/NetQD

.NET implementation of the double-double and quad-double technique for achieving almost 128-bit and 256-bit floating point precision types.

Primary LanguageC#MIT LicenseMIT

NetQD

.NET port of the QD library implementing the double-double and quad-double technique for achieving almost 128-bit and 256-bit floating point precision types.

See original paper by David H. Bailey Yozo Hida and Xiaoye S. Li for mathematical details. An unofficial copy of the original C++/Fortran implementation is available e.g. here.

Note that the port is in it's early stages so there may be some bugs.

Build Status

Installing

dotnet add package NetQD

Roadmap

  • Comprehensive set of unit tests that check that nothing has been broken during porting
  • Add more math functions (Exp, Log, Pow...)
  • Fit dotnet IFormatProvider into parsing and printing code

Contributing

Pull Requests are welcome.