Version 1.3 contains numerous developments aimed towards improving the efficiency and robustness of OpenFOAM for large-scale engineering CFD.

• Improved parallel running

  Parallel running of OpenFOAM has been improved significantly with: easy implementation of integral boundary conditions, e.g. fixed mass flow rate; faster performance of the parallelised AMG solver; faster parallel communications.

• Increased speed

  Significant improvements in speed and memory storage through careful code optimisation.

• Better numerics

  Several discretisation schemes have been added for better accuracy and better stability on "bad" (e.g. tetrahedral) meshes.

• Single precision option

  OpenFOAM can be switched from double to single precision for all floating point numbers to reduce memory requirements, e.g. for large steady-state external aerodynamics calculations (F1).

• OpenMPI

  Support for and shipped with OpenMPI (although still released compiled with lam-MPI by default).

The other main changes have been predominately driven by commercial demand and comprise of the following:

• Development tools

  New, automatic stack trace on code abort; optional initialisation of dynamically allocated memory to trap use of uninitialised floating point variables; processor-labelled error messages for easier debugging.

• Application development

  The developments of the solvers centre around low and high speed aerodynamics, heat transfer and buoyancy-driven flows. There are many new utilities, particularly for mesh conversion/manipulation and post-processing.

• Mesh handling

  There has a major reorganisation and development of mesh structures to improve capability for generation, topological change, sliding interfaces, etc.

• Switchable base system of units

  Physical constants are no longer hard-coded but read in from file enabling the use of a different base system of units from the default (SI).

• Changes to models

  There are some changes in particular relating to wall functions of turbulence models.

• Library restructure

  The finite volume method is separated from the main library so that other discretisation techniques may be easily supported; libraries have been reordered to prevent cyclic dependencies.

• Compiler

  Released compiled with GCC 4.1, a version with superior optimisation, better adherence to the C++ standard and improved error messaging; support for the Intel C++ 9.0 compiler.