Skip to content

fastjet/internal/delaunay: implement hierarchical delaunay insertion

Sebastien Binet requested to merge Bastiantheone:hierarchical_insertion into master

Created by: Bastiantheone

This PR implements the insertion using the hierarchical delaunay algorithm.

The benchmark numbers are significantly worse than the version in the first PR without the predicates. The cpu profile shows that swapDelaunay takes up most of the time. That is where the incircle function is called. If possible improving the incircle function would be the best for the performance.

Merge request reports