To enable utilisation of distributed-memory systems, parallelization has been performed based on the message passing paradigm. The distribution of work and data is based on the domain decomposition concept, and Message Passing Interface (MPI) is used for communication between processors. The computational mesh is partitioned into a number of physical subdomains, and the work associated with each subdomain is processed by an individual processor. The data exchange between processors is based on the halo-layer approach with overlapping elements.