Library for arbitrary precision arithmetic, operating on signed integers, rational numbers and floating point numbers