Skip to Content.
Sympa Menu

cgal-discuss - Re: [cgal-discuss] about running time related to CGAL triangiulation

Subject: CGAL users discussion list

List archive

Re: [cgal-discuss] about running time related to CGAL triangiulation


Chronological Thread 
  • From: Andreas Fabri <>
  • To:
  • Subject: Re: [cgal-discuss] about running time related to CGAL triangiulation
  • Date: Sun, 01 Feb 2009 11:07:02 +0100

Kwok Jasper wrote:
Thank you for your reply.

I have already optimized the compiler and made sure that it does not have the debugging option selected before.
May I ask is it possible that it is related to the memory in my PC?

Currently, I am using Window XP and have only less than 10 GB left in my C drive.
May I ask if the running may be improved through increasing the memory left?
If it may help, how many GB shall I increase?

Hi Jasper,

This has nothing to do with the drive size, only with the RAM,
but even this should not pose any problem for 362272 points.

As Monique pointed out, you should either insert all points at
the same time. If you must insert them one after the other,
you should use the Triangulation_hierarchy_2. There is an
example in CGAL-3.4/examples/Triangulation_2 illustrating it.

Another thing is the point set itself. If it happens to be a
grid the algorithm may be slow. We work on this.

Finally, can you put the data set on a server so that we can
try to reproduce the timings.

best regards,

andreas


Thanks



> Date: Sun, 1 Feb 2009 09:43:45 +0100
> From:

> To:

> Subject: Re: [cgal-discuss] about running time related to CGAL triangiulation
>
> Kwok Jasper wrote:
> > Excuse me.
> >
> > I have tried to use the 3D Delaunay Triangulation in CGAL in Visual C++
> > with the following declaration.
> >
> > typedef CGAL::Exact_predicates_inexact_constructions_kernel K1;
> > typedef CGAL::Delaunay_triangulation_3<K1> Triangulation;
>
> Also, use t.insert(begin(),end()). It uses spatial sorting, which makes
> the construction much faster. See the manual.
>
> > After I have set the compiler optimization in Visual C++,
> > when I insert 362272 points into the triangulation, it takes more than
> > 20 minutes
> > and after the insertion, when I use the iterator such as the
> > Finite_facets_iterator, it also takes a lot of time.< br>
>
> > May I ask if the running time is normal?
>
> no, it is not normal
>
> > Or in fact it should be faster in the normal case?
>
> It should be much faster (a few seconds, depending on your computer)
> Be careful to compile your code with the appropriate options (optimized
> and without debugging options).
>
> > Thank you very much for your attention
> >
> > ------------------------------------------------------------------------
> > ?? Windows Liveā„¢ Photos ??????? ?????????
> > <http://www.microsoft.com/windows/windowslive/events.aspx>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss

------------------------------------------------------------------------
??MSN Buddy??????d? ??MSN Buddy???! <http://buddy.hk.msn.com/>




Archive powered by MHonArc 2.6.16.

Top of Page