GPU Cluster Implementation of FMM-FFT for Large-Scale Electromagnetic Problems

Author

Vinh Dang ; Nguyen, Quang M. ; Kilic, Ozlem

Author_Institution

EECS, Catholic Univ. of America, Washington, DC, USA

Volume

13

fYear

2014

fDate

2014

Firstpage

1259

Lastpage

1262

Abstract

The fast multipole method (FMM) combined with fast Fourier transform (FFT) is investigated for the solution of large-scale electromagnetic problems, which require high computational capability that cannot be accommodated using conventional computing systems. The implementation is parallelized on a 13-node graphics processing unit (GPU) cluster that populates Nvidia Tesla M2090 GPUs. The experimental results based on our FMM-FFT implementation on GPUs demonstrate up to 957 times speedup compared to that of the single-core, single-node CPU implementation. The implementation details and the performance achievements in terms of accuracy, speedup, and scalability are discussed.

Keywords

computational electromagnetics; electromagnetic wave scattering; fast Fourier transforms; graphics processing units; FMM-FFT; GPU cluster; Nvidia Tesla M2090 GPUs; fast Fourier transform; fast multipole method; graphics processing unit; large-scale electromagnetic scattering problems; single-core single-node CPU; Antennas; Electromagnetics; Graphics processing units; Method of moments; Rough surfaces; Scalability; Surface roughness; Fast Fourier transform (FFT); fast nultipole method (FMM); graphics processing unit (GPU) clusters; iterative solvers; method of moments (MoM);

fLanguage

English

Journal_Title

Antennas and Wireless Propagation Letters, IEEE

Publisher

ieee

ISSN

1536-1225

Type

jour

DOI

10.1109/LAWP.2014.2332972

Filename

6843872