libMesh: A C++ Finite Element Library

Dear Derek,

On Wed, 15 Nov 2006, Derek Gaston wrote:

> So far I'm not having any luck though.  When I try to do something like:
>
>    EquationSystems overkill_es(overkill_mesh);
>    overkill_es.read(overkill_sol_file,
>                     libMeshEnums::READ,
>                     EquationSystems::READ_DATA);
>
> It bombs out saying "ERROR: no system named Porous Media found!"
> Where Porous Media is the name of the system contained in the xda
> file.

The following works for me:

Mesh* mesh = new Mesh(3);
mesh->read(meshFileName);
EquationSystems* equationSystems = new EquationSystems(*mesh);
equationSystems->read(systemsFileName,libMeshEnums::READ,EquationSystems::READ_DATA|EquationSystems::READ_ADDITIONAL_DATA|EquationSystems::READ_HEADER);

Best Regards,

Tim

Tim,

You are absolutely correct!  Thanks!

I had already tried that... but it errored out with a weird error.
Turns out that I generated the solution files with a libmesh that had
compiled in infinite element support... meanwhile I was trying to read
the solution using a libmesh that didn't have infinite element
support... and this was screwing everything up.  So when it was
breaking I thought it was because I couldn't do what I wanted... not
because there was something else wrong.

Everything is working now!

Thanks for the help!  I really appreciate it!

Derek

On 11/15/06, Tim Kr=F6ger <tim@...> wrote:
> Dear Derek,
>
> On Wed, 15 Nov 2006, Derek Gaston wrote:
>
> > So far I'm not having any luck though.  When I try to do something like=
:
> >
> >    EquationSystems overkill_es(overkill_mesh);
> >    overkill_es.read(overkill_sol_file,
> >                     libMeshEnums::READ,
> >                     EquationSystems::READ_DATA);
> >
> > It bombs out saying "ERROR: no system named Porous Media found!"
> > Where Porous Media is the name of the system contained in the xda
> > file.
>
> The following works for me:
>
> Mesh* mesh =3D new Mesh(3);
> mesh->read(meshFileName);
> EquationSystems* equationSystems =3D new EquationSystems(*mesh);
> equationSystems->read(systemsFileName,libMeshEnums::READ,EquationSystems:=
:READ_DATA|EquationSystems::READ_ADDITIONAL_DATA|EquationSystems::READ_HEAD=
ER);
>
> Best Regards,
>
> Tim
>
>

Dear Derek,

>> Mesh* mesh = new Mesh(3);
>> mesh->read(meshFileName);
>> EquationSystems* equationSystems = new EquationSystems(*mesh);
>> equationSystems->read(systemsFileName,libMeshEnums::READ,EquationSystems::READ_DATA|EquationSystems::READ_ADDITIONAL_DATA|EquationSystems::READ_HEADER);

I should add that as per my experience, it is better to use xdr files 
than xda, because in xda files all numbers are truncated to 6 digits. 
Very small cells are likely to degenerate due to these round-off 
errors.

Best Regards,

Tim

Hmmm... very good point.

I have xdr's for one of my program's output.... but only xda's for the
other.  That could be a problem.  Let's hope not though... because
rerunning all of those results could take weeks... sigh.

Derek

On 11/15/06, Tim Kr=F6ger <tim@...> wrote:
> Dear Derek,
>
> >> Mesh* mesh =3D new Mesh(3);
> >> mesh->read(meshFileName);
> >> EquationSystems* equationSystems =3D new EquationSystems(*mesh);
> >> equationSystems->read(systemsFileName,libMeshEnums::READ,EquationSyste=
ms::READ_DATA|EquationSystems::READ_ADDITIONAL_DATA|EquationSystems::READ_H=
EADER);
>
> I should add that as per my experience, it is better to use xdr files
> than xda, because in xda files all numbers are truncated to 6 digits.
> Very small cells are likely to degenerate due to these round-off
> errors.
>
> Best Regards,
>
> Tim
>
>