A Parallel Implementation of the Durand-Kerner Algorithm for Polynomial Root-Finding on GPU

Abstract

In this article we present a parallel implementation of the Durand-Kerner algorithm to find roots of polynomials of high degree on a GPU architecture (Graphics Processing Unit). We have implemented both a CPU version in and a GPU compatible version with CUDA. The main result of our work is a parallel implementation that is 10 times as fast as its sequential counterpart on a single CPU for high degree polynomials that is greater than about 48,000.