Performance evaluation of Jacobi iterative solution for sparse linear equation system on multicore and manycore architectures

Abstract

One of the common and pressing challenges in solving real-world problems in various domains, such as in smart cities, involves solving large sparse systems of linear equations. Jacobi iterative method is used to solve such systems in case if they are diagonally dominant. This research focuses on the parallel implementation of the Jacobi method to solve large systems of diagonally dominant linear equations on conventional CPUs and Intel Xeon Phi co-processor. The performance is reported on the two architectures with a comparison in terms of the execution times.