These benchmarks compare the results for delaunator before and after a refactor of the underlying data representation.
All tests were run on a 2018 Razer Blade with an intel i7-8750H cpu, using node.js v10.15.0.
"before" results were tested on current master branch of delaunator.
"after" results were tested on my soa (structure of arrays) branch
uniform:
20000 average: before 11.8ms, after 5.318ms - 122% faster
1000000 average: before 938.27ms, after 466.973ms - %100 faster
gaussian:
20000: before 10.405ms, after 4.787ms - %117 faster
1000000: before 899.979ms, after 459.642ms - %96 faster
grid:
20000 average: before 9.665ms, after 5.528ms - 75% faster
1000000 average: before 835.14ms, after 499.08ms - 67% faster
degenerate:
20000 average: before 4.638ms, after 3.05ms- 52% faster
1000000 average: before 323.286ms, after 124.749ms - %159 faster
Below here are per-run results for each part of the comparison.
uniform:
20000: 6.255ms, 5.002ms, 4.701ms
average: 5.318ms
1000000: 488.443ms, 450.989ms, 461.488ms
average: 466.973ms
gaussian:
20000: 4.831ms, 4.673ms, 4.857ms
average: 4.787ms
1000000: 517.156ms, 425.174ms, 436.596ms
average: 459.642ms
grid:
20000: 5.680ms, 5.472ms, 5.432ms
average: 5.528
1000000: 542.662ms, 476.904ms, 477.676
average: 499.08
degenerate:
20000: 2.670ms, 4.504ms, 1.976ms
average: 3.05ms
1000000: 182.673ms, 94.173ms, 97.401ms
average: 124.749ms
uniform:
20000: 14.652, 10.382, 10.365
average: 11.8ms
1000000: 925.398, 953.461, 936.248
average: 938.27ms
gaussian:
20000: 9.809, 11.050, 10.357
average : 10.405ms
1000000: 910.472, 875.217, 914.248
average : 899.979ms
grid:
20000: 9.78, 9.522, 9.694
average : 9.665ms
1000000: 842.117, 847.033, 816.269
average : 835.14ms
degenerate:
20000: 4.947, 4.361, 4.607
average : 4.638ms
1000000: 327.96, 360.135, 281.765
average : 323.286ms