Benchmarking Real Work (avoid setup bias)

Problem

You want to benchmark a function without accidentally measuring setup or allocation work.

Idiom: Keep Setup Outside the Benchmarked Function

Move any setup outside the function passed to micro, so the benchmark measures only the work you care about.

var {*} = import(
	"slug.time",
	"slug.benchmark",
)

var f = fn() {
	sleep(100)
}

micro("sleep 100ms", f)
	/> printResult

This produces more stable and meaningful results.