Merge consecutive MPI_Allreduce operations into one

SGpp / DisCoTec

MPI-based code for distributed HPC simulations with the sparse grid combination technique. Docs->(https://discotec.readthedocs.io/)

https://sparsegrids.org/

GNU Lesser General Public License v3.0

8 stars 7 forks source link

Merge consecutive MPI_Allreduce operations into one #61

Closed cniethammer closed 1 year ago

cniethammer commented 1 year ago

MPI_Allreduce is an expensive operation. Therefore, merge consecutive MPI_Allreduce operations by combining single elements into a multi element buffer and extract the results afterwards.