[Gmsh] Bug when loading 2 files resulting from a domain decomposition solve

Christophe Geuzaine cgeuzaine at ulg.ac.be
Mon Jan 28 17:13:19 CET 2013


Hello Nico - In your new files the meshes have different names (Mesh_2_1 and Mesh_2_2), so we treat them as independent models.

You can check this e.g. in "Visibility->Tree browser", and there you can choose which entities (of both models) to display...

On 28 Jan 2013, at 13:42, Karin&NiKo <niko.karin at gmail.com> wrote:

> Dear Christophe,
> 
> First I would like to thank you warmly for your  responsiveness : the new feature is great.
> 
> I have 2 remarks : 
> - I think their is a little bug when you rebuild the Physicals from the med result file. In order to check it, I have attached an archive containing the initial mesh + the 2 partitioned meshes (you can see their are 2 new Physicals called joint_1 and joint_2 that are the interfaces between the subdomains)
> 
> - I would like to know if it is possible to produce the following view : I would like to load the different results files, and, on the global view, display the interfaces (the join_* Physicals)? I have not succeeded in doing that (when trying to display the interfaces, Gmsh also displays the view on these Physicals).
> 
> Regards,
> Nicolas
> 
> 2013/1/27 Christophe Geuzaine <cgeuzaine at ulg.ac.be>
> 
> Hello Nico - Could you test with the next nightly? When reading MED files we now use the filename as a kind of "partition indicator". This should fix the problem that you encountered, which is linked to the fact that your MED file has not a full profile, but uses implicit entity indexing.
> 
> Christophe
> 
> On 25 Jan 2013, at 14:55, Karin&NiKo <niko.karin at gmail.com> wrote:
> 
> > Dear Gmsh Gurus,
> >
> > The attached archive contains 2 med files resu.proc.0.med and resu.proc.1.med, that result from a parallel solve with a domain decomposition approach.
> > But, when loaded in Gmsh, the visu is all but correct : this can be easily checked by loading the files separately.
> >
> > Is it a bug? Perhaps this feature is not in Gmsh yet.
> >
> > Thanks,
> > Niko
> > _______________________________________________
> > gmsh mailing list
> > gmsh at geuz.org
> > http://www.geuz.org/mailman/listinfo/gmsh
> 
> --
> Prof. Christophe Geuzaine
> University of Liege, Electrical Engineering and Computer Science
> http://www.montefiore.ulg.ac.be/~geuzaine
> 
> 
> 
> 
> <MAIL.7z>

-- 
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science 
http://www.montefiore.ulg.ac.be/~geuzaine