Use the algorithms in the C++ compiler's Math library to get up to 1e18 precision in the cdf function.