[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Getfem-users] Problems with icare
From: |
renard |
Subject: |
Re: [Getfem-users] Problems with icare |
Date: |
Thu, 10 Aug 2006 16:11:46 +0200 |
User-agent: |
Internet Messaging Program (IMP) 3.2.5 |
Quoting John Burnell <address@hidden>:
> Hi
>
> I recently downloaded getfem++ 2.0.1 and compiled it. I tried a couple
> of the test programs and they worked okay. I then tried to run the
> navier-stokes solver icare (in contrib/icare). The program failed with a
> "dimension mismatch" error.
>
> I tried to find the problem with the debugger and found a problem in the
> proper_update routine in icare.h - the indices vector was not being
> cleared so each time proper_update was called, indices grew larger and
> larger.
>
> Fixing that problem gave another "dimension mismatch" error, but I don't
> understand this one. It occurs when on the first call of
>
> getfem::standard_solve(MSM, mixed_dyn, iter);
>
> on line 480 of icare.cc. This then results in a call sequence
>
> mdpb.compute_tangent_matrix at line 446 of getfem_model_solvers.h
> pb.compute_tangent_matrix at line 363 of getfem_model_solvers.h
> pb.compute_tangent_matrix at line 363 of getfem_model_solvers.h
> ((mdbrick_abstract*)sub_bricks[i])->compute_tangent_matrix(MS, i1, j1);
> at line 473 of getfem_modeling.h
> ((mdbrick_abstract*)sub_bricks[i])->compute_tangent_matrix(MS, i1, j1);
> at line 473 of getfem_modeling.h
> do_compute_tangent_matrix(MS, i0, j0); at line 477 of getfem_modeling.h
> gmm::copy(get_B(), gmm::sub_matrix(MS.tangent_matrix(), SUBI, SUBJ)); at
> line 1569 of getfem_modeling.h
>
> Which is where the dimension mismatch occurs. The first matrix has
> dimension of 786x1 and the second 1572x1.
>
> At the moment I don't understand exactly what this sequence of calls is
> doing, so I thought I would ask if anyone knows what is going on?
>
> I have compiled getfem++ under Linux and Windows with Mingw and get the
> same error.
>
I think the problem is due to the fact that we made the getfem++ archive with a
non working icare.cc (someone worked on the files at the corresponding period).
Could you please try to download the icare's files from the gna repository:
http://svn.gna.org/viewcvs/getfem/trunk/getfem%2B%2B/contrib/icare/
and see if it better works ?
Sincerly,
Yves.