<HTML>
<HEAD>
<TITLE>Re: [Paraview] holes in distributed polydata</TITLE>
</HEAD>
<BODY>
<FONT FACE="Calibri, Verdana, Helvetica, Arial"><SPAN STYLE='font-size:11pt'>I can’t find anything wrong other then there must be some sort of precision issue that is interfering with the decision of when to draw a polygon that is smaller than a pixel. I just checked in a change to IceT that prevents shifting around the projection matrix in single-display mode. That should fix the problem.<BR>
<BR>
-Ken<BR>
<BR>
<BR>
On 1/15/10 11:07 AM, "burlen" <<a href="burlen.loring@gmail.com">burlen.loring@gmail.com</a>> wrote:<BR>
<BR>
</SPAN></FONT><BLOCKQUOTE><FONT FACE="Calibri, Verdana, Helvetica, Arial"><SPAN STYLE='font-size:11pt'>Using the dataset below I can reproduce the issue on my workstation<BR>
using 4 processes and both hardware and mesa rendering.<BR>
<BR>
burlen wrote:<BR>
> Ken,<BR>
><BR>
> Here is a dataset that I am able to reproduce with:<BR>
> <a href="http://nashi-submaster.ucsd.edu/movies/PV/holes.tar.gz">http://nashi-submaster.ucsd.edu/movies/PV/holes.tar.gz</a><BR>
><BR>
> run with 96 procs. Set the view to +y. If the long holes aren't<BR>
> appearing in your view I found that smaller holes can show up by<BR>
> rotating the plane in small increments about either the x or z axis.<BR>
> These ones are pretty small so you have to look closely.<BR>
><BR>
> Burlen<BR>
><BR>
> Moreland, Kenneth wrote:<BR>
>> Burlen,<BR>
>><BR>
>> Is there any possible way you can send me some data or replicate the<BR>
>> problem with something like the Mandelbrot source? So far I have not<BR>
>> been able to replicate it exactly.<BR>
>><BR>
>> -Ken<BR>
>><BR>
>><BR>
>> On 1/5/10 2:34 PM, "burlen" <<a href="burlen.loring@gmail.com">burlen.loring@gmail.com</a>> wrote:<BR>
>><BR>
>> Ken,<BR>
>> Some bad news, this patch didn't solve the problem, the holes<BR>
>> returned<BR>
>> on my first run.<BR>
>> Burlen<BR>
>><BR>
>><BR>
>> Moreland, Kenneth wrote:<BR>
>> > That was not intended to be a solution, but rather a<BR>
>> diagnostic. My<BR>
>> > guess is that there are precision errors in the rasterization<BR>
>> when the<BR>
>> > viewport is shifted. Could you restore vtkIceTRenderManager and<BR>
>> try<BR>
>> > the attached patch to IceT?<BR>
>> ><BR>
>> > -Ken<BR>
>> ><BR>
>> ><BR>
>> > On 12/10/09 12:26 PM, "burlen" <<a href="burlen.loring@gmail.com">burlen.loring@gmail.com</a>> wrote:<BR>
>> ><BR>
>> > Hi Ken,<BR>
>> ><BR>
>> > it seems to have solved the problems. I say that with fingers<BR>
>> > crossed, I<BR>
>> > haven't seen holes any since your suggested changes, where before<BR>
>> > I was<BR>
>> > seeing them quite often, popping up from time to time.<BR>
>> ><BR>
>> > Burlen<BR>
>> ><BR>
>> > Moreland, Kenneth wrote:<BR>
>> > > Hmm. It is possible that the “floating viewport” feature of IceT<BR>
>> > could<BR>
>> > > be causing troubles with precision. Could you try adding<BR>
>> > ><BR>
>> > > icetDisable(ICET_FLOATING_VIEWPORT);<BR>
>> > ><BR>
>> > ><BR>
>> > > somewhere in the vtkIceTRenderManager::UpdateIceTContext()<BR>
>> method and<BR>
>> > > see if the problem goes away?<BR>
>> > ><BR>
>> > > -Ken<BR>
>> > ><BR>
>> > ><BR>
>> > > On 12/7/09 10:11 AM, "burlen" <<a href="burlen.loring@gmail.com">burlen.loring@gmail.com</a>> wrote:<BR>
