On Mon, 16 Nov 2015, Miguel Angel Salazar de Troya wrote:
> My idea is to output an EquationSystems after a simulation in which I might
> have refined/coarsened the mesh and calculated the solutions of the
> System's. I would like to have the option to read that output and restart
> the simulation where I left it off. So far what I've done is simply this
>
> equation_systems.write("eq_output ")
>
> And then to read it
>
> equation_systems.read("eq_output ")
>
> But for some reason, the dofs of my Systems are doubled. I was thinking it
> might be because I create all the Systems, init() them and then I make a
> check to see if I want to read "eq_output" . That could the reason, but I
> don't want to include all the calls to add_system() inside of a huge
> conditional. Is there any way to do this in a cleaner way? Maybe an
> overwrite flag to pass to EquationSystems::read()
I believe you'll want something like
libMesh::EquationSystems::READ_DATA |
libMesh::EquationSystems::READ_ADDITIONAL_DATA
for the flags. Leaving off READ_HEADER tells the read() method that
you've already set up the systems yourself and it shouldn't try to do
so from the restart file headers.
---
Roy
------------------------------------------------------------------------------
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users
