libmultival is a C++ class library for multi-precision floating-point interval arithmetic. It uses libmplite to represent the interval bounds. (Source: http://freecode.com/)