release kernel (e.g. not a -rc). We do not accept bug reports for unreleased kernels.
  • I am running on a stable kernel release.
Hardware: GPU

NVIDIA GeForce RTX 4090

Describe the bug

All desktop animations (virtual-desktop switch/slide, dragging a window) stutter intermittently. Crucially this happens at a steady frame rate — KWin's built-in FPS overlay reads a stable 98–105 FPS (cap 100) throughout, with no drops. So it's a frame-pacing problem, not throughput.

Root cause appears to be GPU power management: under sustained desktop compositing the GPU never raises its clocks. I sampled nvidia-smi clocks at 100 ms resolution for ~50 s while continuously dragging a window, and used kwin_wayland's CPU time (from /proc//stat deltas) as an independent witness that real compositing work was happening:

  • kwin_wayland CPU: 17–37% of one core for 113 of 144 samples (definitely under load)
  • GPU during that entire window: P8, 210 MHz core, 405 MHz memory, util 0–7%, ~8 W — 144/144 samples, zero transitions.

At the 405 MHz memory-clock floor, the occasional heavier compositor frame misses the 10 ms (100 Hz) deadline → the intermittent hitch. GPU utilization stays low enough (0–7%) that the clock governor never ramps, so the GPU sits at its idle memory clock indefinitely even while actively compositing.

To Reproduce
  1. KDE Plasma 6.7.3 Wayland on the open 610.43.03 module, RTX 4090, high-refresh display.
  2. Continuously drag a window or rapidly switch virtual desktops.
  3. Observe intermittent stutter; confirm with nvidia-smi -lms 100 --query-gpu=pstate,clocks.gr,clocks.mem,utilization.gpu --format=csv that the GPU stays at P8 / 405 MHz memory throughout.

Forcing the clocks up eliminates the stutter completely:
sudo nvidia-smi -pm 1
sudo nvidia-smi -lmc 10501

Bug Incidence

Always

nvidia-bug-report.log.gz

nvidia-bug-report.log.gz

More Info

No response

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No source file or test is named. Start by reproducing the Wayland/KWin stutter with nvidia-smi sampling and review the attached nvidia-bug-report.log.gz; done would require identifying and correcting the open-module clock-management behavior, then validating compositor load without forced clocks.

Written by the indexing model from the issue text.

Assessment

Tech stack
c, linux
Domain
desktop-dev, operating-systems, performance
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.