We can run rsync in `dry-run` mode and then run rsync in parallel against resulting files/dirs, might end up being faster. Just an idea.