AN HYBRID OPENMP–MPI PARALLELIZATION OF THE PRINCETON OCEAN MODEL

This paper deals with the parallelization of one of the most popular threedimensional oceanographic model: the Princeton Oceean Model (POM). The parallelization is achieved using standard tools like MPI and OpenMP, to ensure portability of the code. A robust and ecien t domain decomposition method is used to solve, in principle, large scale examples on clusters of shared memory machines. A comparison between the pure MPI and the hybrid MPI+OpenMP is shown in terms of elapsed time for a standard seamount problem.