[morfeus80]

Hallo,
I'm going to run a case with a dual core processor: have I to make the partition of the grid and start a parallel process or not? If I have to do, what commands have I to type to run the case?

[mike_jaworski]

Hi Mattia,
My understanding is that if you want to use both cores of your processor you will have to split the mesh and run it with the mpirun command. There are explicit directions on doing this in the user guide for the third tutorial case called "DamBreak" which uses interFoam. It's very straight forward to run if you don't have to mess with connecting to other machines and the steps in the tutorial are fairly explicit.

Good luck!
Mike J.

[wikstrom]

Reference pressure for e.g. buoyantSimpleFoam:

Hi,

How is the reference pressure cell selection handled in parallel runs?

I use OpenFOAM-1.4.1 from the svn-repository (proposedly including bugfixes).

The following description is based on problems with buoyantSimpleFoam. As example the tutorial case 'hotRoom' is used.

1) Serial case runs fine. The pressure solution (iteration numbers and residuals) is clearly effected by the choice of pdRefCell.

2) Parallel case runs fine, as long as pdRefCell is a number smaller than the number of cells per process.

3) Using a larger pdCellRef cause floating point exception. Which might be understandable IF the pdRefCell is not translated during a parallel run.

So, how should the reference pressure be handled for parallel runs? For now it seems like it has to be positioned inside the processor0 domain(??).

cheers
Niklas