Yes, that's true. I'm trying to use an external eigensolver called FEAST,
instead of SLEPc. So instead of

_system->solve();

I call

assemble_eigensystem(*_equation_systems, _system_name);
_system->matrix_A->close();
_system->matrix_B->close();

where assemble_eigensystem() is the function which assembles the system.
Then I call _system->matrix_A->mat() to get the PETSc matrix.

On Mon, Oct 20, 2014 at 4:40 PM, John Peterson <[email protected]> wrote:

> On Mon, Oct 20, 2014 at 1:42 PM, Harshad Sahasrabudhe
> <[email protected]> wrote:
> > It's a CondensedEigenSystem. I initialize the EquationSystems object that
> > contains that System, so I guess it is initialized? I also assemble and
> > close the PetscMatrix before calling mat().
>
> The CondensedEigenSystem uses two different matrices,
> condensed_matrix_A and condensed_matrix_B.
>
> It looks like they are not created until you call solve(), see
> condensed_eigen_system.C line 97.
>
> Is one of these the matrix you are trying to call mat() on?
>
> --
> John
>
------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://p.sf.net/sfu/Zoho
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to