>> > ><BR>
>> > > Hi Ken,<BR>
>> > > For that figure you mention I turned on "surface with edges" to<BR>
>> > > show the<BR>
>> > > cell size better. Sorry I can see how that could be<BR>
>> confusing. But<BR>
>> > > just<BR>
>> > > to clarify, there aren't actually any holes in the surface.<BR>
>> > ><BR>
>> > > Here is another zoom in of the same area where "surface with<BR>
>> > edges" is<BR>
>> > > off and you can see that there are no holes.<BR>
>> > > <a href="http://nashi-submaster.ucsd.edu/movies/PV/bug-zoom.png">http://nashi-submaster.ucsd.edu/movies/PV/bug-zoom.png</a><BR>
>> > ><BR>
>> > > Now I also have hit a case where after running through D3 I<BR>
>> got a<BR>
>> > hole<BR>
>> > > at the process boundary. this run had 80 processes, the surface<BR>
>> shown<BR>
>> > > has dimensions of 5.5 x 10 units with 1500 x 2727quads with side<BR>
>> > > 0.0036<BR>
>> > > units.<BR>
>> > > <a href="http://nashi-submaster.ucsd.edu/movies/PV/bug-d3.png">http://nashi-submaster.ucsd.edu/movies/PV/bug-d3.png</a><BR>
>> > ><BR>
>> > > I am only seeing this with the small quads and in parallel at<BR>
>> process<BR>
>> > > boundaries.<BR>
>> > ><BR>
>> > > Burlen<BR>
>> > ><BR>
>> > ><BR>
>> > > Moreland, Kenneth wrote:<BR>
>> > > > Burlen,<BR>
>> > > ><BR>
>> > > > For the zoom in, you say there are no holes/lines, but in the<BR>
>> > image I<BR>
>> > > > see a grid of lines. It looks like you have a bunch of little<BR>
>> quads<BR>
>> > > > with spacing in between them. Is this the case? If so, then<BR>
>> the<BR>
>> > > “hole”<BR>
>> > > > artifacts you see on the bottom of the screen are probably<BR>
>> simply<BR>
>> > > > aliasing artifacts. They are places where the pixel happens to<BR>
>> > align<BR>
>> > > > right where the gap is.<BR>
>> > > ><BR>
>> > > > I can’t think of an easy way around this (other than to<BR>
>> modify your<BR>
>> > > > data to remove the gaps, if that makes sense). Anti-aliasing<BR>
>> > > > techniques such as oversampling or smoothing would probably<BR>
>> fix the<BR>
>> > > > problem, but they would also break the parallel rendering so<BR>
>> > they are<BR>
>> > > > no good.<BR>
>> > > ><BR>
>> > > > -Ken<BR>
>> > > ><BR>
>> > > ><BR>
>> > > > On 12/5/09 12:18 AM, "burlen" <<a href="burlen.loring@gmail.com">burlen.loring@gmail.com</a>> wrote:<BR>
>> > > ><BR>
>> > > > its ugly but I get a lot better performance by splitting the<BR>
>> > work up<BR>
>> > > > dynamically with a small grain size. in the run shown below<BR>
>> > there are<BR>
>> > > > only 16 processes but there are a whole lot of process<BR>
>> boundaries.<BR>
>> > > ><BR>
>> > > > I was able to reproduce it on a second system today.<BR>
>> > > ><BR>
>> > > > these holes are pretty non-deterministic in where they show<BR>
>> up.<BR>
>> > > moving<BR>
>> > > > the camera they can show up in different places. Which makes<BR>
>> > sense if<BR>
>> > > > this is related to some parallel rendering/finite precision<BR>
>> issue<BR>
>> > > with<BR>
>> > > > all those process boundaries. The small size of the quads are<BR>
>> > also a<BR>
>> > > > factor, because I didn't ever notice it before when using<BR>
>> larger<BR>
>> > > > quads.<BR>
>> > > ><BR>
