Initialize all shader caches concurrently

[marco-carvalho]

No description provided.

[github-actions]

Download the artifacts for this pull request:

• ryujinx-Release-1.1.0+dcef21e-linux_arm64
• ryujinx-Release-1.1.0+dcef21e-linux_x64
• ryujinx-Release-1.1.0+dcef21e-macos_universal
• ryujinx-Release-1.1.0+dcef21e-win_x64

Old GUI (GTK3)

• gtk-ryujinx-Release-1.1.0+dcef21e-linux_arm64
• gtk-ryujinx-Release-1.1.0+dcef21e-linux_x64
• gtk-ryujinx-Release-1.1.0+dcef21e-win_x64

GUI-less (SDL2)

• sdl2-ryujinx-headless-Release-1.1.0+dcef21e-linux_arm64
• sdl2-ryujinx-headless-Release-1.1.0+dcef21e-linux_x64
• sdl2-ryujinx-headless-Release-1.1.0+dcef21e-macos_universal
• sdl2-ryujinx-headless-Release-1.1.0+dcef21e-win_x64

Only for Developers

• ryujinx-Debug-1.1.0+dcef21e-linux_arm64
• sdl2-ryujinx-headless-Debug-1.1.0+dcef21e-linux_arm64
• gtk-ryujinx-Debug-1.1.0+dcef21e-linux_arm64
• ryujinx-Debug-1.1.0+dcef21e-linux_x64
• sdl2-ryujinx-headless-Debug-1.1.0+dcef21e-linux_x64
• gtk-ryujinx-Debug-1.1.0+dcef21e-linux_x64
• ryujinx-Debug-1.1.0+dcef21e-macos_universal
• sdl2-ryujinx-headless-Debug-1.1.0+dcef21e-macos_universal
• ryujinx-Debug-1.1.0+dcef21e-win_x64
• sdl2-ryujinx-headless-Debug-1.1.0+dcef21e-win_x64
• gtk-ryujinx-Debug-1.1.0+dcef21e-win_x64

[tgcordell]

What’s the expected behavior change for the end user for this? Some type of performance improvement, but where? During initial game load?

[marco-carvalho]

The change should make shader cache initialization happen concurrently instead of one by one. This means the game should start up faster because it takes less time to get everything ready at the beginning. So, users should notice a quicker load time when they first launch the game.

[Foul-Tarnished]

The change should make shader cache initialization happen concurrently instead of one by one. This means the game should start up faster because it takes less time to get everything ready at the beginning. So, users should notice a quicker load time when they first launch the game.

Each game has only a single shader cache though ? so nothing to parallelize
Did you ask copilot to make ryujinx faster or what lmao

at least add some benchmarks, if it change anything