On Wed, 18 Feb 2009, Roy Stogner wrote:

> I'm swamped right now.  The next step is to reduce ex10 to a simpler
> case (maybe solving du/dt = 0 in nD with some non-zero ICs?) where I
> can step through the whole projection in gdb, but it's likely to be
> weeks before I can get to that.

I apparently underestimated the productivity benefits of trying to
avoid twiddling my thumbs while the supercomputer I'm using undergoes
maintenance.

The problem in ex10 (and in my reduced version of it) was this line:

*system.old_local_solution = *system.current_local_solution;

which correctly copied local but not ghosted degree of freedom values.


Explicitly correcting ghost dofs at the end of operator=() fixed it:

   // Make sure ghost dofs are copied
   if (this->type() == GHOSTED)
     this->close();


And ex10 seems to be working correctly now as well, as do the rest of
the example codes.  I'll keep trying to trigger bugs with my
application codes, and I'll do some performance benchmarking, before I
consider turning on ghosted vectors in SVN.  But additional user
testing would be nice too.  I think I'll make --enable-ghosted-vectors
a configure option to make the code easier to play with while we're
still not confident about it.


Besides, this can't be the right fix, can it?  Why should any parallel
communication be required to copy one consistent ghosted vector to
another?  It seems as if VecCopy is simply failing to copy the whole
vec, and I don't like not understanding why.
---
Roy

------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
Libmesh-devel mailing list
Libmesh-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libmesh-devel

Reply via email to