On Tue, 12 Jan 2010, Rahul Sampath wrote:

> Does the global dof id change if the number of processors change

Yes.  Every repartitioning or post-refinement/coarsening step
renumbers ids, in fact, so that we can have efficient partitions with
contiguous id blocks.

> or does it remain the same for a given mesh immaterial of the number
> of processors used?

There is an option to avoid renumbering nodes and elements when
reading a new mesh; that combined with --disable-libHilbert may give
you an unchanged id.

Actually, though, this is an outstanding bug - we currently ensure
consistent global ids between mesh and solution files by using
libHilbert to give us a unique ordering.  But a bug in our libHilbert
interface can occasionally prevent that.  --disable-libHilbert along
with skipping renumbering is the best current workaround, but the
final solution is going to have to be something that is guaranteed to
(temporarily) preserves even inefficient mesh numberings, but then
have the user manually do an efficient repartition/renumbering with
the current number of processors.  Ben's been working on that, but
we've both been busy lately.
---
Roy

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev 
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to