The availability of parallel computing provides an opportunity for solving increasingly complex problems.

OpenFOAM employs domain decomposition to run cases on more than one processor. The domain is automatically decomposed into a number of sub-domains, each of which is solved on a separate processor. The communication between processors uses the MPI communications protocol (or shmem on Cray platforms).

OpenFOAM has been used for calculations with 10 million cells and has run on a 256 processor Cray T3E.

OpenFOAM displays excellent scaling performance, i.e. reduction in computing time with increase in number of processors. The table below presents timing data for a 3-D linear stress analysis problem on a Diesel injector valve seat with 360,000 cells. The calculation was performed on a 24 CPU Silicon Graphics Origin 2000 in non-dedicated mode, i.e. other jobs were running on the machine during the test.