back
1 comments
In normal rasterization you can simply render your geometry with a single active shader (one vertex, one pixel etc). This is because you have direct control over when a specific type of geometry is rendered.

With raytracing you do not control when which geometry is rendered. Thus unless you want to simply use a single UberShader for the entire call to DispatchRays you need to link the geometry with the specific raytrace shader. (shader binding table https://developer.nvidia.com/sites/default/files/pictures/20...)