Exactly! I don't see it as a major limitation since the mesh will be equally distributed on each processor.