[vtkusers] VTK pipeline & rendering strategies

Andrea Gavana andrea.gavana at gmail.com
Tue Nov 29 00:21:47 EST 2016


Hi Bill,

Thank you for your answer. I think I may try something like that, although
I am more interested in cells rather than points. As in many problems in
CFD, the cell scalars are the important part in the visualization, and
ignite were for me I could easily obliterate the points - their only
function is to define the cell boundaries, I'm not using them for anything.

vtkDataSetSurfaceFilter may not be a good idea anymore - as mentioned by
David. In any case, if I have to loop through all the cells and setting
their scalars based on the original cell ID it's going to be slow. Python
is not famous for the speed of its for loops :-) .

But beside the filters, I'm still puzzled by the actual call to Render that
is taking up so much time. And the grid I have is not particularly big by
industry standards, some grids can easily go up to millions of cells.

Andrea.

On Mon, 28 Nov 2016 at 21:46, Bill Lorensen <bill.lorensen at gmail.com> wrote:

> Here is a half-baked idea (that I have not tested). And I may have
> completely misunderstood your problem...
>
> You can delay the application of the scalars until after
> DataSetSurfaceFilter. That filter has an option to pass the original
> cell and/or point ids. So after an update to the
> DataSetToSufaceFilter, create your scalars and populate the new points
> with the scalars at the original point id. It should be pretty fast
> since you will only be settng scalar values at surface points and not
> interior points. How many triangles are produced by
> DataSetToSufaceFilter?
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtkusers/attachments/20161129/997ffe41/attachment.html>


More information about the vtkusers mailing list