CGAL users discussion list

[Rash <>]

It looks like that AABBTree is locked when an other thread is using it. 
When you are using a single tree per thread you get the full speed. I am 
using the CGAL::Simple_cartesian<double> Kernel.

Runtime 16 secs, independ of #pragma omp parallel for, only one core is 
used in system manager (ubuntu):

        #pragma omp parallel for
        for(size_t y = 0; y < number_of_ray_y; y++)
        {
            for(size_t x = 0; x < number_of_ray_x; x++)
            {
                Ray_intersection hit = 
tree.first_intersection(rays[this->transformCoordinates(y,x)]);
            }
        }

Runtime 4 secs 4 cores are used:

#pragma omp parallel
{
        Tree localtree(faces(mesh).first, faces(mesh).second, mesh);
        std::cout << "Start: " << omp_get_wtime() << std::endl;

        size_t tid = omp_get_thread_num();
        size_t number_of_threads = omp_get_num_procs();
        for(size_t y = tid; y < number_of_ray_y; y=y+number_of_threads)
        {
                if(y >= number_of_ray_y)
                {
                        break;
                }

                for(size_t x = 0; x < number_of_ray_x; x++)
                {
Ray_intersection hit = 
localtree.first_intersection(rays[this->transformCoordinates(y,x)]);
                }
        }
}


This is very troubling, an acceleration structure can have a pretty huge 
memory footprint.  Am I using something wrong?

    typedef CGAL::AABB_face_graph_triangle_primitive<Polyhedron> Primitive;
    typedef CGAL::AABB_traits<Kernel, Primitive> Traits;
    typedef CGAL::AABB_tree<Traits> Tree;

Kind regards
Rashid