lkml.org 
[lkml]   [2011]   [Nov]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [patch 4/7] sched: convert rq->avg_idle to rq->avg_event
From
Date
On Wed, 2011-11-23 at 13:57 +0100, Mike Galbraith wrote:
> On Wed, 2011-11-23 at 13:27 +0100, Peter Zijlstra wrote:

> > Now I'm not saying this all isn't worth it, just saying my brain is
> > having difficulty seeing how it all makes sense :-)
>
> They make sense only in that one cheap number generator bandaids three
> owies. It's fugly but effective :)

Addendum:

That number represents scheduler busyness. If we're "this" busy, it's
not worth entering nohz, there's unlikely to be enough overlap to be
worth going after at the expense of L2 misses fro L3 equipped CPUs, and
we can't afford to futz around with load balancing just now.

"this" is arbitrary, but in the select_idle_sibling() case at least,
it's impossible to know what any wakee will do with the CPU, so you're
stuck with arbitrary no matter what you use to shut the thing off.

-Mike



\
 
 \ /
  Last update: 2011-11-23 15:23    [W:0.125 / U:0.200 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site