A simple parallelization technique with MPI for ocean circulation models

Abstract

This paper proposes a comparatively simple technique based on overlapping domain decomposition for the transformation of serial computational algorithms of ocean circulation models into parallel ones. The advantage of the proposed technique is that only additional procedures using basic MPI functionality are required instead of a significant transformation of serial code. The method is fairly efficient for computations on comparatively inexpensive clusters, which are usually affordable for small companies. The impacts of various factors, which may affect the method efficiency are studied. © 2009 Elsevier Inc. All rights reserved.