Advertisement

Parallelization Algorithms for Three-Body Interactions in Molecular Dynamics Simulation

  • Jianhui Li
  • Zhongwu Zhou
  • Richard J. Sadus
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4330)

Abstract

Two force decomposition algorithms are proposed for parallelizing three-body interactions in Molecular Dynamics (MD) simulations. The first algorithm divides the entire 3D force matrix into equal sized force cubes that are assigned to parallel processors. In the second strategy, the force matrix is decomposed into slices of two-dimensional force matrixes, and those slices are distributed among processors cyclically. The proposed decomposition algorithms are implemented using MPI and tested in computational experiments. The performances of proposed decomposition methods are studied and compared with computational load theoretical analysis. Both theoretical prediction and computation experiments demonstrate that the load balance is a key factor that impacts the parallel performance of the examined system, and the cyclic force decomposition algorithm produced reasonably good overall performances.

Keywords

Load Balance Parallel Performance Decomposition Algorithm Load Imbalance Force Element 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Bosko, J.T., Todd, B.D., Sadus, R.J.: Journal of Chemical Physics, 121, 1091–1096 (2004)Google Scholar
  2. 2.
    Rentsch, R., Brinksmeier, E., Li, J.: Nonlinear Dynamics of Production Systems, pp. 245–263. Wiley-VCH, Chichester (2003)Google Scholar
  3. 3.
    Plimpton, S.: Journal of Computational Physics 117, 1–19 (1995)Google Scholar
  4. 4.
    Roy, S., Jin, R.Y., Chaudhary, V., Hase, W.L.: Computer Physics Communications 128, 210–218 (2000)Google Scholar
  5. 5.
    Schreiber, H., Steinhauser, O., Schuster, P.: Parallel Computing 18, 557–573 (1992)Google Scholar
  6. 6.
    Smith, W.: Computer Physics Communications 62, 229–248 (1991)Google Scholar
  7. 7.
    Boyer, L.L., Pawley, G.S.: Journal of Computational Physics 78, 405–423 (1988)Google Scholar
  8. 8.
    Brunet, J.P., Edelman, A., Mesirov, J.P.: SIAM Journal of Scientific and Statistical Computing 14(5), 1143–1158 (1993)MATHCrossRefMathSciNetGoogle Scholar
  9. 9.
    Plimpton, S.: Journal of Computational Chemistry 17(3), 326–337 (1996)CrossRefGoogle Scholar
  10. 10.
    Fincham, D.: Molecular Simulation, vol. 1, pp. 1–45 (1987)Google Scholar
  11. 11.
    Gupta, S.: Computer Physics Communications, vol. 70, pp. 243–270 (1992)Google Scholar
  12. 12.
    Matsuoka, O., Clement, E., Yoshimine, M.: Journal of Chemical Physics 64(4), 1351–1361 (1976)CrossRefGoogle Scholar
  13. 13.
    Marcelli, G., Sadus, R.J.: Journal of Chemical Physics 111, 1533–1540 (1999)CrossRefGoogle Scholar
  14. 14.
    Li, J., Zhou, Z., Sadus, R.J.: Modified Force Decomposition Strategies for Three-Body Interactions in Molecular Dynamics Simulations. Computer Physics Communications (to be published, 2006)Google Scholar
  15. 15.
    Axilrod, B.M., Teller, E.: Journal of Chemical Physics 11(6), 299–300 (1943)CrossRefGoogle Scholar
  16. 16.
    Sadus, R.J.: Molecular Simulation of Fluids, Theory, Algorithms and Object-Orientation. Elsevier, Amsterdam (2002)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Jianhui Li
    • 1
  • Zhongwu Zhou
    • 1
  • Richard J. Sadus
    • 1
  1. 1.Centre for Molecular SimulationSwinburne University of TechnologyHawthornAustralia

Personalised recommendations