[Paraview] Re: ParaView Digest, Vol 31, Issue 25

Sun Nov 19 16:14:16 EST 2006

Hello Fabian,

I have a patch for ParaView/VTK allowing you to read EnSight gold files
with polyhedra. You can find the patch (for ParaView's CVS for version
2.5 as of March 3, 2006) at the following URL:
http://yvan.fournier.free.fr/paraview/nfaced/

To apply the patch on a current CVS version, overwrite the EnSight
reader files in ParaView's src/VTK/IO directory with the ones from
the patch, modify the CVS/Entries file in the same subdirectory
so that the following versions are given:

vtkEnSightGoldBinaryReader.cxx/1.64/ 
vtkEnSightGoldReader.cxx/1.56/ 
vtkEnSightGoldReader.h/1.20/ 
vtkEnSightReader.cxx/1.61/ 
vtkEnSightReader.h/1.27

then update your CVS directory (to merge changes to those files)
These files have changed little over time (yesterday's versions
were at vtkEnSightReader.cxx/1.63/ and vtkEnSightReader.h/1.28,
others were unchanged), and I can send you updated versions if
you prefer. In any case, you will need to recompile ParaView
(or parts of it at least).

This patch is also entered in ParaView's bug tracker as Bug #2971.
Polyhedra are converted to VTK_CONVEX_POINT_SET, which may not
be the best way to do things, but at least works.

For your info, I ran a 480,000 cell mesh generated by Comet Design
(a precursor of the meshing tools in Star CCM+ I believe, and
in any case a CD-Adapco product from 2-3 years ago) through
our CFD code's pre-processor, generating EnSight Gold files.
EnSight read it quite easily, while ParaView seemed to split it
into 12 million or so tetrahedra, leading to memory consumption
close to 2 Gb and making further operations impossible on a
32-bit machine, but it worked, and works fine for smaller cases
(except for artifacts at junctions between "classical" and
"polyhedral" elements appearing as false boundaries (which
dissapear when the tetralization filter is applied to the whole
set, and should not be an issue with data from Star-CD or Star-CCM+,
which is purely polyhedral (as far as I know).

Good luck !

	Yvan Fournier
	EDF R&D
	Fluid Mechanics, Power Generation, and Environment
	6 Quai Watier, 78400 Chatou, France

	yvan.fournier at edf.fr

P.S. I will be out of office most of the coming week, so if you have
any questions, please use this e-mail adress (yvan.fournier at free.fr),
as I cannot connect to my work adress from outside my office.

On Sun, 2006-11-19 at 12:00 -0500, f.braennstroem at gmx.de wrote:

> Date: Sun, 19 Nov 2006 10:05:59 +0100
> From: Fabian Braennstroem <f.braennstroem at gmx.de>
> Subject: [Paraview] Re: read starcd ensight mesh
> To: paraview at paraview.org
> Message-ID: <20061119090559.GA4887 at ddorf.de>
> Content-Type: text/plain; charset=us-ascii
> 
> Hi Berk,
> 
> thanks for your help. I exported the data in ensight format.
> I did not know that I can export in any vtk format!? The
> strange thing is that even with the old starcd 3.26 version
> the ensight exported tetra mesh makes the same problem. I'll
> try to convert it to a vtk format using python-vtk. Maybe
> that helps.
> Does anybody work on polyhedra support for vtk right now?
> Would be nice :-)
> 
> Thanks and Greetings!
> Fabian
> 
> 
> * On 18 Nov 2006 * Berk Geveci wrote:
> 
> > Hi Fabian,
> > 
> > Which file format are you using? Legacy VTK? It should work with the
> > tetra mesh. I will not work with the polyhedra mesh unfortunately. VTK
> > (hence ParaView) does not currently support polyhedral cells. Although
> > starccm+ visualization engine is built on top of VTK, they developed
> > the polyhedra support independently and are keeping it proprietary. I
> > hope that in the future we will have the opportunity to add polyhedra
> > support to VTK.
> > 
> > -berk
> > 
> > On 11/18/06, Fabian Braennstroem <f.braennstroem at gmx.de> wrote:
> > >Hi,
> > >
> > >sorry, but I want to ask again, if anybody has used paraview
> > >together with starccm+ or starcd version 4.00 or 3.26.
> > >Now, I got paraview 2.4.4 installed in my office on an 64bit
> > >machine.  I
> > >tried to load polyhedra results from starccm+ and starcd
> > >version 4.00. This time it does not crash, but nothing
> > >happens; I just see the variables. I tried to change the
> > >export to binary, but I got the same result. Later on I used
> > >a tetra mesh from the old version with the same result.
> > >
> > >I wonder, if nobody uses starcd together with paraview and
> > >has the same problems like I do!? I am the first at our
> > >company, so I hope that there will be more of us in the next
> > >months, but for that it has to work with starcd...
> > >
> > >Would be nice to get any hint!
> > >Greetings!
> > >Fabian
> > >
> > >--
> > >_______________________________________________
> > >ParaView mailing list
> > >ParaView at paraview.org
> > >http://www.paraview.org/mailman/listinfo/paraview
> > >
> > 
> > 
> > -- 
> >  Berk Geveci
> >  Kitware Inc.
> > 28 Corporate Drive
> > Clifton Park, NY, 12309
> > _______________________________________________
> > ParaView mailing list
> > ParaView at paraview.org
> > http://www.paraview.org/mailman/listinfo/paraview
> 
> -- 