CGAL users discussion list

[Raphael Grimm <>]

Hi,
i was playing around with Min_sphere_d and encountered some behavior i 
do not understand (see attached files).

In both cases my code creates a Min_sphere_d using 
Min_sphere_annulus_d_traits_3 and Epick and then it inserts three points 
and prints the radius.

In the first case i used doubles to initialize the Point_3 and the 
output is correct.
In the second case i used floats to initialize the Point_3 and the 
output is nan.

The only difference is the type of the values used to initialize the 
Point_3.
This means the error has to be caused by the differences in accuracy, 
but this should not really create a difference.

Has anyone any idea why this happens and how i can prevent this problem?
Since the data i am using is represented as floats, I can't really 
switch everything to double.

I appreciate any help.

Best regards,

Raphael


PS.: I am on master (54f2a119f9598f884f98df9ee1fa63aa07451c4c).


#include <iostream>
#include <cstdlib>

#include <CGAL/Exact_predicates_inexact_constructions_kernel.h>
#include <CGAL/Min_sphere_annulus_d_traits_3.h>
#include <CGAL/Min_sphere_d.h>

typedef CGAL::Epick                            K;
typedef CGAL::Min_sphere_annulus_d_traits_3<K> Traits;
typedef CGAL::Min_sphere_d<Traits>             Min_sphere;
typedef K::Point_3                             Point;

int main ()
{
    std::cout << "\ndirect insert double (point set 1)\n";
    {
        Min_sphere  ms;
        ms.insert(Point{132.973,   8.457, -34.103});
        ms.insert(Point{147.746,  -2.665, -36.310});
        ms.insert(Point{132.973,   8.457, -34.103});
        std::cout << "r^2 " << ms.squared_radius() << " c " << ms.center() << "\n";
    }
    std::cout << "\ndirect insert float (point set 1)\n";
    {
        Min_sphere  ms;
        ms.insert(Point(132.973f,  8.457f, -34.103f));
        ms.insert(Point(147.746f, -2.665f, -36.310f));
        ms.insert(Point(132.973f,  8.457f, -34.103f));
        std::cout << "r^2 " << ms.squared_radius() << " c " << ms.center() << "\n";
    }
    std::cout << "\nindirect insert double (point set 1)\n";
    {
        Min_sphere  ms;
        Point p1(132.973,  8.457, -34.103);
        Point p2(147.746, -2.665, -36.310);
        Point p3(132.973,  8.457, -34.103);
        std::cout << "p1 " << p1 << " p2 " << p2 << " p3 " << p3 << '\n';
        ms.insert(p1);
        ms.insert(p2);
        ms.insert(p3);
        std::cout << "r^2 " << ms.squared_radius() << " c " << ms.center() << "\n";
    }
    std::cout << "\nindirect insert float (point set 1)\n";
    {
        Min_sphere  ms;
        Point p1(132.973f,  8.457f, -34.103f);
        Point p2(147.746f, -2.665f, -36.310f);
        Point p3(132.973f,  8.457f, -34.103f);
        std::cout << "p1 " << p1 << " p2 " << p2 << " p3 " << p3 << '\n';
        ms.insert(p1);
        ms.insert(p2);
        ms.insert(p3);
        std::cout << "r^2 " << ms.squared_radius() << " c " << ms.center() << "\n";
    }
    std::cout << "\ndirect insert float (point set 2)\n";
    {
        Min_sphere  ms;
        ms.insert(Point(132.f,  8.f, -34.f));
        ms.insert(Point(147.f, -2.f, -36.f));
        ms.insert(Point(132.f,  8.f, -34.f));
        std::cout << "r^2 " << ms.squared_radius() << " c " << ms.center() << "\n";
    }
    return 0;
}
direct insert double (point set 1)
r^2 86.7028 c 140.36 2.896 -35.2065

direct insert float (point set 1)
r^2 -nan c -nan -nan -nan

indirect insert double (point set 1)
p1 132.973 8.457 -34.103 p2 147.746 -2.665 -36.31 p3 132.973 8.457 -34.103
r^2 86.7028 c 140.36 2.896 -35.2065

indirect insert float (point set 1)
p1 132.973 8.457 -34.103 p2 147.746 -2.665 -36.31 p3 132.973 8.457 -34.103
r^2 -nan c -nan -nan -nan

direct insert float (point set 2)
r^2 82.25 c 139.5 3 -35