Benchmark of Acceleware vs XFdtd for Field Simulations of Microstrip Patch Antenna

We benchmark the accuracy and the speed of the Acceleware FDTD library vs XFdtd in simulating a microstrip patch antenna. The benchmark is based on the sampled electric and magnetic fields. The results show that Acceleware performs almost three times faster than XFdtd. The relative mean error is less than 2.0%.

I. Introduction

The finite-difference time-domain (FDTD) method is a well-established and widely used technique in solving electromagnetic (EM) problems [1]. Fast FDTD computations including those in many FDTD commercial solvers were achieved by utilizing graphics processing units (GPUs) [2]–[5]. However, it is necessary to verify the accuracy of the simulation of the electric (E) and magnetic (H) fields before fabricating real devices. In this work, the simulations on two established software packages are presented, Acceleware FDTD library and Remcom's XFdtd. The Acceleware FDTD library is a bundle of C/C++ functions that can be used to build applications to solve EM problems [6]. In contrast, Remcom's XFdtd is a full-featured EM simulation solver software package [7]. In this paper, the Acceleware FDTD library is benchmarked against XFdtd regarding the simulation of the electric (E) and magnetic (H) fields of a patch antenna. The hardware and implementation are described below.