Sounds like everyone has, so I'll chip in how I did it in *my* version :-)

I populate a hash with a filelist and sort based on size, so that I don't need to compare two files with the same filesize. But since I was comparing images from my webcam that were very small, there tends to be many files sized the same 87kb or whatever, so I still had do the hashing...

Oh yeah, and the first thing to do after getting all the filesizes was to get rid of the zero byte or corrupt files...