Home

This is a fast C++ implementation of adaptive precision arithmetics as described by Shewchuk, Jonathan Richard. "Adaptive precision floating-point arithmetic and fast robust geometric predicates." Discrete & Computational Geometry 18.3 (1997): 305-363.

It uses compile-time reasoning about the size of the expansions for maximum performance. A simpler version of the code is provided for OpenCL enabled devices which do not support C++ (yet). The code is compatible with Windows (compiles since Visual C++ 6.0 and upwards), Linux and OS X.

Project Members:

• Lukas Polok (admin)