This guy has done some arbitrary precision floating point classes ... maybe something useful there. It's all template code, but other than that, it's ok. http://www.hvks.com/Numerical/arbitrary_precision.htm