lkml.org 
[lkml]   [2013]   [Apr]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH] sched: wake-affine throttle
From
Date
On Thu, 2013-04-11 at 10:44 +0200, Mike Galbraith wrote: 
> On Thu, 2013-04-11 at 16:26 +0800, Michael Wang wrote:
>
> > The 1:N is a good reason to explain why the chance that wakee's hot data
> > cached on curr_cpu is lower, and since it's just 'lower' not 'extinct',
> > after the throttle interval large enough, it will be balanced, this
> > could be proved, since during my test, when the interval become too big,
> > the improvement start to drop.
>
> Magnitude of improvement drops just because there's less damage done
> methinks. You'll eventually run out of measurable damage :)
>
> Yes, it's not really extinct, you _can_ reap a gain, it's just not at
> all likely to work out. A more symetric load will fare better, but any
> 1:N thing just has to spread far and wide to have any chance to perform.
> > Hmm...that's an interesting point, the workload contain different
> > 'priority' works, and depend on each other, if mother starving, all the
> > kids could do nothing but wait for her, may be that's the reason why the
> > benefit is so significant, since in such case, mother's little quicker
> > respond will make all the kids happy :)
>
> Exactly. The entire load is server latency bound. Keep the server on
> cpu, the load performs as best it can given unavoidable data miss cost.

(ie serial producer, parallel consumer... choke point lies with utterly
unscalable serial work producer)



\
 
 \ /
  Last update: 2013-04-11 11:41    [W:0.066 / U:0.172 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site