Introduction to GPU Computing and CUDA Programming: A Case Study on FDTD [EM Programmer´s Notebook]

Author

De Donno, Danilo ; Esposito, Alessandra ; Tarricone, Luciano ; Catarinucci, Luca

Author_Institution

Univ. of Salento, Lecce, Italy

Volume

52

Issue

3

fYear

2010

fDate

6/1/2010 12:00:00 AM

Firstpage

116

Lastpage

122

Abstract

The recent advent of general-purpose graphics-processing units (GPGPUs) as inexpensive arithmetic-processing units brings a relevant amount of computing power to modern desktop PCs. This thus providing an interesting pathway to the acceleration of several numerical electromagnetic methods. In this paper, we explain how to exploit GPGPU features by examining how the computational time of the Finite-Difference Time-Domain Method can be reduced. The attainable efficiency is demonstrated by providing numerical results achieved on a two-dimensional study of a human-antenna interaction problem.

Keywords

computer graphic equipment; coprocessors; finite difference time-domain analysis; parallel architectures; CUDA programming; GPU computing; compute unified device architecture; finite-difference time-domain method; general-purpose graphics-processing units; human-antenna interaction problem; inexpensive arithmetic-processing units; numerical electromagnetic methods; Finite difference methods; Graphics processing unit; Parallel processing; Parallel programming; Time domain analysis; Compute Unified Device Architecture (CUDA); FDTD methods; Parallel processing; general purpose graphics processing unit (GPU); parallel programming;

fLanguage

English

Journal_Title

Antennas and Propagation Magazine, IEEE

Publisher

ieee

ISSN

1045-9243

Type

jour

DOI

10.1109/MAP.2010.5586593

Filename

5586593