[Paraview-developers] VTK and ParaView with OSMesa
Burlen Loring
burlen.loring at gmail.com
Wed Nov 27 14:31:14 EST 2013
Thank you guys for taking a look at the wiki page.
> Any thoughts on the remaining failures? I'ld like to fix or if
> necessary suppress whatever remains and then promote them to the
> expected section of the VTK dashboard.
awesome! for the PV dashboard with llvmpipe 9.2.2 the tests failing
aren't rendering related which is very nice. for the VTK dashboards
there's a mix of issues which appear to be Mesa bugs. If we have time we
could report them to mesa developers. Here's the list from the rocky
nightly dashboard.
I think this is a Mesa bug but could be fixed by adding a baseline.
vtkViewsInfovisCxx-TestConeLayoutStrategy
<http://open.cdash.org/testSummary.php?project=11&name=vtkViewsInfovisCxx-TestConeLayoutStrategy&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220589417&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220589417&build=3115094>
630ms Completed (Failed)
This is the only one that's not a Mesa bug. the test needs to check an
IsSupported method and skip if not. see test output for more info.
vtkRenderingParallelCxx-MPI-TestPShadowMapPass
<http://open.cdash.org/testSummary.php?project=11&name=vtkRenderingParallelCxx-MPI-TestPShadowMapPass&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220589340&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220589340&build=3115094> 2s
490ms Completed (Failed)
in these mesa crashes, probably just disable them.
vtkFiltersHybridPython-largeImageOffset
<http://open.cdash.org/testSummary.php?project=11&name=vtkFiltersHybridPython-largeImageOffset&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220588466&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220588466&build=3115094> 1s
410ms Completed (OTHER_FAULT)
vtkFiltersHybridPython-largeImageParallel
<http://open.cdash.org/testSummary.php?project=11&name=vtkFiltersHybridPython-largeImageParallel&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220588467&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220588467&build=3115094> 1s
370ms Completed (OTHER_FAULT)
vtkRenderingVolumePython-cursor3D
<http://open.cdash.org/testSummary.php?project=11&name=vtkRenderingVolumePython-cursor3D&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220589385&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220589385&build=3115094> 1s
300ms Completed (OTHER_FAULT)
these are more general mesa rendering bugs, not sure if they could be
tweaked to work or just need to be disabled.
[flag] vtkRenderingCorePython-TestOpacityVectors
<http://open.cdash.org/testSummary.php?project=11&name=vtkRenderingCorePython-TestOpacityVectors&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220589164&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220589164&build=3115094> 1s
280ms Completed (Failed)
vtkRenderingOpenGLCxx-TestTransformCoordinateUseDouble
<http://open.cdash.org/testSummary.php?project=11&name=vtkRenderingOpenGLCxx-TestTransformCoordinateUseDouble&date=2013-11-26>
Failed
<http://open.cdash.org/testDetails.php?test=220589304&build=3115094>
Passed
<http://open.cdash.org/testDetails.php?test=220589304&build=3115094>
200ms Completed (Failed)
Hope this helps,
Burlen
On 11/26/2013 04:54 AM, David E DeMarle wrote:
> Thanks Burlen! Nice work as always.
>
> Any thoughts on the remaining failures? I'ld like to fix or if
> necessary suppress whatever remains and then promote them to the
> expected section of the VTK dashboard.
>
> David E DeMarle
> Kitware, Inc.
> R&D Engineer
> 21 Corporate Drive
> Clifton Park, NY 12065-8662
> Phone: 518-881-4909
>
>
> On Sat, Nov 23, 2013 at 5:51 PM, Utkarsh Ayachit
> <utkarsh.ayachit at kitware.com <mailto:utkarsh.ayachit at kitware.com>> wrote:
>
> Great post, Burlen.
>
> Utkarsh
>
> On Mon, Nov 18, 2013 at 1:24 PM, Burlen Loring
> <burlen.loring at gmail.com <mailto:burlen.loring at gmail.com>> wrote:
> > Hi All,
> >
> > VTK 6.1 will take full advantage of OpenGL features in OSMesa.
> In the past
> > this hadn't been the case, a number of key algorithms or classes
> that
> > algorithms depended on were explicitly disabled when OSMesa was
> detected.
> > I've been investigating use of the new 9.2.2 Mesa release in
> which OSMesa
> > can take advantage of the VMWare's Gallium llvmpipe back-end.
> The advantages
> > of this back-end over the classic OSMesa back-end are better OpenGL
> > extension support, threading, and use of LLVM/clang for shader
> compilation.
> > The llvmpipe backend is quite a bit faster on shader heavy
> algorithms and
> > supports Surface LIC while the classic back-end does not. In
> the recent
> > Mesa 9.2.2 release it finally works flawlessly with pvbatch.
> >
> > The Mesa 9.2.2 release configured with the Gallium llvmpipe
> OSMesa state
> > tracker should be the recommended Mesa release for off-screen
> rendering in
> > VTK and ParaView and I've updated the wiki with build
> instructions and some
> > rendering benchmarks.
> > http://paraview.org/Wiki/ParaView/ParaView_And_Mesa_3D
> > I've left a link to the old page, if there are no objections I'd
> like to
> > remove it.
> >
> > I've set up ParaView and VTK dashboards for the new 9.2.2 Mesa
> release
> >
> http://open.cdash.org/index.php?project=VTK&filtercount=1&showfilters=1&field1=site/string&compare1=63&value1=rocky
> >
> http://open.cdash.org/index.php?project=ParaView&filtercount=1&showfilters=1&field1=site/string&compare1=63&value1=rocky
> >
> > If you guys see anything wrong with it , please let me know.
> >
> > Thanks
> > Burlen
> > _______________________________________________
> > Paraview-developers mailing list
> > Paraview-developers at paraview.org
> <mailto:Paraview-developers at paraview.org>
> > http://public.kitware.com/mailman/listinfo/paraview-developers
> _______________________________________________
> Paraview-developers mailing list
> Paraview-developers at paraview.org
> <mailto:Paraview-developers at paraview.org>
> http://public.kitware.com/mailman/listinfo/paraview-developers
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview-developers/attachments/20131127/b64595f9/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: flag.png
Type: image/png
Size: 3462 bytes
Desc: not available
URL: <http://public.kitware.com/pipermail/paraview-developers/attachments/20131127/b64595f9/attachment-0001.png>
More information about the Paraview-developers
mailing list