Interesting stuff. Not sure if I read this right that it‘s 16 und 32 bit values of integers that get sorted. If yes, I‘d love to see if the GPU implementation can beat a competitive Radix sort implementation on a CPU.
It's 32 32-bit values which get sorted. I don't think a GPU sort would beat a CPU sort at this scale, even if you don't take kernel launch time into account. CPUs are simply too fast for (super-)small data, especially with AVX-512.
But if we're talking about a larger amount of data, that would be a different story, i.e. as part of a normal gpu mergesort.
It is also useful if your data already lives on the GPU memory. For example, when you need to z-sort a bunch of particles in a 3d renderer particle system.
