Re: page aging for _BUFFER_ cache?

Linus Torvalds (torvalds@transmeta.com)
Fri, 27 Mar 1998 10:42:24 -0800 (PST)


On Fri, 27 Mar 1998, Rik van Riel wrote:
>
> And: Bandwidth is all, or is it?

Nope. Bandwidth is easy to measure with "bonnie" etc, but if something
helps latency I tend to prioritize that. "Interactive feel" is mostly more
important than whether a compile finishes in 5:05 or 5:10 minutes.

That's why I like the new kswapd behaviour: it seems to (at least on my
machine) correctly background itself so that latency appears good. And it
keeps enough free pages around that most applications seem to never need
to really try to free pages on their own.

Linus

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu