On Fri, Jun 13, 2014 at 6:20 PM, Miguel Angel Salazar de Troya
<salazardetr...@gmail.com> wrote:
> I was able to run the mesh refinement for just one processor, it seems to
> work fine. However, when I run it with two processors, I get this error:
>
> Assertion `coarse->parent()' failed.
> [1] src/error_estimation/uniform_refinement_estimator.C, line 545, compiled
> Jun 10 2014 at 14:35:08
>
> The code snippet for that line is:
>
>           for (; elem_it != elem_end; ++elem_it)
>             {
>               // e is necessarily an active element on the local processor
>               const Elem* elem = *elem_it;
>
>               // Find the element id for the corresponding coarse grid
> element
>               const Elem* coarse = elem;
>               dof_id_type e_id = coarse->id();
>               while (e_id >= max_coarse_elem_id)
>                 {
>                   libmesh_assert (coarse->parent());
>                   coarse = coarse->parent();
>                   e_id = coarse->id();
>                 }
>
>           ...
>
> Could you help me out with what the code is doing here? I am struggling
> figuring out how to debug a program in parallel. I ask the question to see
> if you guys could help me in the mean time. What is the element "id" ? What
> is its purpose? Any idea why it is failing? I haven't introduced any
> modification in the mesh or elem data structures. Thanks in advance.

You are supposed to be on an active element, which means you have a
parent, but parent() == NULL, hence the error.

I'm not really sure how you could get to this state, because I can't
see the rest of your changes...

By the way, the line numbers you quoted are already off from master.
If you are serious about doing development, you should really start
using git and a libmesh fork.  Then we can actually look at your
branch, see the entire diff, and have a better idea of where bugs
might be.

-- 
John

------------------------------------------------------------------------------
HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions
Find What Matters Most in Your Big Data with HPCC Systems
Open Source. Fast. Scalable. Simple. Ideal for Dirty Data.
Leverages Graph Analysis for Fast Processing & Easy Data Exploration
http://p.sf.net/sfu/hpccsystems
_______________________________________________
Libmesh-users mailing list
Libmesh-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to