I've started to learn Vulkan and your tutorials are phenomenal for a beginner like me. Thanks a lot for the great resource!
I've a question w.r.t descriptor indexing. See the following quote:
And while descriptor handling is still one of the most verbose parts, using descriptor indexing simplifies this significantly and makes it easier to scale. With that feature we can go for a "bindless" setup, where all textures are put into one large array and indexed in the shader rather than having to create and bind descriptor sets for each and every texture.
It mentions "where all textures are put into one large array", but is that really true?
- Is it the textures that are put into a large array or is it the descriptors?
- If it's textures, then how do big applications are able to dynamically add/remove textures into the array?
- What is meant by textures here? Are they images?
- Is this array continuous or scattered like
std::vector on the VRAM / Unified Memory?
I've started to learn Vulkan and your tutorials are phenomenal for a beginner like me. Thanks a lot for the great resource!
I've a question w.r.t descriptor indexing. See the following quote:
It mentions "where all textures are put into one large array", but is that really true?
std::vectoron the VRAM / Unified Memory?