1# Benchmarks (Taffy 0.3 and Yoga)
2
3- Taffy `0.2` results come from commit `456653c2c618ea2b3390c4d717c38edbdf684ae4`
4- Taffy `0.3` and Yoga results come from commit `71027a8de03b343e120852b84bb7dca9fb4651c5`
5- Yoga benchmarks are run using version `0.4.0` of the `yoga` crate
6- Run on a 2021 MacBook Pro with M1 Pro processor.
7- Taffy benchmarks are using criterion (10 iterations).
8
9## Flexbox Benchmarks
10
11| Benchmark                                    | Yoga      | Taffy 0.2 | Taffy 0.3 | % change (0.2 -> 0.3) |
12| ---                                          | ---       | ---       | ---       | ---                   |
13| yoga/1_000 nodes (3-level hierarchy)         | 402.58 µs | 364.60 µs | 329.04 µs | -                     |
14| yoga/10_000 nodes (4-level hierarchy)        | 4.1988 ms | 3.6236 ms | 4.3486 ms | +20.01%               |
15| yoga/100_000 nodes (5-level hierarchy)       | 45.804 ms | 39.467 ms | 38.559 ms | -                     |
16| wide/1_000 nodes (2-level hierarchy)         | 737.77 µs | 470.52 µs | 505.99 µs | -                     |
17| wide/10_000 nodes (2-level hierarchy)        | 7.1007 ms | 7.1008 ms | 8.3395 ms | +17.44%               |
18| wide/100_000 nodes (2-level hierarchy)       | 135.78 ms | 241.34 ms | 247.42 ms | -                     |
19| deep/4000 nodes (12-level hierarchy))        | 2.2333 ms | 4.7362 ms | 1.7400 ms | -63.26%               |
20| deep/10_000 nodes (14-level hierarchy)       | 5.9477 ms | 10.729 ms | 4.4445 ms | -58.57%               |
21| deep/100_000 nodes (17-level hierarchy)      | 76.755 ms | 151.53 ms | 63.778 ms | -57.91%               |
22| super-deep/1000 nodes (1000-level hierarchy) | 555.32 µs | 1.6444 ms | 472.85 µs | -71.24%               |
23
24## CSS Grid Benchmarks
25
26`NxN/12345` signifies that each node is an N by N grid, and that there are 12345 total nodes in the tree.
27
28| Benchmark                                    | Taffy 0.3 |
29| ---                                          | ---       |
30| grid/wide/31x31/961                          | 491.65 µs |
31| grid/wide/100x100/10000                      | 7.0741 ms |
32| grid/wide/316x316/99856                      | 104.54 ms |
33| grid/deep/2x2/1024                           | 1.6624 ms |
34| grid/deep/3x3/6561                           | 8.1927 ms |
35| grid/deep/2x2/16384                          | 32.700 ms |
36| grid/superdeep/1x1/100                       | 172.83 µs |
37| grid/superdeep/1x1/1000                      | 2.0197 ms |