>> > > > I saved the data as a legacy file and opening it on my desktop<BR>
>> > > > there are<BR>
>> > > > no issues, so its definitely a parallel only issue. Also<BR>
>> running<BR>
>> > > > through<BR>
>> > > > D3 seems to fix it, but the issue may still be there because<BR>
>> > with the<BR>
>> > > > minimal number of process boundaries its much less likely to<BR>
>> > get the<BR>
>> > > > camera in just the right position.<BR>
>> > > ><BR>
>> > > > Berk Geveci wrote:<BR>
>> > > > > Ouch. That's very distributed :-) Does the problem go away<BR>
>> > when you<BR>
>> > > > > decrease the number of partitions?<BR>
>> > > > ><BR>
>> > > > > On Thu, Dec 3, 2009 at 10:55 AM, burlen<BR>
>> <<a href="burlen.loring@gmail.com">burlen.loring@gmail.com</a>><BR>
>> > > > wrote:<BR>
>> > > > ><BR>
>> > > > >> I'm seeing lines where the background shows through a<BR>
>> surface<BR>
>> > > > polydata of<BR>
>> > > > >> quads. When I zoom into the region to investigate the<BR>
>> holes are<BR>
>> > > > gone. Moving<BR>
>> > > > >> the image around the holes appear in different places. They<BR>
>> > > > depend on camera<BR>
>> > > > >> position. In this surface there are 2.5E6 quads. the<BR>
>> area is<BR>
>> > > > 10x16 units and<BR>
>> > > > >> the number of quads is 1250x2000. each quad has 0.008<BR>
>> units on a<BR>
>> > > > side. I<BR>
>> > > > >> hadn't seen the holes before going to this higher<BR>
>> resolution.<BR>
>> > > > It's likely<BR>
>> > > > >> that the hole is near a process boundary, in my polydata<BR>
>> filter<BR>
>> > > > each process<BR>
>> > > > >> adds his quads to his output polydata, in this run the<BR>
>> quads are<BR>
>> > > > distributed<BR>
>> > > > >> in strips of 512 as needed.<BR>
>> > > > >><BR>
>> > > > >> 3 holes/lines in bottom half of the image (black background<BR>
>> > > > shows through):<BR>
>> > > > >> <a href="http://nashi-submaster.ucsd.edu/movies/PV/bug.png">http://nashi-submaster.ucsd.edu/movies/PV/bug.png</a><BR>
>> > > > >><BR>
>> > > > >> zoom in no holes/lines:<BR>
>> > > > >> <a href="http://nashi-submaster.ucsd.edu/movies/PV/bug-zoom-2.png">http://nashi-submaster.ucsd.edu/movies/PV/bug-zoom-2.png</a><BR>
>> > > > >><BR>
>> > > > >> process boundaries (from process id filter):<BR>
>> > > > >> <a href="http://nashi-submaster.ucsd.edu/movies/PV/bug-procs.png">http://nashi-submaster.ucsd.edu/movies/PV/bug-procs.png</a><BR>
>> > > > >><BR>
>> > > > >> Should PV be able to handle a polydata distributed like<BR>
>> this?<BR>
>> > > > >><BR>
>> > > > >><BR>
>> > > > >><BR>
>> > > > >> _______________________________________________<BR>
>> > > > >> Powered by www.kitware.com<BR>
>> > > > >><BR>
>> > > > >> Visit other Kitware open-source projects at<BR>
>> > > > >> <a href="http://www.kitware.com/opensource/opensource.html">http://www.kitware.com/opensource/opensource.html</a><BR>
>> > > > >><BR>
>> > > > >> Please keep messages on-topic and check the ParaView<BR>
>> Wiki at:<BR>
>> > > > >> <a href="http://paraview.org/Wiki/ParaView">http://paraview.org/Wiki/ParaView</a><BR>
>> > > > >><BR>
>> > > > >> Follow this link to subscribe/unsubscribe:<BR>
>> > > > >> <a href="http://www.paraview.org/mailman/listinfo/paraview">http://www.paraview.org/mailman/listinfo/paraview</a><BR>
