Problem with mpirun
 

Now I am using k-epsilon model to calculate the pollutant transfer. I set the pollutant as a user defined scalar and also set the open boundary in the solver, however, I find I will have different result when I am using one core or using parallel calculation.

And I check the file, I finde in each part data, I can find the boudary,e.g, uniform 0, in inflow, but after reconstruct, disappeared.

Are there any suggession about decompose and reconstruct and parallel calculation?

Thank you!