Oooog. Boy is my face red. I was recalculating the main SHA1 each time through the loop. A minor refactoring mistake with a major impact...
With the base string hash pulled out of the loop, it's looking more like 1.7M SHA1 + HD calcs/core/sec.
Yeah, mine's Perl, and running on a slower processor, so the difference feels about right.
Of course my main calculation is only one line:
$hamming_dist = unpack("%160b*", sha1($attempt_phrase) ^ $challenge_sha);
:)
It's a brutally simple algorithm, really. XOR + bitcount, which only iterates for each 1 bit, so the lower the total hamming #, the faster it finds it.
It's in C, so everything works in unsigned long blocks. Compiled with -O3 it's:
real 0m0.968s
for 1M each on two forked processes, and 80 lines of code.