A PARALLEL DIRECTIONAL FAST MULTIPOLE METHOD

被引:10
|
作者
Benson, Austin R. [1 ]
Poulson, Jack [2 ]
Tran, Kenneth [3 ]
Engquist, Bjoern [4 ,5 ]
Ying, Lexing [1 ,2 ]
机构
[1] Stanford Univ, ICME, Stanford, CA 94305 USA
[2] Stanford Univ, Dept Math, Stanford, CA 94305 USA
[3] Microsoft Corp, Redmond, WA 98052 USA
[4] Univ Texas Austin, Dept Math, Austin, TX 78712 USA
[5] Univ Texas Austin, ICES, Austin, TX 78712 USA
来源
SIAM JOURNAL ON SCIENTIFIC COMPUTING | 2014年 / 36卷 / 04期
基金
美国国家科学基金会;
关键词
parallel; fast multipole methods; N-body problems; scattering problems; Helmholtz equation; oscillatory kernels; directional; multilevel; ELECTROMAGNETIC SCATTERING; COLLECTIVE COMMUNICATION; INTEGRAL-EQUATIONS; ALGORITHM;
D O I
10.1137/130945569
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
This paper introduces a parallel directional fast multipole method (FMM) for solving N-body problems with highly oscillatory kernels, with a focus on the Helmholtz kernel in three dimensions. This class of oscillatory kernels requires a more restrictive low-rank criterion than that of the low-frequency regime, and thus effective parallelizations must adapt to the modified data dependencies. We propose a simple partition at a fixed level of the octree and show that, if the partitions are properly balanced between p processes, the overall runtime is essentially O N log N/p + p. By the structure of the low-rank criterion, we are able to avoid communication at the top of the octree. We demonstrate the effectiveness of our parallelization on several challenging models.
引用
收藏
页码:C335 / C352
页数:18
相关论文
共 50 条
  • [1] A PARALLEL VERSION OF THE FAST MULTIPOLE METHOD
    GREENGARD, L
    GROPP, WD
    [J]. COMPUTERS & MATHEMATICS WITH APPLICATIONS, 1990, 20 (07) : 63 - 71
  • [2] The parallel continuous fast multipole method.
    Johnson, BG
    Graham, RL
    [J]. ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY, 1997, 213 : 25 - PHYS
  • [3] APPLICATION OF FAST MULTIPOLE METHOD FOR PARALLEL MUFFLERS
    Wu, C-H
    Wang, C-N
    [J]. JOURNAL OF MECHANICS, 2012, 28 (01) : 153 - 162
  • [4] A parallel fast multipole method for elliptic difference equations
    Liska, Sebastian
    Colonius, Tim
    [J]. JOURNAL OF COMPUTATIONAL PHYSICS, 2014, 278 : 76 - 91
  • [5] Parallel implementation of fast multipole method based on JASMIN
    Cao XiaoLin
    Mo ZeYao
    Liu Xu
    Xu XiaoWen
    Zhang AiQing
    [J]. SCIENCE CHINA-INFORMATION SCIENCES, 2011, 54 (04) : 757 - 766
  • [6] Parallel implementation of fast multipole method based on JASMIN
    XiaoLin Cao
    ZeYao Mo
    Xu Liu
    XiaoWen Xu
    AiQing Zhang
    [J]. Science China Information Sciences, 2011, 54 : 757 - 766
  • [7] Parallel implementation of fast multipole method based on JASMIN
    CAO XiaoLin 1
    2 Laboratory of Computational Physics in Institute of Applied Physics and Computational Mathematics
    [J]. Science China(Information Sciences), 2011, 54 (04) : 757 - 766
  • [8] Parallel Fast Multipole Boundary Element Method for Crustal Dynamics
    Quevedo, Leonardo
    Morra, Gabriele
    Mueller, R. Dietmar
    [J]. 9TH WORLD CONGRESS ON COMPUTATIONAL MECHANICS AND 4TH ASIAN PACIFIC CONGRESS ON COMPUTATIONAL MECHANICS, 2010, 10
  • [9] A parallel implementation of the fast multipole method for Maxwell's equations
    Havé, P
    [J]. INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN FLUIDS, 2003, 43 (08) : 839 - 864
  • [10] Parallel Fast Multipole Method accelerated FFT on HPC clusters
    Mehta, Chahak
    Karthi, Amarnath
    Jetly, Vishrut
    Chaudhury, Bhaskar
    [J]. PARALLEL COMPUTING, 2021, 104