A perl script that pulls data and does an MD5 hash or something similar may be a reasonable way to do this (assuming the size of the tables isn't overly large).

Another way (which only works if both servers are quiescent) is to bulk-copy all of the data to OS files and checksum the resulting files (be careful of row ordering, however). This won't tell you where the differences are, but it will tell you if the data is identical.

Michael