Post

Interesting take: "Benchmakrs should never be longer than 300ms" matklad.github.io/2026/10/05/b...

matklad.github.ioBenchmark In MillisecondsHow long should a micro benchmark run? My rule of thumb is to tweak the input size until the benchmark takes about 300ms, for the following reasons:

20/10/26, 12:46 pm0 reposts6 likes

Jacob "kurtextrem" Groß

@kurtextrem.de

Interesting take, but why not take it a step further and measure instruction count? (like "tak" and similar tools do). Ofc a bit hard with JS but not impossible

Samual

@samual.uk

not as useful as wall clock time since not all instructions take the same number of clock cycles