On Wed, 25 Jun 2008, Benjamin Kirk wrote:

> It can't be a bug in my redistribution, since it is never called!
> So obviously that is the problem.  Repartitioning is performed, but the mesh
> is not redistributed.  This causes Roy's assertion above to break.

Oh, that would definitely break things!

Hehe; we should always be redistributing after repartitioning in
parallel, but I just assumed that that was already being done.
Whoops!

> I'll comment-out the repartitioning call and check that in to svn, and at
> the same time work here to get the redistribution integrated.

Sounds good.

> It is has been a while since I've touched ParallelMesh -- can you remind me
> what is known not to work?

At the moment, there are some adaptive coarsening special cases that I
expect to break, IIRC because I couldn't figure out how to keep the
remote_elem neighbor pointers correct.  (That might be a serious
problem for redistribution too...)  Uniform and adaptive refinement
should be working.  I'm not sure about uniform coarsening.
---
Roy

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Libmesh-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-devel

Reply via email to