Anders Logg wrote:
> On Fri, Mar 06, 2009 at 04:30:56PM +0100, Patrick Riesen wrote:
>> Anders Logg wrote:
>>> On Thu, Mar 05, 2009 at 04:55:38PM +0100, Patrick Riesen wrote:
>>>> hi all
>>>>
>>>> i'm stuck on the following problem:
>>>>
>>>> i'd like to compute a term as
>>>>
>>>> eta = (gamma)^(-2/3), or general eta = gamma^(n) and n is real
>>>>
>>>>
>>>> in my code, gamma is dependent on a velocity function from a previous
>>>> interation, so gamma = gamma(u0), and how to compute gamma stems from a
>>>> form-file, so i do:
>>>>
>>>> _M_gamma = new gammaFunctional(u0);
>>>>
>>>> Function f_gamma(&mesh, *M_gamma, 0);
>>>>
>>>>
>>>> my question is now
>>>>
>>>> how can i create the other function f_eta = f_gamma^(2/3) in a nice way?
>>>>
>>>> what i thought about is to create a derived function class Eta where
>>>> overload the evaluation function in a way as
>>>>
>>>> double eval(const double* x) const
>>>> {
>>>> return (gamma.eval(x)*gamma.eval(x))^(-1/3)
>>>> }
>>>>
>>>> please tell me if this is a possible approach, or else advise me how to
>>>> handle this :-)
>>>>
>>>>
>>>> thanks a lot for the support,
>>>>
>>>> patrick
>>>>
>>>> (btw i do this together with dolfin-grade2 so i use dolfin0.8.1 )
>>> I'd suggest operating on the vectors of dofs. Send in the gamma
>>> function to the constructor of your Function subclass, pick out the
>>> vector of dofs, iterate over the dofs and just insert the values into
>>> the vector using pow(x[i], 2.0/3.0).
>>>
>>> Use the get/set methods in Vector to first pick out arrays, operate on
>>> the arrays and then put them back.
>>>
>> hi anders,
>> i understand how to create the f_eta class and hand in the f_gamma but
>> i'm uncertain on what to do inside f_eta exactly. can you give me some
>> more help, e.g. in a few pseudocode statements?
>> thanks a lot for the help,
>> patrick
>
> In the constructor of EtaFunction, do something like this:
>
> const Vector& gamma_vector = gamma.vector();
> const Vector& eta_vector = vector();
>
> for (uint i = 0; i < eta_vector.size(); i++)
> eta_vector.setitem(i, pow(gamma_vector[i], 3.0));
>
now i see the link between the gamma function and the Eta function.
many thanks for the instruction,
patrick
_______________________________________________
DOLFIN-dev mailing list
[email protected]
http://www.fenics.org/mailman/listinfo/dolfin-dev
