A short guide on how it all works will help you not only run benchmark but also add your own case.
The candidates folder contains the store (state) implementation of the tested state managers, which will be used in all benchmarks.
The cases folder contains test variants.
The scripts folder contains the scripts for running the cases. Note that candidates and cases should be independent and used only in the launch script. For convenience, next to the mjs script is the sh script. It is needed to run a particular case for all candidates.
<case_name>.mjs is a cli-like that processes the passed parameters and, depending on them, runs the case for any of the candidates.
The results of the case execution are written to results/<CASE_NAME>.json.
One by one adds n elements to the array x times. Where n is a number from the array of numbers [1, 10, 100, 1000, 10000, 100000, 1000000, 2000000, 5000000, 10000000,
50000000] (countOfElements), and x is the number of iterations (1 by default). If n = 5; x = 2, that means to add 5 elements 2 times. The element is an object {foo: "bar", baz: "qux"}. Between iterations the storage is reset (empty array).
The average value for iterations is calculated and written as the result.
Think of this case as a TODO list with a simple structure, e.g. {title: string, notes: string}.
In additional to main rules, there are a few things you need to know.
- Keep layers isolated (
candidates,cases) - Add a
shscript next tomjsscript - Adapt the new case for all existing candidates or leave a message in the PR as to why you don't.
- Test the new candidate in all cases