Benchmark suites to profile performance of Roassal3.
This project uses https://github.com/smarr/SMark to run and report benchmarks.
First, load Roassal3.
Metacello new
baseline: 'Roassal3';
repository: 'github://ObjectProfile/Roassal3';
load.
Second, load the benchmarks.
Metacello new
baseline: 'RS3Benchmarks';
repository: 'github://tinchodias/RS3Benchmarks';
load
Note Roassal3 is not added as a dependency on purpose. The intention is to avoid loading confusions when benchmarking multiple versions of Roassal3. This way, the user will explicitly choose which version is being measured.
Evaluate:
RSBenchChartBuilder exampleFirstRenderings
Wait a minute until benchmarks are completed, and a window will show up:
Evaluate:
RSLabelBenchs new runOnNumberOfRenderings
You will find a series of reports like this on Transcript:
a SmallDictionary(#numberOfLabels->100 #numberOfRenderings->0 #layoutSelector->#applyForceBasedLayout:)
Report for: RSLabelBenchs
Benchmark Labels
Labels total: iterations=10 runtime: 347.10ms +/-0.83
The first line shows the parameters of that benchmark run (100 labels, 0 renderings, force based layout).
The code is licensed under MIT.