On Thu, 3 Sep 2015, Miguel Angel Salazar de Troya wrote:

> Is libmesh able to handle adaptive mesh refinement/coarsening in dynamic
> problems over multiple processors?

Yes.

> I can imagine there must be a significative overhead because of the
> load unbalances if we adaptively refine the mesh every other time
> step.

It depends on your application.  IIRC the overhead of adapting at all
(and so forcing sparsity pattern recomputation, matrix reallocation,
vector reallocation and projection, etc) is comparable to the extra
overhead of repartitioning and communicating.

> Taking things further, could we save the history of these
> coarsening/refinements in order to perform an adjoint analysis
> (backwards in time)? I'm just wondering if this is possible, I
> currently do not need the details.

Possible, but currently unassisted by the library.  The only
internal support for transient adjoints comes via FEMSystem (that
being the only framework that gives the library a rich enough handle
on the time integration) and it currently only works for small
problems and with fixed meshes.

We have some ideas for changing that, if you want to be a collaborator
and/or an alpha tester.

> Last but not least, last year the mesh refinement was not available for
> vector finite elements such as LAGRANGE_VEC, is this still not possible?

I'm afraid it's not.  There's a refactoring branch (for handling AMR
on mixed-dimensional manifolds) which should make it much easier to
add AMR on vector elements as well, but even that's on hold for a bit
while higher priority stuff buries me.
---
Roy

------------------------------------------------------------------------------
Monitor Your Dynamic Infrastructure at Any Scale With Datadog!
Get real-time metrics from all of your servers, apps and tools
in one place.
SourceForge users - Click here to start your Free Trial of Datadog now!
http://pubads.g.doubleclick.net/gampad/clk?id=241902991&iu=/4140
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to