Skip to content

Comment on Core scheduling lands in Linux 5.14parent

Comments

Yea, I'm curious what they even mean by 'performance' in this context.

It would obviously impact whole-system performance, since this setting would intentionally limit what can run on each core leading to some idle cycles. That's expected though, and is merely a consequence of using this feature for security.

I'm curious if it measurably slows down the scheduler and impacts performance from there.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.