High Scalability Multipole Method for the Analysis of Hundreds of Millions of Unknowns

Conference: EuCAP 2009 - 3rd European Conference on Antennas and Propagation
03/23/2009 - 03/27/2009 at Berlin, Germany

Proceedings: EuCAP 2009

Pages: 4Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Taboada, J. M.; Landesa, L. (Dept. Tecnologías de los Computadores y de las Comunicaciones, Escuela Politécnica, Universidad de Extremadura, 10071 Cáceres, Spain)
Bértolo, J. M.; Obelleiro, F.; Rodríguez, J. L. (Dept. Teoría de la Seal y de las Comunicaciones, E.T.S.I. Telecomunicación, Universidad de Vigo, 36200 Pontevedra, Spain)
Mouriño, J. C.; Gómez, A. (Supercomputing Center of Galicia, CESGA, Avda. de Vigo s/n 15705, Santiago de Compostela, Spain)

In this work, we present an efficient parallel implementation of the fast multipole method (FMM) combined with the fast Fourier transform (FFT). The good scaling propensity of the FMM-FFT, combined with a careful parallelization strategy, has shown to be very effective when using large parallel high performance supercomputers. A challenging problem with more than 0.5 billion unknowns has been solved. This is the largest problem analyzed in computational electromagnetics to date, which demonstrates that the proposed implementation of the FMM-FFT constitutes a real alternative to the more frequently used multilevel FMM algorithm (MLFMA).