[Paraview] ParaView 3.8.0-RC1 Contour problem

[Moreland, Kenneth]

John,

I believe the generate normals option requests a layer of ghost cells.  Are you sure your reader is responding properly to that?

-Ken


On 4/16/10 9:53 AM, "John Patchett" <patchett2002 at gmail.com> wrote:

Good Morning,
  I have encountered a problem with the contour filter with computer
normals set on.

My configuration is a Red Hat Enterprise Linux Server release 5.4
(Tikanga) server running 32 pvserver processes, openmpi 1.4.1, Mesa
7.6 (OSMesa).
a Fedora release 11 (Leonidas) workstation running the client.
Both are CVS ParaView-3-8 builds updated after the calculator fix was
applied (very fast thank you).

I load a dataset with 2 vectors and a scalar all floats using VPIC reader.
I run 1 of the vectors through the calculator to compute magnitude.
  Result array has range [.250244, .572473]
I then run that result into a contour, which will do 1 of 2 things:
If I un-check compute normals : contour works as expected (attached).
If compute normals is checked :
   it takes a long time to produce a bad contour (attached) with
errant lines OR the server segfaults

If the server segfaults it is worthy of note that the information tab
of the Calculator then reports the Result array as having
range[0.0,3.57...+38]
If I delete the contour and the calculator, create a new calculator,
calculate magnitude it reports range as [0.0,3.57...+38]

Seems like a problem ...

I will work at finding a more simple example...

Thanks,
--John Patchett.



   ****      Kenneth Moreland
    ***      Sandia National Laboratories
***********
*** *** ***  email: kmorel at sandia.gov
**  ***  **  phone: (505) 844-8919
    ***      web:   http://www.cs.unm.edu/~kmorel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20100416/c76d4ad7/attachment.htm>