[vtkusers] Corrupted OpenGL graphics in render window for prism ia64

[Remik Ziemlinski]

Okay, I ran 'make Experimental', but after 6 hours it kept spinning the 
CPU at 99%.  Is that normal behaviour?  The output wasn't verbose, only 
indicated by dots and stopping after ~151,000k of "Output".  Within 
ccmake, I used the site name prism1.gfdl.noaa.gov.

Do I need to setup any additional software to submit the test?

Remik

Mathieu Malaterre wrote:
> Remik Ziemlinski wrote:
> 
>> Hi all,
>>   I tried compiling vtk 4.4 as well as the cvs version on a Prism vis. 
>> system with ATI Firegl X3 cards using GCC 3.2.3, however all the 
>> renderings are corrupted with random patches along the perimeter, plus 
>> trails left by props when interacted with.  A Mesa equivalent build 
>> doesn't have such problems, nor do other OpenGL built binaries, such 
>> as SPECviewperf, Qt demos, or the like.  Has anyone seen this before?
>>
>> I can see from the vtk Dashboard that 
>> prism.sci.utah.edu/Linux-sgi-ia64-gcc3.2 has 833 tests as N/A (these 
>> are precisely the tests that fail for our Prism).  Is this being 
>> looked into, and are there any hints as to how to resolve it?
>>
>> uname -a gives:
>> Linux prism1 2.4.21-sgi306r6 #1 SMP Sat Jun 18 14:43:35 PDT 2005 ia64 
>> ia64 ia64 GNU/Linux
>>
> 
> N/A are for tests that are not run. 
> prism.sci.utah.edu/Linux-sgi-ia64-gcc3.2 does not wrap VTK in neither 
> Tcl nor python therefore has a very limited subset of test running.
> 
> If you have an internet connection from the machine you are having 
> problem. I would suggest you submit a dashboard, something like:
> 
> $ make Experimental (from your build tree)
> 
> Let me know if you decide to do it (specify the name of the machine)
> 
> Thanks,
> Mathieu

-- 
Remik Ziemlinski                        Raytheon Company
NOAA/Geophysical Fluid Dynamics Lab.    Remik.Ziemlinski at noaa d0t gov
P.O. Box 308                            1.609.452.6500 ext. 6977
Princeton, NJ 08542 USA                 1.609.987.5063 fax