>> > > > >><BR>
>> > > > >><BR>
>> > > ><BR>
>> > > > _______________________________________________<BR>
>> > > > Powered by www.kitware.com<BR>
>> > > ><BR>
>> > > > Visit other Kitware open-source projects at<BR>
>> > > > <a href="http://www.kitware.com/opensource/opensource.html">http://www.kitware.com/opensource/opensource.html</a><BR>
>> > > ><BR>
>> > > > Please keep messages on-topic and check the ParaView Wiki at:<BR>
>> > > > <a href="http://paraview.org/Wiki/ParaView">http://paraview.org/Wiki/ParaView</a><BR>
>> > > ><BR>
>> > > > Follow this link to subscribe/unsubscribe:<BR>
>> > > > <a href="http://www.paraview.org/mailman/listinfo/paraview">http://www.paraview.org/mailman/listinfo/paraview</a><BR>
>> > > ><BR>
>> > > ><BR>
>> > > ><BR>
>> > > ><BR>
>> > > > **** Kenneth Moreland<BR>
>> > > > *** Sandia National Laboratories<BR>
>> > > > ***********<BR>
>> > > > *** *** *** email: <a href="kmorel@sandia.gov">kmorel@sandia.gov</a><BR>
>> > > > ** *** ** phone: (505) 844-8919<BR>
>> > > > *** web: <a href="http://www.cs.unm.edu/~kmorel">http://www.cs.unm.edu/~kmorel</a><BR>
>> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> > <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> > > <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> > > ><BR>
>> > ><BR>
>> > ><BR>
>> > ><BR>
>> > ><BR>
>> > ><BR>
>> > > **** Kenneth Moreland<BR>
>> > > *** Sandia National Laboratories<BR>
>> > > ***********<BR>
>> > > *** *** *** email: <a href="kmorel@sandia.gov">kmorel@sandia.gov</a><BR>
>> > > ** *** ** phone: (505) 844-8919<BR>
>> > > *** web: <a href="http://www.cs.unm.edu/~kmorel">http://www.cs.unm.edu/~kmorel</a><BR>
>> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> > <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> > ><BR>
>> ><BR>
>> ><BR>
>> ><BR>
>> ><BR>
>> ><BR>
>> > **** Kenneth Moreland<BR>
>> > *** Sandia National Laboratories<BR>
>> > ***********<BR>
>> > *** *** *** email: <a href="kmorel@sandia.gov">kmorel@sandia.gov</a><BR>
>> > ** *** ** phone: (505) 844-8919<BR>
>> > *** web: <a href="http://www.cs.unm.edu/~kmorel">http://www.cs.unm.edu/~kmorel</a><BR>
>> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>> ><BR>
>><BR>
>><BR>
>><BR>
>><BR>
>> **** Kenneth Moreland<BR>
>> *** Sandia National Laboratories<BR>
>> ***********<BR>
>> *** *** *** email: <a href="kmorel@sandia.gov">kmorel@sandia.gov</a><BR>
>> ** *** ** phone: (505) 844-8919<BR>
>> *** web: <a href="http://www.cs.unm.edu/~kmorel">http://www.cs.unm.edu/~kmorel</a> <<a href="http://www.cs.unm.edu/%7Ekmorel">http://www.cs.unm.edu/%7Ekmorel</a>><BR>
>><BR>
><BR>
<BR>
<BR>
<BR>
</SPAN></FONT></BLOCKQUOTE><FONT FACE="Calibri, Verdana, Helvetica, Arial"><SPAN STYLE='font-size:11pt'><BR>
</SPAN></FONT><FONT SIZE="2"><FONT FACE="Consolas, Courier New, Courier"><SPAN STYLE='font-size:10pt'><BR>
**** Kenneth Moreland<BR>
*** Sandia National Laboratories<BR>
*********** <BR>
*** *** *** email: <a href="kmorel@sandia.gov">kmorel@sandia.gov</a><BR>
** *** ** phone: (505) 844-8919<BR>
*** web: <a href="http://www.cs.unm.edu/~kmorel">http://www.cs.unm.edu/~kmorel</a><BR>
</SPAN></FONT></FONT><FONT FACE="Calibri, Verdana, Helvetica, Arial"><SPAN STYLE='font-size:11pt'><BR>
</SPAN></FONT>
</BODY>
</HTML>