Currently we look for MPI ghosts only using the bounding boxes of the domains. It would be much more optimal (i.e. it would create less unnecessary ghosts) to use only the active cells of the pruned tree. Note that this would become very similar to the exporting of particles...
Currently we look for MPI ghosts only using the bounding boxes of the domains. It would be much more optimal (i.e. it would create less unnecessary ghosts) to use only the active cells of the pruned tree. Note that this would become very similar to the exporting of particles...