• DocumentCode
    2058635
  • Title

    An efficient solution for the parallelization of the multilevel fast multipole algorithm

  • Author

    Qiao, Tianzhu ; Zhang, Jian F. ; Cui, Tie Jun

  • Author_Institution
    State Key Lab. of Millimeter Waves, Southeast Univ., Nanjing, China
  • fYear
    2013
  • fDate
    23-26 Sept. 2013
  • Firstpage
    175
  • Lastpage
    178
  • Abstract
    In this paper, an implementation of parallel MLFMA for scattering problems is presented. MLFMA is an efficient algorithm for accelerating the calculation of matrix-vector-multiplication. Using the hybrid strategy, the parallel version can be efficiently realized on distributed memory systems. Tree construction, communication, etc., are the key techniques which are discussed in details here. Numerical results show the efficiency of the MLFMA is improved by the parallelization.
  • Keywords
    distributed memory systems; electromagnetic wave scattering; parallel algorithms; trees (mathematics); distributed memory systems; hybrid strategy; matrix-vector-multiplication; multilevel fast multipole algorithm; parallel MLFMA; scattering problems; tree construction; Algorithm design and analysis; Antennas; Interpolation; MLFMA; Memory management; Program processors; Scattering;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Direct and Inverse Problems of Electromagnetic and Acoustic Wave Theory (DIPED), 2013 XVIIIth International Seminar/Workshop on
  • Conference_Location
    Lviv
  • ISSN
    2165-3585
  • Print_ISBN
    978-966-02-6765-7
  • Type

    conf

  • Filename
    6653860