When putting data into the MPI shared object store we should try to use aligned memory access. This should improve performance.
When putting data into the MPI shared object store we should try to use aligned memory access. This should improve performance.