Martin,

the problem statement should be OK, because I chose the problem from step-11 
(f=-2, g=1). But I found the solution by myself. I have to choose finite 
elements of order 1, then it runs well. At the moment I don't know why, but I 
will figure out.

Thank you for your help,
Markus



On Fri March 20 2009 7:02 pm, Martin Kronbichler wrote:
> Markus,
>
> I tried your program and could not easily resolve the convergence
> problem. I even tested step-11 with distribute_local_to_global instead
> of the condense function on the matrix, and there everything works fine
> (with Neumann conditions all over the boundary), so I don't think the
> problem is due to the ConstraintMatrix::distribute_local_to_global
> function itself. I'd guess that something else (something with the
> problem statement) makes the system matrix and RHS incompatible and
> destroys the convergence of the iterative solver (not even BiCGStab
> converges on your system), like e.g. that the Neumann boundary function
> and the right hand side function do not satisfy the condition stated in
> the tutorial step-11 introduction. Do you think anything like that
> could cause the error (or some other incompatibility)? Or has anyone
> else an idea?
>
> Best regards,
> Martin
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to