The GNU Multiple Precision Library implements arbitrary precision arithmetic for integers, floating point numbers, and rationals\cite{granlund2014}. The MPFR library is based on GMP but implements IEEE-754 rounding\cite{fousse2007mpfr}.