Periodic boundary conditions on multi processors

[carlodev]

I am having some strange behaviour on periodic boundary conditions. In the case in this repo, I have generated the .msh file using Gmsh and imposed the periodic boundary condition on the plane z = 0, and the parallel plane on the other side for a z = const. The code used to generate the model is in the repo (.sh file), the OpenMesh.jl is the one used for the serial case. In the folder n64 there is the model printed by the code using 64 cores, and so on for the other folders. As you can see, if look at the zm in the _0 files, there are some points also over the other plane, and the number and distribution of this "random extra points" change with the number of processor; there are none in the sequential case.

Thanks for the support

repo: https://github.com/carlodev/naca_3d_problem.git