I would like to know if ComputeShader.SetBuffer has any overhead.
I have a ComputeBuffer that I want to share between kernels since its rather large and I will need to set the Buffer for each kernel that needs to modify it.
Setting it to multiple kernels seems to work fine but I am wondering if there are any gotchas or if there is a cost to setting the buffer to multiple kernels(like... does it have a completely separate copy in memory per kernel or does it just share the same memory between kernels).
↧