Andrii Sydorchuk wrote:
I've just added S-Hull triangulation (non-robust version) to the benchmarks and it appears that it's only a bit faster for the input sizes of around 100 points and only on Linux-64, for all the other cases Boost.Polygon Voronoi dominates (especially for large input data sets).
Updated benchmark page: http://svn.boost.org/svn/boost/sandbox/gtl/doc/voronoi_benchmark.htm
Great, that's exactly the analysis that I wanted to see! Can you update the "Logarithmic Execution Time" graph to include the s-hull results? For completeness you might also like to include q-hull, as I suspect it is probably robust. But the important thing is that you have compared your algorithm with the competitor that claims to be the current fastest. Regards, Phil.