Parallelization of MLFMA with Composite Load Partition Criteria and Asynchronous Communication
Keywords:
Parallelization of MLFMA with Composite Load Partition Criteria and Asynchronous CommunicationAbstract
This paper describes an efficient parallelization of the multi-level fast multipole algorithm (MLFMA) for fast solution of very large scale electromagnetic scattering problems. Computation in the MLFMA can be divided into several stages. To accomplish load balance at any time, load partition criteria are adjusted according to different features of every phase. Meanwhile, an asynchronous communication method is designed to overlap the communication with computation and thus the communication cost in parallelization is reduced. Numerical results show that good parallel efficiency is obtained in the presented parallelization of MLFMA. With our parallel MLFMA, a scattering problem with nearly 5, 300, 000 unknowns is solved in about six hours using 8 CPUs on SGI O350 server.
Downloads
References
J. Song level fast-
l., “Fast solution of
scat res by
loca lectric
polynomial representation for the
translation operators of an MLFMA,”
etics Society Journal, vol. 23,
criteria have been designed to achieve
balance at any time and an asynchronous
ommunication method has bc
reduce the communication cost. Numerical
results show the good parallel efficiency
achieved by the proposed parallel scheme.
With our PMLFMA, a scattering problem with
about 5.3 million unknowns has been solved,
demonstrating its ability in solving very large
scale problems. Our future work is to improve
the load partition for the translation phase in
coarse levels so that higher parallel efficiency
can be achieved.
ACKNOWLEDGEMENT
This work is supported partly by NSFC (No.
, research funding (No. 9140A0301
DZ0235, 9140A07030109DZ02), and the
Programme of Introducing Talents of
Discipline to Universities under Grant b07046.
and W. Chew, “Multi
multipole algorithm for solving combined
field integral equations of electromagnetic
scattering,” Microwave and Optical Tech.
Lett., vol. 10, no. 1, pp. 14-19, Sept. 1995.
W. Chew, T. Cui and J. Song, “A FAFFA-
MLFMA algorithm for electromagnetic
scattering,” IEEE Transactionas on
Antennas and Propagation, vol. 50, no. 11,
pp. 1641-1649, Nov. 2002.
T. Cui, W. Chew and G. Chen et al.,
“Efficient MLFMA, RPFMA, and FAFFA
algorithms for EM scattering by very large
structures,” IEEE Transactions on
Antennas and Propagation, vol. 52, no. 3,
pp. 759-770, Mar. 2004.
Hu and Z. Nie et a] J.
tering from conducting structu
l MLFMA based on improved e
field integral equation,” IEEE
Transactions on Electromagnetic
Compatibility, vol. 50, no. 4, pp. 940-945,
Nov. 2008.
S. Velamparambil and W. Chew, “A fast
Microwave and Optical Tech. Lett. , vol.
, no. 5, pp. 298-303, Mar. 2001.
H. Zhao, J. Hu and Z. Nie, “Block storing
method for efficient storage of near group
impedance in MLFMA,” Electronics
Letters, vol. 44, no. 20, pp. 1171-1173,
Sept. 2008.
S. Velamparambil, W. Chew and J. Song,
“10 million unknowns: is it that big?”
IEEE Antennas and Propag. Mag. vol. 45,
no. 2, pp. 43-58, Apr. 2003.
S. Velamparambil and W. Chew,
“Analysis and performance of a
distributed memory multilevel fast
multipole algorithm,” IEEE Transactions
on Antennas and Propagation vol. 53, no.
, pp. 2719-2727, Aug. 2005.
I. Gonzalez, E. Garcia, F. S. de Adana and
M. F. Catedra, “MONURBS: A
parallelized fast multipole multilevel code
for analyzing complex bodies modeled by
NURBS surfaces,” Applied Computational
Electromagn
no. 2, pp. 134-142, June 2008.
T. Iwashita, M. Shimasaki and J. Lu,
“Parallel ICCG solvers for a finite element
eddy-current analysis on heterogeneous
parallel computation environment,”
Applied Computational Electromagnetics
Society Journal , vol. 22, no. 2, pp. 195-
, July 2007.
W. R. Dearholt and S. P. Castillo,
“Electromagnetic scattering problems
utilizing a direct, parallel solver,” Applied
Computational Electromagnetics Society
Journal, vol. 22, no. 3, pp. 395-413, Nov.
R. F. Harrington, Field computation by
Moment Methods, Robert E. Krieger, FL,
J. Hu, Z. Nie and X. Gong, “Solving
electromagnetic scattering and radiation
by FMM with curvilinear RWG basis,”
Chinese Journal of Electronics, vol. 12,
no. 3, pp. 457-460, July 2003.
J. Dull, K. Gallivan, J. Song and W.
ACES JOURNAL, VOL. 25, NO. 2, FEBRUARY 2010
Chew, “Parallel fast multipole capacitance
solver,” IEEE Antennas and Propagation
Symp., Atlanta, June 1998.
