For the best exploration of Perl's hash (and other) data structures I have seen, see Gisle Aas' PerlGuts Illustrated. Then perlguts, perlxstut, perlxs et al.

Another, possibly easier, method would be to embed a perl interpreter into the C-program and use that to interface to the hash.

Or, perhaps you could stick with the existing Perl script, but use Inline::C (or XS) to speed up the heavy processing, and leave interfacing to the shared hash in Perl.