Skip to content

Windows

-> Note: For executables see Pyaket's FAQ

Q: Wrong GPU being used

Hybrid systems might offload lightweight tasks to the weaker or integrated gpu for performance, battery life, and efficiency reasons. Since ShaderFlow is a 'generic' Python application that simply uses moderngl contexts, the operating system is being lazy (in a good way) with its choice.

Using the integrated card in a hybrid system can cause crashes or poor performance.

  • Open the Nvidia Control Panel, go to Manage 3D settings and either:
    1. Find the Python interpreter or project executable being run
    2. Go for the Global Settings tab (discouraged)
  • Set the preferred graphics to the Nvidia GPU, apply settings.

Unknown, consider improving this answer if you have one!

Unknown, consider improving this answer if you have one!

On every scene initialization, it prints the renderer in use, a good output is eg.:

$ shaderflow basic main
β”‚0'00.936β”œβ”€info β”‚ Initializing scene Basic with backend WindowBackend.GLFW
β”‚0'01.186β”œβ”€info β”‚ OpenGL Renderer: NVIDIA GeForce RTX 3060/PCIe/SSE2
...

Q: Artifacts in uint16 textures

Seems to only happen in AMD cards, setting texture.anisotropy = 1 fixes it.

Related: DepthFlow#83 β€’ DepthFlow#84