An FE-BI-MLFMA with GPU acceleration for electromagnetic scattering analysis

Author

Jian Guan ; Su Yan ; Jian-Ming Jin

Author_Institution

Dept. of Electr. & Comput. Eng., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA

fYear

2014

fDate

6-11 July 2014

Firstpage

747

Lastpage

748

Abstract

A GPU-incorporated finite element-boundary integral algorithm with the acceleration of the multilevel fast multipole algorithm is presented for computing electromagnetic scattering from objects with complex structures and materials. The radar cross sections (RCS) are calculated to demonstrate the accuracy and efficiency of the proposed algorithm. Compared with the 8-threaded CPU-FE-BI-MLFMA (Intel Xeon W3520 @2.67GHz), this algorithm can achieve 27.3 times speedup in matrix assembly and 12.1 times speedup in matrix-vector product on a GeForce GTX 660 with 2GB memory.

Keywords

electrical engineering computing; electromagnetic wave scattering; finite element analysis; graphics processing units; integral equations; matrix algebra; radar cross-sections; vectors; FE-BI-MLFMA; GPU acceleration; GPU-incorporated finite element-boundary integral algorithm; GeForce GTX 660; Intel Xeon W3520; RCS; electromagnetic scattering analysis; frequency 2.67 GHz; matrix assembly; matrix-vector product; multilevel fast multipole algorithm; radar cross sections; Acceleration; Equations; Graphics processing units;

fLanguage

English

Publisher

ieee

Conference_Titel

Antennas and Propagation Society International Symposium (APSURSI), 2014 IEEE

Conference_Location

Memphis, TN

ISSN

1522-3965

Print_ISBN

978-1-4799-3538-3

Type

conf

DOI

10.1109/APS.2014.6904703

Filename

6904703