CGAL users discussion list

[<>]

I had the same issue. I needed to compute the convex hull of a set of points on a sphere.
 
This may not work for you, but I ended up using the triangulation routine after which I only consider infinite cells.
 
I can't quote exact numbers, but the convex hull algorithm performed very poorly on my machine (in the order of hundreds of ms for perhaps 100 points?). I tried qhull, which did the same number of points in mere milliseconds.
 
The triangulation takes about 20ms for 50-100 points and what's more, it supports insertion as well as deletion of points.

> To:
> From:
> Date: Mon, 24 Nov 2008 17:31:19 +0100
> Subject: [cgal-discuss] Convex_hull_d performance
>
> Dear all,
>
> I am using CGAL::Convex_hull_d to compute the convex hull of a small set of
> points in some n-dimensional space. Here, small means 2*n points, where n is
> the number of dimensions of the space (3 <= n <= 6). Unfortunately, compute the
> convex hull is the bottleneck of my implementation. Actually, the program
> spends 87% of the time calling CGAL::Convex_hull_d.insert().
>
> Today, my implementation takes several minutes to process simple data sets and
> I want to make it run faster. My question is: There is some trick that I can
> use to improve CGAL’s performance assuming a small number of input points? I
> believe that the bottleneck inside CGAL is memory allocation.--
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss

---

Vanaf nu heb je je vrienden overal bij! Messenger voor Mobile