Order is not guaranteed for per-region output, both for threaded and MPI output. Sort would require an extra kernel and some MPI reductions
Order is not guaranteed for per-region output, both for threaded and MPI output. Sort would require an extra kernel and some MPI reductions