Skip to content

Latest commit

 

History

History
95 lines (72 loc) · 6.9 KB

README.md

File metadata and controls

95 lines (72 loc) · 6.9 KB
Raw

React Component Benchmark NPM version CI status

This project aims to provide a method for gathering benchmarks of component tree mount, update, and unmount timings.

Please note that the values returned are estimates. Since this project does not hook into the React renderer directly, the values gathered are not 100% accurate and may vary slightly because they're taken from a wrapping component. That being said, running a large sample set should give you a confident benchmark metric.

Motivation

Historically, React has provided react-addons-perf in order to help gain insight into the performance of mounting, updating, and unmounting components. Unfortunately, as of React 16, it has been deprecated. Additionally, before deprecation, it was not usable in production React builds, making it less useful for many applications.

Usage

See the examples directory for ideas on how you might integrate this into your own project, whether in your user-interface or your automated tests.

Quick Start

import { Benchmark } from 'react-component-benchmark';

function MyComponentBenchmark() {
	const ref = React.useRef();

	const handleComplete = React.useCallback((results) => {
		console.log(results);
	}, []);

	const handleStart = () => {
		ref.current.start();
	};

	return (
		<div>
			<button onClick={handleStart}>Run</button>
			<Benchmark
				component={MyComponent}
				componentProps={componentProps}
				onComplete={handleComplete}
				ref={ref}
				samples={50}
				timeout={10000}
				type="mount"
			/>
		</div>
	);
}

In tests

See examples/tests for various test integrations.

Build a UI

Benchmark props

key type description
component typeof React.Component The component that you would like to benchmark
componentProps object Properties to be given to component when rendering
includeLayout boolean Estimate the amount of time that the browser's rendering engine requires to layout the rendered HTML. Only available for 'mount' and 'update' benchmark types. Default: false
onComplete (x: BenchResultsType) => void Receives the benchmark results when the benchmarking is complete
samples number Samples to run (default 50)
timeout number Amount of time in milliseconds to stop running (default 10000)
type string One of 'mount', 'update', or 'unmount'. Also available from BenchmarkType.

Results

Note: All times and timestamps are in milliseconds. High resolution times provided when available.

key type description
... ...ComputedResult All values from the ComputedResult table
startTime number Timestamp of when the run started
endTime number Timestamp of when the run completed
runTime number Amount of time that it took to run all samples.
sampleCount number The number of samples actually run. May be less than requested if the timeout was hit.
samples Array<{ start, end, elapsed }> Raw sample data
layout ComputedResult The benchmark results for the browser rendering engine layout, if available (see includeLayout property)

ComputedResult

key type description
max number Maximum time elapsed
min number Minimum time elapsed
median number Median time elapsed
mean number Mean time elapsed
stdDev number Standard deviation of all elapsed times
p70 number 70th percentile for time elapsed: mean + stdDev
p95 number 95th percentile for time elapsed: mean + (stdDev * 2)
p99 number 99th percentile for time elapsed: mean + (stdDev * 3)