Optimising the Parallelisation of OpenFOAM Simulations

Abstract

The OpenFOAM computational fluid dynamics toolbox allows parallel computation of fluid dynamics simulations on a theoretically unlimited number of processor cores. The software utilises a message passing interface (MPI) library in order to facilitate communication between parallel processes as the computation is completed. In order to maximise the parallel efficiency of the computation, the way in which the MPI application distributes the problem among the processors needs to be optimised. This report examines the performance of a given high performance computing cluster with several OpenFOAM cases, running using a combination of MPI libraries and corresponding MPI flags. The effect of hardware variations on the speed of the computation is also briefly investigated. The results show a noticeable improvement in simulation time when certain flags are given to the MPI library in order to control the binding and distribution of processes amongst the cluster.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 2014
Accession Number
ADA612337

Entities

People

  • Shannon Keough

Organizations

  • Defence Science and Technology Group

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Cell Count
  • Central Processing Units
  • Computational Fluid Dynamics
  • Computations
  • Computers
  • Engineering
  • Flow
  • Fluid Dynamics
  • Fluid Flow
  • High Performance Computing
  • Local Area Networks
  • Networks
  • Operating Systems
  • Optimization
  • Parallel Computing
  • Parallel Processing
  • Specifications

Fields of Study

  • Computer science

Readers

  • Parallel and Distributed Computing.
  • Systems Analysis and Design