Thread Subject: Delaunay

Hi I was testing the new delaunay triangulator from the CGAL library available from R2009a.

These are my results:
The new one is faster than delaunayn of 3-4 factors on uniform dataset.
Instead, for sparse dataset is terribly slower, for a medium number of points I had to stop the test before it finished.

Test were run on 2-D 3-D models with increasing number of points up to 1e5.

Have you some explanation on that? and what is your opinon?

Thanks in advance

Hi Luigi,

For randomly distributed datapoints the new triangulation functions are 3-4
times faster.
For highly-degenerate datasets, such as datapoints arranged in a regular
grid, the performance is worse, but the code is robust in tringulating this
data.
We are aware of this issue and we actively working on addressing it.
Thank you for your feedback.

Regards,

Damian

"Luigi Giaccari" <giaccariluigi@msn.com> wrote in message
news:guh53h$22b$1@fred.mathworks.com...
> Hi I was testing the new delaunay triangulator from the CGAL library
> available from R2009a.
>
> These are my results:
> The new one is faster than delaunayn of 3-4 factors on uniform dataset.
> Instead, for sparse dataset is terribly slower, for a medium number of
> points I had to stop the test before it finished.
>
> Test were run on 2-D 3-D models with increasing number of points up to
> 1e5.
>
> Have you some explanation on that? and what is your opinon?
>
> Thanks in advance