CGAL users discussion list

[Edoardo Milotti <>]

Hi Thierry,

the example that you quote is just what I used in my code because I had not parallelization in mind. However a simple way to achieve this is just to use a thread per node (finite vertex) and query the neighboring nodes. This limits speed to the maximum number of threads that you can actually use in your processor. I do not have the correct code at hand, but given a little time I could give you some hint. 

Best, Edoardo  

On 26 Jan 2018, at 17:48, Thierry Fredrich <> wrote:

Dear all,

I really love the Documentation of CGAL, you can find almost everything explained there. Unfortunately the parallel features are not so well documented,yet. Meanwhile I know how to create a Triangulation in parallel (the examples and this mailing-list helped ;-) ).

But after an other few days of trying to traverse a Triangulation, I get stuck and need some help or inspiration. Is there any example showing this? I tried with Compact_container, arrays, openMP and tbb. No success. (see https://github.com/thierry3000/VBL/blob/806098f993d6a96da448b6c6fad37aab043bcab5/tools/tests/parallel_insertion_in_delaunay_3.cpp)

The example (Triangulation_3/info_insert_with_pair_iterator.cpp) is exactly what I need. There the loop is traversed by "for (vit = T.finite_vertices_begin(); vit != T.finite_vertices_end(); ++vit)". But this is a pure serial execution. Is there anybody who knows how to do this in parallel manner?

Best Thierry