Introduction

SWAN can be installed on various architectures, including laptops and supercomputers. The portability of the SWAN source code is guaranteed by the use of standard ANSI Fortran 90. (See also the manual Programming rules.) Hence, virtually all Fortran compilers can be used for installing SWAN. It should be noted that there are two Fortran commands used in the source code of SWAN (v41.41+) which were introduced in later versions of Fortran: stream I/O (Fortran 2003 standard) and execution OS command line (Fortran 2008 standard). They are, however, supported by currently maintained Fortran compilers, including gfortran and Intel Fortran.


The SWAN source code also supports parallelization, which enables a considerable reduction in the wall-clock time for relatively large CPU-demanding calculations. Two parallelization strategies are available:

• The computational kernel of SWAN contains a number of OpenMP compiler directives, so that users can optionally run SWAN on a cluster or laptop with multi shared-memory processors.
• A message passing modelling is employed based on the Message Passing Interface (MPI) standard that enables communication between independent processors. Hence, users can optionally run SWAN on a Linux cluster containing memory-distributed nodes.

The SWAN software can be build in the usual way via GNU make or from scratch. This building process is explained in Section 3.2. However, since version 41.41, the option to install SWAN using CMake is supported and is elaborated in Section 3.3.