[Paraview] VALUES or THRESHOLDS selections

[David E DeMarle]

For Thresholds selections it runs vtkExtractSelectedThresholds, which
has the simpler implementation. It simply iterates through the data
looking at the values.

For Value selections, it ends up running vtkExtractSelectedIDs. In
that filter, a copy of the data array is first sorted by value, and
then searched. The copy is supposed to stick around for the later
calls, which makes sorting it worthwhile in the long run. I assume the
copy is not sticking around for some reason in your case, or less
likely that the sort algorithm itself is running out of memory.

David E DeMarle
Kitware, Inc.
R&D Engineer
28 Corporate Drive
Clifton Park, NY 12065-8662
Phone: 518-371-3971 x109



On Thu, Jul 2, 2009 at 5:36 PM, Jean Favre<jfavre at cscs.ch> wrote:
> I have a filter doing cell selection from an array of material ids. The
> usage pattern is to select a single material id, i.e. a single number.
>
> To debug and check my filter's output, I ran the same with the Selection
> Inspector. However, ParaView only enables THRESHOLDS selection in the
> Selection Inspector. Is there a reason the alternative, selection by
> VALUES (described in a very good article on page 3 of The Source, April
> 2009) is not available?
>
> I am running into a problem with my filter's CELL selection. If I use
> THRESHOLDS, and set the SelectionList to an array of size 2 (both values
> being equal since I want to select a single material ID), the filter
> executes really fast and gives the correct result. If instead, I use
> VALUES, and set the SelectionList to an array of size 1, execution is
> extremely slow. I measured
> 2.4 secs for 27,000 cells
> 253 secs for 270,000 cells
> and a seg fault for 2,700,000 cells.
>
> Every time I tried to stop execution, I am inside the QuickSort class.
> Are there issues with that? Isn't the complexity of the sorting supposed
> to be n*log(n)? that is not what I see when I get execution times from
> 2.4 to 253 secs when going from 27000 to 270000 cells.
>
> Note that I get the correct results with VALUE-selection (even though
> very slowly), identical to the selection by THRESHOLDS (except for the
> last case with the seg fault)
>
> any experience from other users?
>
> Jean--
> Swiss National Supercomputer Center
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://www.paraview.org/mailman/listinfo/paraview
>