Improve OpenGL text drawing performance.

[unpacklo]

Description

Improve text drawing performance with OpenGL by using a static instance of QOpenGLPaintDevice. I still haven't narrowed down exactly what is causing the performance problem but I can see creating a new QOpenGLPaintDevice causes every drawText() call to create and destroy GL buffer objects.

Yields about a 2x speedup for my typical setup which draws lots of lines for the celestial sphere.

Fixes # (issue)

Screenshots (if appropriate):

Before, 10 FPS:

After, 20 FPS:

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• This change requires a documentation update

How Has This Been Tested?

Manually ran the change with a locally built stellarium and profiled with Instruments.

Test Configuration:

• Operating system: macOS 13.1
• Graphics Card: Radeon Pro 560X 4 GB
  Intel UHD Graphics 630 1536 MB

Checklist:

• My code follows the code style of this project.
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation (header file)
• I have updated the respective chapter in the Stellarium User Guide
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules

[github-actions]

Great PR! Please pay attention to the following items before merging:

Files matching src/**/*.cpp:

• Are possibly unused includes removed?

This is an automatically generated QA checklist based on modified files.

[github-actions]

Thanks for adding your first pull request to Stellarium. If you have questions, please do not hesitate to contact us.

[alex-w]

Thanks!

[10110111]

Even though this improves performance, this is still a hack, so it deserves a comment explaining why this object is static and what OS environment needs this.

[gzotti]

I tend to agree with @10110111 on this. The Qt docs of this class warns to be careful when interacting with the GL state. Can we guarantee that nothing can break when we leave a static object for re-use? (Or is it better to use an instance member variable in StelPainter and take care of this?)

[10110111]

Instance member will be worse because it will partially restore the performance dip, since StelPainter is created and destroyed multiple times per frame.

[gzotti]

Maybe then the StelPainter should be kept alive and handed around for a longer time?

[10110111]

Maybe then the StelPainter should be kept alive and handed around for a longer time?

It's not as easy, because StelPainter restores GL state on destruction. This would require a large refactoring.

[gzotti]

Yes, I know. But if creation&destruction, i.e., state switching&restoring, happen so often, of course this is a plausible cause of slowdown. If we can hand around a StelPainter for a longer time, it should save a lot of overhead.

[alex-w]

I got report about performance regression in Windows also :(

[10110111]

I got report about performance regression in Windows also

Does this PR improve the situation?

[alex-w]

Does this PR improve the situation?

• 1.2 release with equatorial grid - 17-35 fps
• 1.2 release with equatorial grid, no atmosphere, DSO, meridian, no ground - 17-20 fs
• the current branch with equatorial grid - 17-40 fps
• the current branch with equatorial grid, no atmosphere, DSO, meridian, no ground - 17-20 fs

macOS 13.2.1, MBP, M1

[alex-w]

@unpacklo could you share your config.ini file?

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.

[10110111]

The code that this PR changed has now been removed, hopefully the performance will be acceptable now. The PR itself is therefore obsolete.

[gzotti]

We can close this, thanks to all participants!