songsenand
/
cgal
mirror of https://github.com/CGAL/cgal
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

AABB tree: no more bias for large models as the internal KD-tree is built initially and not at the first query

This commit is contained in:
Pierre Alliez 2009-07-09 13:25:25 +00:00
parent 991e936861
commit 953f277f91
2 changed files with 6 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
AABB_tree/demo/AABB_tree/benchmarks.cpp
View File

@ -66,9 +66,6 @@ void Scene::benchmark_distances(const int duration)
tree.accelerate_distance_queries();
std::cout << "done (" << time.elapsed() << " ms)" << std::endl;
// one dummy query (remove bias for large models)
tree.closest_point(CGAL::ORIGIN);
// benchmark
bench_closest_point(tree,duration);
bench_squared_distance(tree,duration);

12
AABB_tree/doc_tex/AABB_tree/performances.tex
View File

@ -12,9 +12,9 @@ The surface triangle mesh chosen for benchmarking the tree construction is the k
Number of triangles & Construction (in ms) & Construction with internal KD-tree (in ms)\\
\hline
14,400 & 156 & 157 \\
57,600 & 328 & 313 \\
230,400 & 1,141 & 1,203 \\
921,600 & 4,813 & 5,047 \\
57,600 & 328 & 328 \\
230,400 & 1,141 & 1,437 \\
921,600 & 4,813 & 5,953 \\
\hline
\end{tabular}
@ -100,9 +100,9 @@ The surface triangle mesh chosen for benchmarking distances is again the knot mo
Nb triangles & Construction (ms) & Closest\_point() & squared\_distance() & closest\_point\_and\_primitive() \\
\hline
14,400 & 157 & 45,132 & 45,626 & 45,770 \\
57,600 & 313 & 21,589 & 21,312 & 21,137 \\
230,400 & 1,203 & 11,063 & 10,962 & 11,086 \\
921,600 & 5,047 & 5,636 & 5,722 & 5,703 \\
57,600 & 328 & 21,589 & 21,312 & 21,137 \\
230,400 & 1,437 & 11,063 & 10,962 & 11,086 \\
921,600 & 5,953 & 5,636 & 5,722 & 5,703 \\
\hline
\end{tabular}