Massively Parallel Fast Multipole Method Solutions of Large Electromagnetic Scattering Problems

We describe a massively parallel version of the single-level fast multipole method (FMM) that employs the fast Fourier transform (FFT) for the translation stage. The proposed FMM-FFT method alleviates the communication bottleneck and has a lower complexity, O(N 4/3 log 2/3 N), as compared to the con...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	IEEE transactions on antennas and propagation 2007-06, Vol.55 (6), p.1810-1816
Hauptverfasser:	Waltz, C., Sertel, K., Carr, M.A., Usner, B.C., Volakis, J.L.
Format:	Artikel
Sprache:	eng
Schlagworte:	Antennas Application software Applied classical electromagnetism Applied sciences Clustering algorithms Computational electromagnetics Concurrent computing Diffraction, scattering, reflection Distributed computing Distributed memory Electromagnetic scattering Electromagnetic wave propagation, radiowave propagation Electromagnetism electron and ion optics Exact sciences and technology fast Fourier transform (FFT) Fast Fourier transforms Fast methods fast multipole method (FMM) Fourier transforms Fundamental areas of phenomenology (including applications) Hardware Iterative algorithms Mathematical models Multicore processing Multilevel Multipoles parallel Physics Radiocommunications Radiowave propagation scalable Telecommunications Telecommunications and information theory Translations
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	We describe a massively parallel version of the single-level fast multipole method (FMM) that employs the fast Fourier transform (FFT) for the translation stage. The proposed FMM-FFT method alleviates the communication bottleneck and has a lower complexity, O(N 4/3 log 2/3 N), as compared to the conventional single level FMM which scales as O(N 3/2 ), where N is the number of unknowns. Through numerical examples we demonstrate that the proposed parallel fast multipole method yields a faster solution time than its multilevel counterpart for very large problems in a distributed memory parallel setting.
ISSN:	0018-926X 1558-2221
DOI:	10.1109/TAP.2007.898511