Subject: | Missing the fastest: crc32 |
Compress::Zlib contains the fast (SSE2 enabled) version for crc32
which blows away all other digests.
1 CPU cycle (+ 3 wait-cycles if not in parallel) per word in modern Intel or ARM CPU's.
zlib uses Adler's parallel version. There's nothing faster.
Note that the other crc32 versions on CPAN, like String::CRC32 and Digest::CRC32 are not recommended.