[totalart]

Hi all,

I am running interfoam with quite a heavy mesh on 24 CPU. And I need to test with several meshing case. For this, what I am doing now is:

1. blockMesh
2. decomposePar(meshes are decomposed)
3. mpirun -np 24 snappyhexmesh.exe -overwrite -parallel
4. reconstructparMesh
5. (prepare fields including alpha1. in 0 directory)
6. setfields
7. decomposePar
8. run the solver in parallel

Only if I can run setfields while decomposed, I can save more time. However, error occurs at step 5 with followings:

1. blockMesh
2. (prepare fields including alpha1. in 0 directory.)
3. decomposePar (mesh and fields in 0 are decomposed)
4. mpirun -np 24 snappyhexmesh.exe -overwrite -parallel
5. mpirun -np 24 setfields.exe -parallel -> ERROR
6. run the solver

Do you know what is wrong for setfields in parallel??

[totalart]

I found the patch created by SHM is not shown in Paraview... It's because the patch is not created when I decompose(this is before running SHM). I added the patch with empty in 'boundary' before decomposing, and it's settled. But error still persists with setfields with parallel. The error message is:

keyword procBoundary16to12 is undefined in dictionary "C:\~~~~~\alpha1::boundaryField"

Boundaryfield for the SHM patch is defined in 0 folder. But I cannot read alpha1 in the decomposed 0, probably it's in binary.
I hope someone can give an advice.

[blue8803]

1. blockMesh
2. (prepare fields including alpha1. in 0 directory.)

** just put boundaryField such as "procBoundary.*" in your alpha1. file

"procBoundary.*"
{
type processor;
value uniform 0;
}


3. decomposePar -copyZero(mesh and fields in 0 are decomposed)
4. mpirun -np 24 snappyHexmesh -overwrite -parallel
5. mpirun -np 24 setFields -parallel -> ERROR
6. run the solver