CGAL users discussion list

[400555 <>]

Thank you for answer !

Manuel, I checked out hidden vertices number - no any hidden vertices at all.

Also I read manual but still don't understand when vertex could become hidden, could anyone explain it in easy words or show an example.

Manuel wrote: Also maybe you profit from some optimizations in Delaunay triangulation..

Manuel, what kind of optimization do you mean? Inside of CGAL engine or on my side? On my own side I don't do any optimization. All what I do - just insert list of points one by one and apply info for each of them.

Sylvain wrote: The lack of static filters for the power tests...

I don't get what does it mean? And is it possible to avoid it?

Sorry for many questions.

Thank you for answerers.

On Thu, Aug 13, 2009 at 11:14 AM, Sylvain Pion <> wrote:

The lack of static filters for the power tests probably explains
a good share of this difference.

Manuel Caroli wrote:

Hi,

the weighted Delaunay triangulation has some additional overhead due to points that can be hidden. If you have a lot of hidden points this might have a considerable effect on running times. Also maybe you profit from some optimizations in Delaunay triangulation but not in the weighted Delaunay triangulation.

Without more information about your problem and your code and I cannot tell you more than the above speculations.

best

Manuel


400555 wrote:

Hello !

Could you say your opinion about:

I tried to compare performance between using CGAL Delaunay and Regular triangulation with weighted points.

I rebuilt it 200.000 times for moving points (I know that rebuilding is not best option) and Delaunay takes 17 sec but weighted one - 64 seconds.
I don't understand why, difference is only that weighted one considers about weight which is included in determinant (which defines "in circle" criteria), please correct me if I am wrong.
So I was not expected so much time difference between them.

Do you know why it happen?

Thank you.

--
Sylvain Pion
INRIA Sophia-Antipolis
Geometrica Project-Team
CGAL, http://cgal.org/

--
You are currently subscribed to cgal-discuss.
To unsubscribe or access the archives, go to
https://lists-sop.inria.fr/wws/info/cgal-discuss