UoB-HPC / stdpar-nbody

https://research-information.bris.ac.uk/en/publications/efficient-tree-based-parallel-algorithms-for-n-body-simulations-u
MIT License
2 stars 0 forks source link

Fuse Hilbert-BVH CalculateMultipoles on HW with par-execution policy #36

Open gonzalobg opened 3 months ago

gonzalobg commented 3 months ago

CalculateMultipoles issues one kernel per level in sequence, which enables it to run in parallel on hardware that only supports par_unseq. On hardware that can run par in parallel, we could use a similar strategy to the Concurrent Octree with a per node counter, to fuse all these kernels into one.