Re: Linux 2.6.0-test6

[Rob Landley <rob@landley.net>]

On Sunday 28 September 2003 23:55, Nick Piggin wrote:

> >I.E. with your new scheduler, priority levels actually have enough of an
> >effect now that things that aren't reniced can be noticeably starved by
> >things that are.
>
> AFAIK, Con's scheduler doesn't change the nice implementation at all.
> Possibly some of his changes amplify its problems, or, more likely they
> remove most other scheduler problems leaving this one noticable.
>
> If X is running at -20, and xmms at +19, xmms is supposed to still get
> 5% of the CPU. Should be enough to run fine. Unfortunately this is
> achieved by giving X very large timeslices, so xmms's scheduling latency
> becomes large. The interactivity bonuses don't help, either.

It's the old latency vs throughput problem.  Nice only has a single linear 
metric, it says you want more or you want less but it doesn't say more or 
less of _what_.

Rob