In the current version of this library,
the basic multiprecision calculations are implemented in i386+ assembly
language. Multplication is performed recursively resulting in
O(lg 3) performance. Modular multiplication can be
performed using algorithms given in
Montgomery (1985).