The Parallelization of the Princeton Ocean Model

  • L. A. Boukas
  • N. Th. Mimikou
  • N. M. Missirlis⋆
  • G. L. Mellor
  • A. Lascaratos
  • G. Korres
Conference paper

DOI: 10.1007/3-540-48311-X_196

Part of the Lecture Notes in Computer Science book series (LNCS, volume 1685)
Cite this paper as:
Boukas L.A., Mimikou N.T., Missirlis⋆ N.M., Mellor G.L., Lascaratos A., Korres G. (1999) The Parallelization of the Princeton Ocean Model. In: Amestoy P. et al. (eds) Euro-Par’99 Parallel Processing. Euro-Par 1999. Lecture Notes in Computer Science, vol 1685. Springer, Berlin, Heidelberg

Abstract

In this paper we present the parallel implementation of the Princeton Ocean Model (POM) using message passing. Domain decomposition techniques are used for the horizontal discretization whereas in the vertical direction each column per grid point is computed. The required interprocessor communication was implemented on the PVM message passing library. Three different data partitioning schemes are studied. It is found that the checkerboard partitioning produces the best results when the number of processors becomes larger than 15, otherwise row block striped partitioning is to be preferred.

Keywords

Princeton Ocean Model domain decomposition message passing 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • L. A. Boukas
    • 1
  • N. Th. Mimikou
    • 1
  • N. M. Missirlis⋆
    • 1
  • G. L. Mellor
    • 2
  • A. Lascaratos
    • 3
  • G. Korres
    • 3
  1. 1.Department of InformaticsUniversity of AthensAthensGreece
  2. 2.Princeton UniversityPrincetonUSA
  3. 3.Department of Applied PhysicsUniversity of AthensAthensGreece

Personalised recommendations