These functions sample multiple RGBA pixels but return only a single component. OpenGL: https://www.opengl.org/sdk/docs/man/html/textureGather.xhtml DirectX: https://msdn.microsoft.com/en-us/library/windows/desktop/ff471525(v=vs.85).aspx
If you were to use this on an RGBA texture, and only gathering the red component, would the texture cache get loaded with the RGBA values, or would it only load the R values?
I'm curious because not all cards in use today support A8 textures (or other single channel 8 bit textures), and a workaround would be to store the data in 1/4 as many pixels by using RGBA and only grabbing the channel needed.
If this sounds obscure, a down to earth practical usage case is distance field textures.
Trying to see if using these functions would be a good alternative to cards that don't support A8.