Aurora Perego

This refactoring aims to reduce code duplication and ease the process of adding new SYCL backends (e.g. `GpuSyclNvidia` and `GpuSyclAmd`). main changes: - removed the inheritance for the SYCL `Acc`,...

[14.0.X] Generalize warp size in pixel clustering kernel

9

#### PR description: In the `FindClus` kernel there was an assumption on the warp size being equal to 32, which is true for NVIDIA GPUs, but not always for AMD...

add warp mask functions

3

Now that oneAPI 2025.3 is out and the bug related to the opportunistic group has been fixed, I've rebased the branch and updated it. Main changes: - warp functions now...