[Libmesh-users] Experience with distributed creation om mesh

[Michael P. <mpo...@pu...>]

Deal Libmesh developers,
as I wrote you before I was trying to create a mesh in a distributed way 
and solve the Poisson equation on it.

Good news: it works
a) a Mesh object  is created (It seems that by default it is a 
SerialMesh object, not a ParallelMesh)
b) nodes and elements that belong to a single partition only are added
c) mesh is prepared for use without renumbering nodes

Then the system is initialized on the mesh, and the solution process 
works. The results I'm getting look okay.

Bad news:
1) If libmesh is compiled in the debug mode there are assertions that fail
2) The serilized output, such as Tecplot does not work

My question:
Is my tactic correct? Or do I need to do something else?
Are you interested to know which assertions fail?

Thank you,
Michael.