64 bit algebra is not really required.

The double floats used by perl internally can represent 53 bit integers on most CPU architectures, enough to store the 12 decimal digits:

my $md5 = ...;

my $dd = 0;
$dd = ($dd * 256 + ord) % 1e12
  for split //, $md5;

print "$dd\n";
[download]

(actually, the biggest number that can appear in this algorithm is 2.56E+14, that can be represented in 48 bits)