NVIDIA / NVIDIA/open-gpu-kernel-modules

Extremely low performance of NVIDIA GPU vs iGPU

Open
#806 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

550.144.03

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

Ubuntu 24.04

Kernel Release

6.8.0-53-generic

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

RTX 4070 Super

Describe the bug

glxgears fps is extremely low for dGPU that is supposed to hardware capable.

Currently using hybrid setup where the monitor is connected to iGPU and dGPU is used on-demand. Directly using the dGPU by changing the monitor connection did not improve the performance.

This is the performance of AMD Radeon Graphics (radeonsi, raphael_mendocino, LLVM 18.1.7, DRM 3.58, 6.8.0-53-generic)

98503 frames in 5.0 seconds = 19700.402 FPS
102508 frames in 5.0 seconds = 20501.584 FPS
102417 frames in 5.0 seconds = 20483.322 FPS
101292 frames in 5.0 seconds = 20258.295 FPS
102314 frames in 5.0 seconds = 20462.562 FPS
101888 frames in 5.0 seconds = 20377.434 FPS

While this is for NVIDIA GeForce RTX 4070 SUPER

53038 frames in 5.0 seconds = 10607.499 FPS
56107 frames in 5.0 seconds = 11221.345 FPS
55948 frames in 5.0 seconds = 11189.562 FPS
53991 frames in 5.0 seconds = 10798.154 FPS
54808 frames in 5.0 seconds = 10961.540 FPS
56220 frames in 5.0 seconds = 11243.844 FPS

Notice, that dGPU is only half better than iGPU. The problem is this poor performance may also affect other applications.

To Reproduce
  1. Run: __NV_PRIME_RENDER_OFFLOAD=1 __GLX_VENDOR_LIBRARY_NAME=nvidia __GL_SYNC_TO_VBLANK=0 glxgears -info
  2. Run: vblank_mode=0 glxgears -info
  3. Compare
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

Start by reviewing nvidia-bug-report.log.gz and reproducing the two glxgears -info commands with the listed PRIME and vblank settings. Compare the NVIDIA offload and AMD iGPU results, including direct monitor connection behavior. Done means the performance discrepancy has a confirmed cause and a validated fix or clearly documented driver limitation.

Written by the indexing model from the issue text.

Assessment

Tech stack
c, linux, ubuntu
Domain
computer-graphics, operating-systems, performance
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.