#962: parallelize Multipole using OpenMP and MPI ------------------------------------+--------------------------------------- Reporter: rhaas | Owner: Type: enhancement | Status: review Priority: minor | Milestone: Component: EinsteinToolkit thorn | Version: Resolution: | Keywords: Multipole ------------------------------------+---------------------------------------
Comment (by rhaas):
It contains threadprivate OMP pragmas for some memory that is allocated
by simpson integration rule which will mean that more memory is allocated than would be for the non-openmp case. The alternative would be to allocate and free the memory each time the function is called (this would be much cleaner).
I removed the threadprivate variables and checked that the tests pass before and after the change (and that hdf5 output does not change before and after).