i don't know of anything that comes with the standard distribution, but you probably want to do a checksum instead of a diff. it will be much faster if you don't care at that point what the actual differences are. you can store the checksum when you download the file, so that you don't have to recompute it when you download the next one.