Re: [Getfem-users] Getfem-users Digest, Vol 86, Issue 4

[Yves Renard]

Dear Marco,

A small fix:

generic_assembly assem;
  assem.set("A=data$1(qdim(#1),qdim(#1),#2); B=data$2(qdim(#1),qdim(#1),#2);
C=data$3(qdim(#1),qdim(#1),#2);
M(#1,#1)+=comp(Base(#2).Base(#2).Base(#2).vBase(#1).vBase(#1))(k,l,m,:,i,:,j).A(j,o,k).B(o,p,l).C(p,q,m);");
  assem.push_mi(mim);
  assem.push_mf(mf_u1);
  assem.push_mf(mf_data);
  assem.push_vec(A);
  assem.push_vec(B);
  assem.push_vec(C);
  assem.push_mat(M);
  assem.assembly(rg);


And to answer to your question, you can declare vectors A,B,C simply as

std::vector<scalar_type> A(nn), B(nn), C(nn);

with nn = mf_data.nb_dof() * d * d, d being the dimensions. The matrices has to be stored in Fortran order inside the vectors and dof by dof.

Yves.





Le 09/10/2013 12:33, Marco Pischedda a écrit :

Dear Yves

thank you for your answer., mf_data is a scalar finite element space.

I have another question:

- what kind of data structures do I have to use to fill the matrix A,B,C? I tried to use a std::vector<vector<vector<scalar_type> > > but it gives me the following error:

error: cannot convert ‘gmm::strongest_value_type<gmm::cs_vector_ref<const double*, const unsigned int*, 0>, std::vector<std::vector<std::vector<double> > > >::value_type {aka std::vector<std::vector<double> >}’ to ‘bgeot::scalar_type {aka double}’ in assignment
AssemblingFunction.cpp:359:1:   instantiated from here
/home/pischedda/Dev/getfem-4.2/src/getfem/getfem_assembling_tensors.h:419:4: error: cannot convert ‘const std::vector<std::vector<double> >’ to ‘bgeot::scalar_type {aka double}’ in assignment

Thanks in advance

Marco

 

2013/10/9 <address@hidden>

Send Getfem-users mailing list submissions to
        address@hidden

To subscribe or unsubscribe via the World Wide Web, visit
        https://mail.gna.org/listinfo/getfem-users
or, via email, send a message with subject or body 'help' to
        address@hidden

You can reach the person managing the list at
        address@hidden

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Getfem-users digest..."


Today's Topics:

   1. Re: Assembing finite element matrices in GetFem++ (Yves Renard)


----------------------------------------------------------------------

Message: 1
Date: Wed, 09 Oct 2013 10:25:29 +0200
From: Yves Renard <address@hidden>
To: address@hidden
Subject: Re: [Getfem-users] Assembing finite element matrices in
        GetFem++
Message-ID: <address@hidden>
Content-Type: text/plain; charset="iso-8859-1"



Dear Marco,

if A(x), B(x) and C(x) are known matrix field described on a finite
element method, you can use the generic assembly to compute your mass
matrix, yes.
It should be similar to the mass matrix assembly in getfem_assembling.h
adding the matrix fields. This should give something like

 generic_assembly assem;
  assem.set("A=data(qdim(#1),qdim(#1),#2); B=data(qdim(#1),qdim(#1),#2);
C=data(qdim(#1),qdim(#1),#2);
M(#1,#1)+=comp(Base(#2).Base(#2).Base(#2).vBase(#1).vBase(#1))(k,l,m,:,i,:,j).A(j,o,k).B(o,p,l).C(p,q,m);");
  assem.push_mi(mim);
  assem.push_mf(mf_u1);
  assem.push_mf(mf_data);
  assem.push_mat(M);
  assem.assembly(rg);

assuming that mf_data is a scalar fem.

Yves.


Le 07/10/2013 16:37, Marco Pischedda a ?crit :
> Dear all,
>
> i'm a new user of GetFem++.
>
> I have a question for assembling a finite element matrix of the
> following type:
>
> \int A(x)B(x)C(x) f(x) \dot v(x) dx
>
> where A(x), B(x), C(x) are known finite element matrices, f(x)  is the
> unknown of the problem and v(x) is the test function. The problem is
> one-dimensional.
>
> There is an example of this type of assembling on the tutorial? It is
> possible to assemble matrices of this type?
>
> Thank's in advance
>
> Marco Pischedda
>
>
> _______________________________________________
> Getfem-users mailing list
> address@hidden
> https://mail.gna.org/listinfo/getfem-users


--

  Yves Renard (address@hidden)       tel : (33) 04.72.43.87.08
  Pole de Mathematiques, INSA-Lyon             fax : (33) 04.72.43.85.29
  20, rue Albert Einstein
  69621 Villeurbanne Cedex, FRANCE
  http://math.univ-lyon1.fr/~renard

---------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: </public/getfem-users/attachments/20131009/89624b72/attachment.html>

------------------------------

_______________________________________________
Getfem-users mailing list
address@hidden
https://mail.gna.org/listinfo/getfem-users


End of Getfem-users Digest, Vol 86, Issue 4
*******************************************

_______________________________________________
Getfem-users mailing list
address@hidden
https://mail.gna.org/listinfo/getfem-users

-- 

  Yves Renard (address@hidden)       tel : (33) 04.72.43.87.08
  Pole de Mathematiques, INSA-Lyon             fax : (33) 04.72.43.85.29
  20, rue Albert Einstein
  69621 Villeurbanne Cedex, FRANCE
  http://math.univ-lyon1.fr/~renard

---------