Arbitrary precision binary floating point arithmetic [1]