NVIDIA / NVIDIA/open-gpu-kernel-modules

High power consumption in idle / unable to turn GPU off

Open
#382 13 comments 2 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

515.76

Does this happen with the proprietary driver (of the same version) as well?

Yes

Operating System and Version

Fedora 36

Kernel Release

5.19.12-200.fc36.x86_64

Hardware: GPU

NVIDIA GeForce RTX 3050 Ti Laptop GPU

Describe the bug

If Nvidia drivers are loaded powertop reports ~14W usage on idle while nvidia-smi shows "No running processes found"

If I blacklist nouveau, nvidia and remove the device using udev rules powertop reports ~7W usage on idle.

Note: From time to time it can boot with Nvidia drivers loaded and show ~7W usage on idle, this is rare but possible. If I wake up the GPU using nvidia-smi it will not go back to low power consumption again.

To Reproduce

Install nvidia drivers using this official fedora guide:
https://rpmfusion.org/Howto/NVIDIA

Bug Incidence

Always

nvidia-bug-report.log.gz

N/A

More Info

cat /proc/driver/nvidia/gpus/0000\:01\:00.0/power output:

Runtime D3 status:          Enabled (fine-grained)
Video Memory:               Active

GPU Hardware Support:
 Video Memory Self Refresh: Supported
 Video Memory Off:          Supported

Power Limits:
 Default:                   40000 milliwatts
 GPU Boost:                 40000 milliwatts

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

Reproduce the idle state on Fedora 36 with the NVIDIA driver, then compare powertop, nvidia-smi, and /proc/driver/nvidia/gpus/0000:01:00.0/power before and after waking the GPU. No source file or test is identified, so trace the relevant power-management entry points and use the reported low-power state as the completion criterion.

Written by the indexing model from the issue text.

Assessment

Tech stack
c, fedora, linux
Domain
operating-systems, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.