What you live or die by when doing this is the speed of your crypt() implementation. The default ones are often horribly slow. If you're only doing 4000 crypts per second, you'll have to wait a long time for those codes. Tripper+ uses the UFC crypt() implementation, and it pushes about 64000 crypts per second per gigahertz on an Athlon processor. I've used some assembly-optimized crypt() code from John the Ripper to get up to 110000 crypts per second per gigahertz.

I don't know how hard it is to interface those from Python, though. I see there's a Perl wrapper for the UFC code, so maybe somebody made one for Python, too.