Skip to Content.
Sympa Menu

cgal-discuss - Re: [cgal-discuss] performance of Delaunay Triangulation of random

Subject: CGAL users discussion list

List archive

Re: [cgal-discuss] performance of Delaunay Triangulation of random


Chronological Thread 
  • From: Sylvain Pion <>
  • To:
  • Subject: Re: [cgal-discuss] performance of Delaunay Triangulation of random
  • Date: Sat, 20 Dec 2008 17:58:55 +0100
  • Organization: INRIA

Hi Ben,

Ben Supnik wrote:
[...]

So...perhaps experiment with different ordering of the points and time the results?

The order of insertion is indeed important.
Fortunately, the insert() function taking an iterator range takes
care of this automagically (using a BRIO, aka random shuffle and
spatial sorting). So, it's not the problem here.

The other issue would be (and this is not as much a factor with DT as other parts of CGAL) to check your optimizer and #define settings carefully...you'll want to make sure that all condition checking is compiled out and that the optimizer is inlining...from what I can tell GCC on OS X won't inline if the optimizer is off, which makes virtually any template code under-perform.

This is most probably the problem here:
I get 4.11s on my Mac with -O2 -DNDEBUG, and 42.31s otherwise.

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

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature




Archive powered by MHonArc 2.6.16.

Top of Page