Approximate calculation of 2D matrix entries via GPU

摘要

Minimizing memory access by converting a given matrix computation into a set of low-order polynomials. The set of polynomials is processed using parallel computational hardware such as graphical processing units.