Advertisement

A Parallel Algorithm with Improved Performance of Finite Volume Method (SIMPLE-TS)

  • Kiril S. Shterev
  • Stefan K. Stefanov
  • Emanouil I. Atanassov
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7116)

Abstract

In this paper a parallel version of the finite volume method SIMPLE-TS for calculation of two-dimensional compressible, viscous gas flows with improved performance is presented. It is demonstrated on a problem regarded to micro gas flows, taking place in Micro-Electro-Mechanical Systems (MEMS). The reorganisation of the parallel algorithm improve the algorithm performance, when more cores are used for calculations on computational grids with relatively small number of nodes or cells. The reorganisation is two-fold: first to reduce the number of communications between the processes, and second to reorder the calculation of some variables in such a way that increases the number of calculations during the communications between the processes. The comparison of speed-up between previous and new parallel versions of SIMPLE-TS was performed on two types of clusters with regard to the communication hardware: the first uses specialised cards with low latency for the interconnections between the computers and the other uses conventional cards for the interconnections. The clusters are a part of the GRID-infrastructure of the European Research Area (ERA).

Keywords

finite volume method SIMPLE-TS gas microflows parallel algorithms GRID 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    InfiniBand trade association, http://www.infinibandta.org
  2. 2.
    Introduction and hardware charachteristics of InfiniBand, http://en.wikipedia.org/wiki/InfiniBand
  3. 3.
    The official MPI (Message Passing Interface) standards documents, errata, and archives of the mpi forum, http://www.mpi-forum.org
  4. 4.
    Bui, T.T.: A parallel, finite-volume algorithm for large-eddy simulation of turbulent flows. Computers & Fluids 29(8), 877–915 (2000), http://dx.doi.org/10.1016/S0045-79309900040-7 zbMATHCrossRefGoogle Scholar
  5. 5.
    Shterev, K.S., Stefanov, S.K.: A Parallelization of Finite Volume Method for Calculation of Gas Microflows by Domain Decomposition Methods. In: Lirkov, I., Margenov, S., Waśniewski, J. (eds.) LSSC 2009. LNCS, vol. 5910, pp. 523–530. Springer, Heidelberg (2010), http://www.springeronline.com/978-3-642-12534-8 CrossRefGoogle Scholar
  6. 6.
    Shterev, K.S., Stefanov, S.K.: Pressure based finite volume method for calculation of compressible viscous gas flows. Journal of Computational Physics 229(2), 461–480 (2010), http://dx.doi.org/10.1016/j.jcp.2009.09.042 MathSciNetzbMATHCrossRefGoogle Scholar
  7. 7.
    Stefanov, S., Roussinov, V., Cercignani, C.: Rayleigh-bénard flow of a rarefied gas and its attractors. I. Convection regime. Physics of Fluids 14(7), 2255–2269 (2002)MathSciNetCrossRefGoogle Scholar
  8. 8.
    Versteeg, H.K., Malalasekra, W.: An Introduction to Computational Fluid Dynamics: The Finite Volume Method, 2nd edn. Prentice Hall, Pearson (2007)Google Scholar
  9. 9.
    Wang, P., Ferraro, R.D.: Parallel multigrid finite volume computation of three-dimensional thermal convection. Computers & Mathematics with Applications 37(9), 49–60 (1999), http://dx.doi.org/10.1016/S0898-12219900113-3 MathSciNetCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Kiril S. Shterev
    • 1
  • Stefan K. Stefanov
    • 1
  • Emanouil I. Atanassov
    • 2
  1. 1.Institute of MechanicsBulgarian Academy of SciencesSofiaBulgaria
  2. 2.Institute of Information and Communication TechnologiesBulgarian Academy of SciencesSofiaBulgaria

Personalised recommendations