Tensor Calculus in Exact Real Computation

programming project in Exact Real Computation

Exact Real Computation is a new paradigm for reliable numerics and rapid numerical prototyping. By object-oriented overloading, a new abstract data type REAL takes care of all the hassles of classical floating-point arithmetic like rounding errors and cancellation. It has been implemented in the C++ library iRRAM [1].

This project evolves around implementing, empirically evaluating and assessing the practical performance of some of the algorithm that we and others have devised, analyzed, and proven efficient theoretically.