Benchmark information updated.

Author: Jaybro

README.md

@@ -8,10 +8,10 @@ See the table below to get an impression of the performance provided by the [KdT
 
 |                                     | Build C++ | Build Python  | Knn C++    | Knn Python  |
 | ----------------------------------- | --------- | ------------- | ---------- | ----------- |
-| [nanoflann][nano] v1.3.2            | 1.5s      | ...           | 3.2s       | ...         |
-| [SciPy KDTree][spkd] v1.6.3         | ...       | 5.0s          | ...        | 547.2s      |
-| [Scikit-learn KDTree][skkd] 0.22.2  | ...       | 12.2s         | ...        | 44.5s       |
-| [pykdtree][pykd] 1.3.6              | ...       | 1.0s          | ...        | 6.6s        |
+| [nanoflann][nano] v1.5.0            | 2.9s      | ...           | 3.2s       | ...         |
+| [SciPy KDTree][spkd] 1.11.0         | ...       | 4.5s          | ...        | 563.6s      |
+| [Scikit-learn KDTree][skkd] 1.2.2   | ...       | 6.2s          | ...        | 42.2s       |
+| [pykdtree][pykd] 1.3.7              | ...       | 1.0s          | ...        | 6.6s        |
 | [OpenCV FLANN][cvfn] 4.6.0          | 1.9s      | ...           | 4.7s       | ...         |
 | PicoTree KdTree v0.8.0              | 0.9s      | 1.0s          | 2.8s       | 3.1s        |
 

docs/benchmark.md

@@ -2,7 +2,9 @@
 
 One of the PicoTree examples contains [benchmarks](../examples/benchmark/) of different KdTree implementations. This page compares the one of PicoTree with respect to [nanoflann](https://github.com/jlblancoc/nanoflann) and describes how to reproduce the input that was used for benchmarking.
 
-Results were generated on: 29-08-2021 using MinGW GCC 10.3 using PicoTree 0.7.4.
+The results described in this document were generated on 29-08-2021 using MinGW GCC 10.3, PicoTree v0.7.4 and Nanoflann v1.3.2.
+
+Note: The performance of PicoTree v0.8.0 released on 30-6-2023 is identical to that of v0.7.4. However, the build algorithm of nanoflann v1.5.0 regressed and has become 90% slower.
 
 # Data sets
 

examples/benchmark/bm_nanoflann.cpp

@@ -130,15 +130,15 @@ BENCHMARK_DEFINE_F(BmNanoflann, RadiusCt)(benchmark::State& state) {
   tree.buildIndex();
 
   for (auto _ : state) {
-    std::vector<std::pair<Index, Scalar>> results;
+    std::vector<nanoflann::ResultItem<Index, Scalar>> results;
     std::size_t sum = 0;
     for (auto const& p : points_test_) {
       benchmark::DoNotOptimize(
           sum += tree.radiusSearch(
               p.data(),
               squared,
               results,
-              nanoflann::SearchParams{0, 0, false}));
+              nanoflann::SearchParameters{0, false}));
     }
   }
 }