> On Jul 30, 2015, at 9:08 PM, "li....@siat.ac.cn" <li....@siat.ac.cn> wrote: > > Although using parallel mesh, a MPI broadcast(mesh) is called after reading > mesh on a single processor, then each processor has a copy of the global > mesh, isn't it?
That is true for serial mesh formats. For meshes which are partitioned ahead of time (parallel ExodusII) each processor will read its own portion. -Ben ------------------------------------------------------------------------------ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
