Fix Kokkos reconstruct tiling and at no-amc optimization

STEllAR-GROUP / octotiger

Astrophysics program simulating the evolution of star systems based on the fast multipole method on adaptive Octrees

http://octotiger.stellar-group.org/

Boost Software License 1.0

48 stars 18 forks source link

Fix Kokkos reconstruct tiling and at no-amc optimization #406

Closed G-071 closed 2 years ago

G-071 commented 2 years ago

This PR provides a performance improvement for the Kokkos reconstruct kernel. The automatic tiling was suboptimal here. Hence we now use manual tiles with 64 workitems each. This lifts the kernel to the performance of its CUDA counterpart.