There are additional checks, warning and error messages that are processed 
in debug mode. So they're specifically in place to help detect problems 
like you're having. For this reason it is *highly* recommended to do your 
development in debug mode. This is mentioned numerous times in the 
documentation.

On the problem you're facing, my first guess is that you've forgotten to 
reinitialise FEValues object for this cell: fe_values.reinit(cell); 

Reading between the lines, it appears to me that you're trying to do 
"cell-based" assembly ("B" capturing all shape function gradients for a 
cell). If this is the case then are you aware that deal.II offers a 
slightly different way of performing the same assembly operation without 
the hassle of constructing an equivalent "B" matrix? In fact, you seem to 
be reverse engineering a lot of the basic operations that FEValues was 
designed to help you with. See some of the tutorials 
<https://www.dealii.org/8.4.1/doxygen/deal.II/step_5.html#Step5assemble_system> 
for examples of how we typically perform assembly, namely accumulating the 
contributions from each shape function individually and, with one call, 
computing the result of "det(J)*h dxi1 dxi2".

Regards,
Jean-Paul

On Sunday, December 18, 2016 at 11:33:25 PM UTC+1, 
[email protected] wrote:
>
> I am running with release mode at the moment (but this should not affect 
> the computation, am I right?). I passed the "update_inverse_jacobians" flag 
> to FEValues. I have only one cell for simplicity and it is a QGauss Q4 
> element with linear shape functions, hence with const QGauss<dim> 
> quadrature_formula (2).
>
> Kind regards,
> S. A. Mohseni
>

-- 
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.

Reply via email to