On 10/12/2017 12:53 PM, Oded Yaakobi wrote:
To my surprise, the solution that was obtained using UMFPACK solver is
very different than the one that was obtained by the original solver of
step-22. Attached are screenshots of the velocity x-component of the
original solution and the one that was obtained from the direct solver.
Does anyone have any explanation to the discrepancy?
At the very least you'll have to fix up hanging nodes via
hanging_node_constraints.distribute (solution);
How do things look after that?
But there is also the issue that in step-22, we put something into the
bottom right block of the matrix (namely, a pressure mass matrix) that
does not belong there. The solver in step-22 doesn't actually solve with
system_matrix, but only with the top left, top right, and bottom left
block. If you want to use a direct solver, you need to remove the code
that puts the mass matrix into that block.
Best
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: [email protected]
www: http://www.math.colostate.edu/~bangerth/
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
