I haven't looked too deeply into this, so this is just a heads up.
I have a util which currently uses md5sum to compare file contents, and in light of the recent vulnerabilities¹ I have decided to use _in addition_ another digest algorithm. So I compared the performance of the digests available in coreutils-5.2.1-31 (fedora core 3). sha1sum = .341s md5sum = .116s sum -s = .047s #SysV sum -r = .168s #BSD (default) cksum = .127s with those in openssl-0.9.7a-40 (man dgst) sha1 = .209s md5 = .110s md2 = 3.070s md4 = .135s sha = .318s rmd160 = .277s Now CRC doesn't seem to be that useful², so I would like to use sha1, but it can be seen above that the coreutils implementation is significantly slower when compared to openssl? cheers. ¹ http://www.cits.rub.de/MD5Collisions/ ² http://cryptography.hyperlink.cz/2004/otherformats.html -- Pádraig Brady - http://www.pixelbeat.org -- _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
