Runner API

Run a Suite

The runSuite function is provided for run a benchmark suite, and the results can be converted to a table using SummaryTable.

import { defineSuite, runSuite, SummaryTable } from "esbench";

const suite = defineSuite(scene => {
	const values = Array.from({ length: 1000 }, Math.random);
	scene.bench("reduce", () => values.reduce((v, s) => s + v));
});

const result = await runSuite(suite);

// Convert the result to a table
const summaryTable = SummaryTable.from([result]);

// Print the table
console.log("\n" + summaryTable.format().toMarkdown());
if (summaryTable.hints.length > 0) {
	console.log("Hints:");
	summaryTable.hints.forEach(n => console.log(n));
}
if (summaryTable.warnings.length > 0) {
	console.log("Warnings:");
	summaryTable.warnings.forEach(n => console.log(n));
}

Parse Results

Despite having SummaryTable to process the results, sometimes we need more primitive data structures. The return value of runSuite is intended to serialize, and it is recommended to use Summary to parse it instead of using it directly.

import { Summary, runSuite } from "esbench";

const results = runSuite(/* ... */);

// ESBench provide a helper `Summary` to parse the results of runSuite.
const summary = new Summary([results]);

console.log("Variables:", summary.vars);
console.log("Metric Descriptions:", summary.meta);

console.log("Case Results:");
for (let i = 0; i < summary.results.length; i++) {
	const result = summary.results[i];
	const metrics = Summary.getMetrics(result);

	console.log(`${i}.`.padEnd(3), JSON.stringify(result));
	console.log("   ", JSON.stringify(metrics));
}

// Get the metrics of specific case.
Summary.getMetrics(summary.find({
	Name: "object",
	exists: "true",
	size: "1000",
	Builder: "None",
	Executor: "node",
}));