[vtk-developers] parallel surface lic broken in opengl 2

[Ken Martin]

I believe there are 4 main LICs in VTK OpenGL2, serial, parallel, serial +
composite_dataset, parallel + composite_dataset. I believe the first three
work and I think trashed pixels have been fixed recently, the VTK tests all
pass. yay!

The parallel + composite_dataset case is known to be broken, more
specifically the class does not exist. PV needs this class in parallel and
I think it falls back to the serial implementation as that class has been
written which is why you see boundary issues.

My gut feeling is that instead of writing that class, the LIC code should
be reworked as a render pass that gets assigned to mappers. That way adding
lic to a mapper is fairly easy. The way I converted the original code to
OpenGL2 combines the LIC complexity into the middle of the mapper
complexity which is a bad mix especially for the parallel +
composite_dataset case.

Maybe it is easier than I think, but currently I suspect it would take a
few weeks of dedicated effort which is why it is still in the queue.




On Tue, Mar 1, 2016 at 12:50 PM, Burlen Loring <burlen.loring at gmail.com>
wrote:

> Hi All,
>
> Writing to report that in latest version of ParaView the surface lic is
> broken with the OpenGL2 back end. I experimented with it using PV5 an
> nvidia gtx 980 with fairly recent driver, and saw a number of issues,
> including completely "trashed" pixels and when rendering in parallel one
> can see "lines" at process boundaries as if the algorithm is serial.
>
> These regression are a shame because it took quite a lot of effort to
> correctly parallelize it, and make it work reliably across a wide spectrum
> of os,gl, and device variations. There is an explanation of the parallel
> algorithm linked below. Hopefully those who are working on porting to
> opengl 2 are aware of and will take a look at these regressions.
>
> Note, that all of the surface LIC classes live VTK but because IceT and
> the compositing classes all live in ParaView, the ctest for parallel
> operation are in ParaView. So another question is, given the importance of
> parallel rendering, should IceT and parallel rendering capability be
> included in VTK?
>
> Burlen
>
>
> https://github.com/burlen/astronum2014/blob/master/paper/loring-astronum-2014.pdf
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Search the list archives at: http://markmail.org/search/?q=vtk-developers
>
> Follow this link to subscribe/unsubscribe:
> http://public.kitware.com/mailman/listinfo/vtk-developers
>
>


-- 
Ken Martin PhD
Chairman & CFO
Kitware Inc.
28 Corporate Drive
Clifton Park NY 12065
518 371 3971

This communication, including all attachments, contains confidential and
legally privileged information, and it is intended only for the use of the
addressee.  Access to this email by anyone else is unauthorized. If you are
not the intended recipient, any disclosure, copying, distribution or any
action taken in reliance on it is prohibited and may be unlawful. If you
received this communication in error please notify us immediately and
destroy the original message.  Thank you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20160301/3e7118b4/attachment.html>