Using double to store large integers is dubious; the largest integer that can be stored reliably in double is much smaller than DBL_MAX. You should use long long, and if that's not enough, you need your own arbitrary-precision code or an existing library.