CGAL users discussion list

[Sagar Gandhi <>]

I understood my mistake. I was using Epick and was also introducing the custom functions - which don't take care of precision.

So issue is resolved. If I use Simple_cartesian<double>, time difference is not considerable. But I am going ahead with Epick, and using CGAL's predicates, as all I have been working on is timing.

I am very curious in understanding - "How exactly CGAL takes care of precision issue?" I would love to have a code reference(s) as well as documentations.

Please point me in the right direction.

Regards,

S. Gandhi

On Mon, Jan 18, 2016 at 12:05 PM, Marc Glisse <> wrote:

On Sun, 17 Jan 2016, s_gandhi wrote:

Please find the code in the attached file named "Custom_vs_CGAL_speed.cpp".
Also find the screenshot of timings on my machine, named "cgal_speeds_1.png"

I am using EPIC,
so should I implement the stuff myself? It gives me better timings.

And if I compare any CGAL's examples with my implementation, I (am) will be
faster.

You should benchmark against Simple_cartesian<double>, not Epick, since you don't seem to aim for exactness.

Please read the link I gave in the previous email and understand that your code will sometimes give the wrong result, which is why we cannot use it in Epick.

--
Marc Glisse

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