[Paraview] Triangulation of Mesh

[Gamble, Kyle]

UNRESTRICTED | ILLIMITÉ

 

Hello All,

 

I am trying to understand why when I use show with surface with edges I am getting bizarre results.  I have used a program called cubit to create an exodus file of my mesh on my local machine with quad8 (quadratic) elements.  To explain the behaviour, I have created a simple case called junk. When I open this exodus file (without results) into Paraview  on my local machine I obtain the mesh shown in the attached image JunkParaviewLocal.jpg.  

 

Now, I run my simulations remotely to another computer used by someone else because it is faster than the computer I currently use locally.  This is done through linux and ssh. If I secure copy the same exodus file to the remote machine and open it in Paraview on the remote machine I obtain the mesh shown in ParaviewJunkRemote.jpg, which is obviously the incorrect representation of the mesh.  I've run the simulation assuming the triangulation was just a graphics issue but the results were based upon the triangular mesh which is incorrect.

 

Why does paraview triangulize the mesh when running remotely?  I ask this because if I run the simulation using the mesh from this exodus file, I obtain results based on the triangular mesh rather than the quadrilateral mesh I designed, and this leads to incorrect results.

 

I thought it may be an issue with the copying of the file so I went to the remote machine with the exodus file on a usb key and copied the file to the hard drive.  I tried opening the file locally on the remote machine and when I click surface with edges, Paraview immediately exits with a seg fault. 

 

Why would paraview exit with a seg fault when clicking surface with edges, when the exodus file opens fine on my local machine?  In addition other exodus files open perfectly fine in Paraview on the remote machine indicating to me there is nothing wrong with the installation of Paraview either.

 

Again this seems like unusual behaviour and I need to be able to run my simulations on the remote computer on the correct mesh.  I do not experience this behaviour using first order quad4 elements.

 

If anyone has any ideas on why this behaviour is observed and how to correct it would be greatly appreciated.

 

Thank you,

 

Kyle Gamble

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20130531/9e34b411/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: JunkParaviewLocal.jpg
Type: image/jpeg
Size: 79201 bytes
Desc: JunkParaviewLocal.jpg
URL: <http://www.paraview.org/pipermail/paraview/attachments/20130531/9e34b411/attachment-0002.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ParaviewJunkRemote.jpg
Type: image/jpeg
Size: 196623 bytes
Desc: ParaviewJunkRemote.jpg
URL: <http://www.paraview.org/pipermail/paraview/attachments/20130531/9e34b411/attachment-0003.jpg>