Arbitrary precision floats

The Float type is an arbitrary precision float. It supports correctly rounded computation of arithmetic (+, -, *, /), square roots and precision conversions. It makes no claims to high performance, but does try to keep it in mind.

Built on ramp, and implemented with a lot of reference to [1].

[1]: Muller, Jean-Michel; Brisebarre, Nicolas; de Dinechin, Florent; Jeannerod, Claude-Pierre; Lefèvre, Vincent; Melquiond, Guillaume; Revol, Nathalie; Stehlé, Damien; Torres, Serge (2010). Handbook of Floating-Point Arithmetic. Birkhäuser. ISBN 978-0-8176-4705-6.