Phi Coprocessor Acceleration Techniques for Computational Electromagnetics Methods

Abstract

In this paper, we introduce the architecture of Intel Xeon Phi coprocessor, programming and acceleration techniques in the computational electromagnetic methods. This paper describes how to develop the acceleration codes based on the Intel Xeon® Phi coprocessors in the parallel format for general computational electromagnetics methods. We also present management of the cores and threads in the Intel Xeon Phi coprocessors to accelerate computational electromagnetics simulations. The Intel Phi coprocessor can be used as a regular CPU and shares the same source codes for Intel Xeon E3 and E5 CPUs with a different compilation option. The examples shown here are for acceleration of the parallel FDTD methods. The Intel Xeon Phi coprocessor is not a GPU and is a general hardware acceleration simulation platform.