[dune-pdelab] dune-pdelab Digest, Vol 28, Issue 2

[Markus Blatt]

Dear Jared,

please ALWAYS reply to the list!

On Mon, Jan 09, 2012 at 11:06:28AM +0100, Jared Okiro wrote:
> On 01/09/2012 09:35 AM, Markus Blatt wrote:
> >Hi,
> >
> >On Mon, Jan 09, 2012 at 09:24:55AM +0100, Jared Okiro wrote:
> >>Thanks for the response. I printed the matrix after the time
> >>loop(just after very few time steps). Here is the line of code that
> >>I used:
> >>
> >>Dune::printvector(std::cout,unew.base(),"const residual","row",4,9,1);
> >>
> >I am a little bit confused. If unew is matrix, you must not use
> >printvector. 
>
> Thanks for the response. When I try to use use any of the above
> methods, I get this error:
> 
> /homes/okiro/DUNE/dune-svn/dune-istl/dune/istl/io.hh:252:5: error:
> ‘coldim’ is not a member of
> ‘Dune::MatrixDimension<Dune::BlockVector<Dune::FieldVector<double,
> 4>, std::allocator<Dune::FieldVector<double, 4> > > >’

Which means that unew IS NOT a matrix, but a vector. Your old usage
was correct then.
BTW: setting columns to 4 when calling printvector is confusing to me.

Of course, we cannot tell anything from the output of the residual. 

Please answer the following questions:
- Which example file are you referring to
  (convection-diffusion/dgdiffusion.cc)? 
- Have you tried tried a test case that is not time dependent?
- At which time step /new iteration is this happening?
- What is the time stepping scheme?
- What is the linear solver?

Markus


-- 
Dr. Markus Blatt - HPC-Simulation-Software & Services http://www.dr-blatt.de
Rappoltsweilerstr. 5, 68229 Mannheim, Germany,  USt-Id: DE279960836
Tel.: +49 (0) 160 97590858  Fax: +49 (0)322 1108991658