Hi Slawa

That would make sense but how would that overcome this issue of the strong 
imposition of the hanging node constraints? Am I missing something?


Cheers
Andrew
 
On 06 Sep 2010, at 2:01 PM, Svyatoslav Gladkov wrote:

> Hi, Andrew!
> 
>   Perhaps  its  possible  to  compute only the material tangent using
>   algorithmic  differentiation  and  compute  all  other parts of the
>   linearized weak form by hand?
> 
> Regards,
> Slawa
> 
> Monday, September 6, 2010, 12:03, you wrote:
> 
> AM> Hi all
> 
> AM> I'm interested in using the Sacado package in Trilinos to
> AM> numerically evaluate the global Newton matrix for a problem in
> AM> which I have have hanging nodes handled via standard hanging-node
> AM> constraints (i.e strongly). What I would like to know is how hard is this 
> to do?
> 
> AM> I have read the warning in tut-33...
> AM> "because we would have to thread the automatic differentiation we
> AM> use to compute the elements of the Newton matrix from the residual
> AM> through the operations of the ConstraintMatrix class. This would
> AM> be possible, but is not trivial, and so we choose this alternative 
> approach."
> 
> AM> Many thanks
> AM> Andrew 
> 

_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to