Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture

Finite-Difference Time-Domain (FDTD) method is a numerical method that can be used to solve electromagnetic problems in time domain. However, this method needs large computer memory and long execution time. Recently, a new scheme of FDTD has been develop which is called the High Speed Low Order (3)...

Full description

Saved in:
Bibliographic Details
Main Authors: Hasan, Mohammad Khatim, Othman, Mohamed, Md. Desa, Jalil, Abbas, Zulkifly, Sulaiman, Jumat
Format: Conference or Workshop Item
Language:English
Published: IEEE 2006
Online Access:http://psasir.upm.edu.my/id/eprint/59104/1/Parallel%20implementation%20for%20HSLO%283%29-FDTD%20with%20message%20passing%20interface%20on%20distributed%20memory%20architecture.pdf
http://psasir.upm.edu.my/id/eprint/59104/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.upm.eprints.59104
record_format eprints
spelling my.upm.eprints.591042018-02-22T07:11:56Z http://psasir.upm.edu.my/id/eprint/59104/ Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture Hasan, Mohammad Khatim Othman, Mohamed Md. Desa, Jalil Abbas, Zulkifly Sulaiman, Jumat Finite-Difference Time-Domain (FDTD) method is a numerical method that can be used to solve electromagnetic problems in time domain. However, this method needs large computer memory and long execution time. Recently, a new scheme of FDTD has been develop which is called the High Speed Low Order (3) FDTD (HSLO(3)-FDTD) method which reduced the speed of FDTD method by 67%. In this paper, we develop the parallel distributed version of the HSLO (3)-FDTD method on Distributed Memory Architecture Machine (Sunfire V1280) with message passing interface. We execute some numerical simulations by a new method with direct-domain (DD) approach to simulate one dimensional free space wave propagation represented by a Gaussian pulse. The simulation is conducted on 2 meter of solution domain truncated by a perfectly conducting boundary condition. The performance of the new schemes are analyzed and compared with the conventional parallel FDTD method in terms of processing time, speed-up, efficiency and cost. The new schemes is shown to be ultra-fast and very efficient with some superlinear speed-up. IEEE 2006 Conference or Workshop Item PeerReviewed text en http://psasir.upm.edu.my/id/eprint/59104/1/Parallel%20implementation%20for%20HSLO%283%29-FDTD%20with%20message%20passing%20interface%20on%20distributed%20memory%20architecture.pdf Hasan, Mohammad Khatim and Othman, Mohamed and Md. Desa, Jalil and Abbas, Zulkifly and Sulaiman, Jumat (2006) Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture. In: International Conference on Computing and Informatics (ICOCI 2006), 6-8 June 2006, Kuala Lumpur, Malaysia. . 10.1109/ICOCI.2006.5276531
institution Universiti Putra Malaysia
building UPM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Putra Malaysia
content_source UPM Institutional Repository
url_provider http://psasir.upm.edu.my/
language English
description Finite-Difference Time-Domain (FDTD) method is a numerical method that can be used to solve electromagnetic problems in time domain. However, this method needs large computer memory and long execution time. Recently, a new scheme of FDTD has been develop which is called the High Speed Low Order (3) FDTD (HSLO(3)-FDTD) method which reduced the speed of FDTD method by 67%. In this paper, we develop the parallel distributed version of the HSLO (3)-FDTD method on Distributed Memory Architecture Machine (Sunfire V1280) with message passing interface. We execute some numerical simulations by a new method with direct-domain (DD) approach to simulate one dimensional free space wave propagation represented by a Gaussian pulse. The simulation is conducted on 2 meter of solution domain truncated by a perfectly conducting boundary condition. The performance of the new schemes are analyzed and compared with the conventional parallel FDTD method in terms of processing time, speed-up, efficiency and cost. The new schemes is shown to be ultra-fast and very efficient with some superlinear speed-up.
format Conference or Workshop Item
author Hasan, Mohammad Khatim
Othman, Mohamed
Md. Desa, Jalil
Abbas, Zulkifly
Sulaiman, Jumat
spellingShingle Hasan, Mohammad Khatim
Othman, Mohamed
Md. Desa, Jalil
Abbas, Zulkifly
Sulaiman, Jumat
Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture
author_facet Hasan, Mohammad Khatim
Othman, Mohamed
Md. Desa, Jalil
Abbas, Zulkifly
Sulaiman, Jumat
author_sort Hasan, Mohammad Khatim
title Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture
title_short Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture
title_full Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture
title_fullStr Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture
title_full_unstemmed Parallel implementation for HSLO(3)-FDTD with message passing interface on distributed memory architecture
title_sort parallel implementation for hslo(3)-fdtd with message passing interface on distributed memory architecture
publisher IEEE
publishDate 2006
url http://psasir.upm.edu.my/id/eprint/59104/1/Parallel%20implementation%20for%20HSLO%283%29-FDTD%20with%20message%20passing%20interface%20on%20distributed%20memory%20architecture.pdf
http://psasir.upm.edu.my/id/eprint/59104/
_version_ 1643836980170063872
score 13.211869