[vtk-developers] TestGaussianBlurPass failures

Francois Bertel francois.bertel at kitware.com
Sun Jan 3 12:45:50 EST 2010


I committed a fix for the second case, when the framebuffer format is
not supported:

cvs commit -m "BUG:Attempt to fix the case where FBO extension is
supported by the requested FBO format is not as for
SINTEF.LabMAC.OSX/VTKCvs_CMake262Carbon_gcc401_Qt443."
vtkGaussianBlurPass.cxx vtkGaussianBlurPass.h
Committer: Francois Bertel <francois.bertel at kitware.com>
Commit to directory "ParaView3/VTK/Rendering" lies in external project VTK
Committer: Francois Bertel <francois.bertel at kitware.com>
/cvsroot/ParaView3/ParaView3/VTK/Rendering/vtkGaussianBlurPass.cxx,v
<--  vtkGaussianBlurPass.cxx
new revision: 1.7; previous revision: 1.6
/cvsroot/ParaView3/ParaView3/VTK/Rendering/vtkGaussianBlurPass.h,v
<--  vtkGaussianBlurPass.h
new revision: 1.5; previous revision: 1.4

Bill, "linux fedora 9 system" is not a relevant information for this
kind of failure. The output of the C++ test LoadOpenGLExtension is.
What is it for you?


For the floating-point exception case, it's probably happening in the
nVidia driver itself. I don't see it on my nVidia card on Ubuntu 9.10
x86_64 (driver 185.18.36, from August 21, 2009). I would recommend to
update the nVidia driver on amber* as the driver version displayed on
the output of LoadOpenGLExtension is 173.14.12 (driver from July 30,
2008).

On Sun, Jan 3, 2010 at 6:20 PM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> Dave,
>
> Are you running the complete command line:
> /home/kitware/Dashboards/MyTests/VTK-gcc/bin/RenderingCxxTests
> TestGaussianBlurPass -D /home/kitware/Dashboards/MyTests/VTKData -T
> /home/kitware/Dashboards/MyTests/VTK-gcc/Testing/Temporary -V
> Baseline/Rendering/TestGaussianBlurPass.png
>
> Maybe there is a problem with the regression testing?
>
> Bill
>
> On Sun, Jan 3, 2010 at 12:15 PM, Dave Partyka <dave.partyka at kitware.com> wrote:
>> Yeah, on amber1.
>>
>> On Sun, Jan 3, 2010 at 12:10 PM, Bill Lorensen <bill.lorensen at gmail.com>
>> wrote:
>>>
>>> Dave,
>>>
>>> Did you run on one of the known failing platforms?
>>>
>>> Bill
>>>
>>> On Sun, Jan 3, 2010 at 11:50 AM, Dave Partyka <dave.partyka at kitware.com>
>>> wrote:
>>> > I tried to run it from gdb yesterday but could not recreate the
>>> > exception.
>>> > Though I was running it remotely over ssh. I am going to try it running
>>> > it
>>> > locally today.
>>> >
>>> > On Sun, Jan 3, 2010 at 11:27 AM, Bill Lorensen <bill.lorensen at gmail.com>
>>> > wrote:
>>> >>
>>> >> Francois,
>>> >>
>>> >> TestGaussianBlurPass is failing on several platforms.
>>> >>
>>> >>
>>> >> http://www.cdash.org/CDash/testSummary.php?project=11&name=TestGaussianBlurPass&date=2010-01-03
>>> >>
>>> >> There are two modes of failure:
>>> >>
>>> >> 1) Floating point exceptions:
>>> >> http://www.cdash.org/CDash/testDetails.php?test=39491121&build=506514
>>> >> Someone needs to run the test inside a debugger and find the source of
>>> >> the floating point divide.
>>> >>
>>> >> 2) Framebuffer not supported:
>>> >> http://www.cdash.org/CDash/testDetails.php?test=40287643&build=506967
>>> >> I also see this failure on my linux fedora 9 system. It seems that
>>> >> there needs to be better checking for whether or not the frame buffer
>>> >> object is properly initialized.
>>> >>
>>> >> Thanks,
>>> >>
>>> >> Bill
>>> >> _______________________________________________
>>> >> Powered by www.kitware.com
>>> >>
>>> >> Visit other Kitware open-source projects at
>>> >> http://www.kitware.com/opensource/opensource.html
>>> >>
>>> >> Follow this link to subscribe/unsubscribe:
>>> >> http://www.vtk.org/mailman/listinfo/vtk-developers
>>> >>
>>> >
>>> >
>>
>>
>



-- 
François Bertel, PhD  | Kitware Inc. Suite 204
1 (518) 371 3971 x113 | 28 Corporate Drive
                      | Clifton Park NY 12065, USA



More information about the vtk-developers mailing list