Dear Luca, Markus and Michael,

  Thank  you  for the suggestions! I'll follow Luca's idea - it sounds
  to    be   the  most general. Computing the derivatives analytically
  with  any of the symbolic calculus packages in my particular case is
  not reasonable, but still I want to have the derivative to be exact,
  and not approximated via the finite differences. In any case, thanks
  you all guys!

Regards,
Slawa

Thursday, September 2, 2010, 17:36, you wrote:

MR> Hi there,

MR> Another possibility which might work for you would be to calculate the
MR> derivative explicitly (possibly using a symbolic math package if the
MR> function is involved) and create a Function<dim> which is the
MR> derivative. Since Function<dim> is user defined it seems like in many
MR> cases the user could do this, however it might not be suitable for
MR> your application.

MR> Cheers,
MR> Michael

MR> On Thu, Sep 2, 2010 at 5:22 PM, Markus Bürg <[email protected]> wrote:
>> Hello Slawa,
>>
>> as far as I know there is no easy way of getting the derivative of a
>> Function<dim> object with Sacado, since (of course) Trilinos does not know
>> the deal.II Function<dim> object. I think the easiest way for you would be
>> to use the AutoDerivativeFunction class from deal.II. If you do not want to
>> do this, please have a look in tutorial step-33 to see how to use Sacado.
>>
>> Best Regards,
>> Markus
>>
>>
>>
>> Am 02.09.10 12:41, schrieb Svyatoslav Gladkov:
>>>
>>> Greetings!
>>>
>>>    I  have  rather  naive  question about automatic differentiation in
>>>    deal.II:
>>> Suppose I have a scalar function derived from Function<dim>. How can I
>>> compute  derivative  (gradient)  of  this  function  in all quadrature
>>> points with Sacado?
>>>
>>> Regards,
>>> Slawa
>>>
>>> _______________________________________________
>>> dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
>>>
>>>
>>
>> _______________________________________________
>> dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
>>
MR> _______________________________________________
MR> dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii




-- 
Best regards,                                       02.09.2010
Svyatoslav Gladkov, Kharkov, UKRAINE                18:19

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

Reply via email to