Hey :)

basically i can't figure out how to use a [u8; 4] instead of a [f32; 4] in the vertex struct, since i can't figure out where to specify the attribute's format

I have the same problem, did you find a solution @ChevyRay ?

@OniDaito Hi!

Sorry for delay with the answer.

Yes, Runtime shader example is one of the options. Current primary use-case of Vulkano is to use precompiled shaders(built in compile time through the shader! macros), but there is a way to load them manually too in which case one will have to configure all descriptors and layouts manually(which is unsafe).

On general note, Vulkano is following a principle(but is not enforces it too strongly) that all shaders normally should be loaded in the beginning and be prepared beforehand. In this case the end-user can gain maximal benefits of GPU capabilities through Vulkan API. GPUs themselves are not very tolerate to shader programs reloading, as it breaks the entire pipeline and all related stuff set up by the device on pipeline creation. Basically GPUs are designed for static programs. That's why Vulkano in turn trying to follow the same approach.

This is, however, not always possible, and there could be edge cases when the API user might want to generate shader's code on the fly. One of the possibility as I mentioned is to use the approach from Runtime shader example. But it could be difficult to maintain. Another approach(a limited one, though) is to use Specialization Constants. The API user in fact can prepare large code of the shader that would specialize depending on Specialization Constants values. One can even organize branching depending on the values of constants, the shader loader will optimizes particular specialization statically and eliminate dead code control flows not covered by particular specilization.

Just as an idea for you. Please take a look at the specialization constants examples in repository for more details.

@OniDaito Hi!

Sorry for delay with the answer.

Yes, Runtime shader example is one of the options. Current primary use-case of Vulkano is to use precompiled shaders(built in compile time through the shader! macros), but there is a way to load them manually too in which case one will have to configure all descriptors and layouts manually(which is unsafe).

On general note, Vulkano is following a principle(but is not enforces it too strongly) that all shaders normally should be loaded in the beginning and be prepared beforehand. In this case the end-user can gain maximal benefits of GPU capabilities through Vulkan API. GPUs themselves are not very tolerate to shader programs reloading, as it breaks the entire pipeline and all related stuff set up by the device on pipeline creation. Basically GPUs are designed for static programs. That's why Vulkano in turn trying to follow the same approach.

This is, however, not always possible, and there could be edge cases when the API user might want to generate shader's code on the fly. One of the possibility as I mentioned is to use the approach from Runtime shader example. But it could be difficult to maintain. Another approach(a limited one, though) is to use Specialization Constants. The API user in fact can prepare large code of the shader that would specialize depending on Specialization Constants values. One can even organize branching depending on the values of constants, the shader loader will optimizes particular specialization statically and eliminate dead code control flows not covered by particular specilization.

Just as an idea for you. Please take a look at the specialization constants examples in repository for more details.

@OniDaito, @shiraori:matrix.org , @derivator

Dear fellows,

I have to apologize I'm not able to maintain the chat regularly. I'm visiting it rarely and noticing new topics came days if not weeks ago. This is my bad, but I just don't have enough resources to response on time. Perhaps, it would be more beneficial for you and for the community if you raise questions and topics in Vulkano repository Issues section. Me and other regular users are watching it more regularly.