Optimization and parameter exploration using GPU based FDTD solvers

Author

Inman, Matthew J. ; Elsherbeni, Atef Z.

Author_Institution

Department of Electrical Engineering, University of Mississippi, 38677-1848, USA

fYear

2008

fDate

15-20 June 2008

Firstpage

149

Lastpage

152

Abstract

Graphical processing units (GPU) has been documented for the implementation of the FDTD technique. The use of these specialized processors for the implementation of numerical codes has been shown to significantly speed up the execution of these codes over standard CPU based solvers. With the execution of the FDTD method being reduced to a matter of seconds, the use of these codes for optimization and parameter exploration becomes possible. This paper detail the use of a developed GPU based FDTD solver for both optimization and through a graphical interface parameter exploration.

Keywords

coprocessors; finite difference time-domain analysis; mathematics computing; optimisation; FDTD solvers; graphical interface parameter exploration; graphical processing units; numerical codes; optimization solver; Central Processing Unit; Code standards; Computer aided manufacturing; Computer languages; Equations; Finite difference methods; Graphics; Kernel; Libraries; Time domain analysis; FDTD; GPU; Optimization;

fLanguage

English

Publisher

ieee

Conference_Titel

Microwave Symposium Digest, 2008 IEEE MTT-S International

Conference_Location

Atlanta, GA

ISSN

0149-645X

Print_ISBN

978-1-4244-1780-3

Electronic_ISBN

0149-645X

Type

conf

DOI

10.1109/MWSYM.2008.4633125

Filename

4633125