All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: Paul Turner <pjt@google.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Alex Shi <alex.shi@intel.com>, Ingo Molnar <mingo@redhat.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Andrew Morton <akpm@linux-foundation.org>,
	Borislav Petkov <bp@alien8.de>,
	Namhyung Kim <namhyung@kernel.org>,
	Mike Galbraith <efault@gmx.de>,
	Morten Rasmussen <morten.rasmussen@arm.com>,
	Vincent Guittot <vincent.guittot@linaro.org>,
	Preeti U Murthy <preeti@linux.vnet.ibm.com>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	LKML <linux-kernel@vger.kernel.org>, Mel Gorman <mgorman@suse.de>,
	Rik van Riel <riel@redhat.com>,
	Michael Wang <wangyun@linux.vnet.ibm.com>
Subject: Re: [PATCH v5 5/7] sched: compute runnable load avg in cpu_load and cpu_avg_load_per_task
Date: Mon, 6 May 2013 13:10:41 +0200	[thread overview]
Message-ID: <20130506111041.GA15446@dyad.programming.kicks-ass.net> (raw)
In-Reply-To: <CAPM31RJg87C0nO_tAaOGLaDCUAB+QmfV4LkNcwnk9j9voNocfg@mail.gmail.com>

On Mon, May 06, 2013 at 03:33:45AM -0700, Paul Turner wrote:
> Yeah, most of the rationale is super hand-wavy; especially the fairly
> arbitrary choice of periods (e.g. busy_idx vs newidle).
> 
> I think the other rationale is:
>   For smaller indicies (e.g. newidle) we speed up response time by
> also cutting motion out of the averages.
> 
> The runnable_avgs themselves actually have a fair bit of history in
> them already (50% is last 32ms); but given that they don't need to be
> cut-off to respond to load being migrated I'm guessing we could
> actually potentially get by with just "instaneous" and "use averages"
> where appropriate?

Sure,. worth a try. If things fall over we can always look at it again.

> We always end up having to re-pick/tune them based on a variety of
> workloads; if we can eliminate them I think it would be a win.

Agreed, esp. the plethora of weird idx things we currently have. If we need to
re-introduce something it would likely only be the busy case and for that we
can immediately link to the balance interval or so.




  reply	other threads:[~2013-05-06 11:12 UTC|newest]

Thread overview: 87+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-05-06  1:45 [PATCH v5 0/7] use runnable load avg in load balance Alex Shi
2013-05-06  1:45 ` [PATCH v5 1/7] Revert "sched: Introduce temporary FAIR_GROUP_SCHED dependency for load-tracking" Alex Shi
2013-05-06  8:24   ` Paul Turner
2013-05-06  8:49     ` Alex Shi
2013-05-06  8:55       ` Paul Turner
2013-05-06  8:58         ` Alex Shi
2013-05-07  5:05         ` Alex Shi
2013-05-06  1:45 ` [PATCH v5 2/7] sched: remove SMP cover for runnable variables in cfs_rq Alex Shi
2013-05-06  4:11   ` Preeti U Murthy
2013-05-06  7:18     ` Alex Shi
2013-05-06  8:01   ` Paul Turner
2013-05-06  8:57     ` Alex Shi
2013-05-06  9:08       ` Paul Turner
2013-05-06 10:47         ` Preeti U Murthy
2013-05-06 15:02         ` Alex Shi
2013-05-07  5:07         ` Alex Shi
2013-05-06  1:45 ` [PATCH v5 3/7] sched: set initial value of runnable avg for new forked task Alex Shi
2013-05-06  8:19   ` Paul Turner
2013-05-06  9:21     ` Alex Shi
2013-05-06 10:17       ` Paul Turner
2013-05-07  2:18         ` Alex Shi
2013-05-07  3:06           ` Paul Turner
2013-05-07  3:24             ` Alex Shi
2013-05-07  5:03               ` Alex Shi
2013-05-09  8:31                 ` Alex Shi
2013-05-09  9:30                   ` Paul Turner
2013-05-09 14:23                     ` Alex Shi
2013-05-08 11:15               ` Peter Zijlstra
2013-05-09  9:34               ` Paul Turner
2013-05-07  9:57             ` Morten Rasmussen
2013-05-07 11:05               ` Alex Shi
2013-05-07 11:20                 ` Paul Turner
2013-05-08 11:34                   ` Peter Zijlstra
2013-05-08 12:00                     ` Paul Turner
2013-05-09 10:55                       ` Morten Rasmussen
2013-05-09  8:22                     ` Alex Shi
2013-05-09  9:24                       ` Paul Turner
2013-05-09 13:13                         ` Alex Shi
2013-05-06 10:22       ` Paul Turner
2013-05-06 15:26         ` Alex Shi
2013-05-06 15:28           ` Peter Zijlstra
2013-05-07  2:19   ` Alex Shi
2013-05-06  1:45 ` [PATCH v5 4/7] sched: update cpu load after task_tick Alex Shi
2013-05-06  1:45 ` [PATCH v5 5/7] sched: compute runnable load avg in cpu_load and cpu_avg_load_per_task Alex Shi
2013-05-06  8:46   ` Paul Turner
2013-05-06 10:19     ` Peter Zijlstra
2013-05-06 10:33       ` Paul Turner
2013-05-06 11:10         ` Peter Zijlstra [this message]
2013-05-07  6:17           ` Alex Shi
2013-06-04  1:45             ` Alex Shi
2013-06-04  1:51               ` [DISCUSSION] removing variety rq->cpu_load ? Alex Shi
2013-06-04  2:33                 ` Michael Wang
2013-06-04  2:44                   ` Alex Shi
2013-06-04  3:09                     ` Michael Wang
2013-06-04  4:55                       ` Alex Shi
2013-05-06 15:00     ` [PATCH v5 5/7] sched: compute runnable load avg in cpu_load and cpu_avg_load_per_task Alex Shi
2013-05-06 18:34       ` Paul Turner
2013-05-07  0:24         ` Alex Shi
2013-05-07  5:12         ` Alex Shi
2013-05-06  1:45 ` [PATCH v5 6/7] sched: consider runnable load average in move_tasks Alex Shi
2013-05-06  8:53   ` Paul Turner
2013-05-06 15:04     ` Peter Zijlstra
2013-05-06 20:59       ` Paul Turner
2013-05-07  5:17         ` Alex Shi
2013-05-08  1:39           ` Alex Shi
2013-05-09  1:24             ` Alex Shi
2013-05-10 13:58               ` Alex Shi
2013-05-09  5:29             ` Alex Shi
2013-05-10 14:03               ` Alex Shi
2013-05-06 15:07     ` Alex Shi
2013-05-06  1:45 ` [PATCH v5 7/7] sched: consider runnable load average in effective_load Alex Shi
2013-05-06  3:34   ` Michael Wang
2013-05-06  5:39     ` Alex Shi
2013-05-06  6:11       ` Michael Wang
2013-05-06  9:39         ` Alex Shi
2013-05-06  7:49       ` Michael Wang
2013-05-06  8:02         ` Alex Shi
2013-05-06  8:34           ` Michael Wang
2013-05-06  9:06             ` Paul Turner
2013-05-06  9:35               ` Alex Shi
2013-05-06  9:59                 ` Preeti U Murthy
2013-05-07  2:43                   ` Michael Wang
2013-05-07  5:43                   ` Alex Shi
2013-05-08  1:33                     ` Alex Shi
2013-05-06 10:00                 ` Paul Turner
2013-05-06  7:10     ` Preeti U Murthy
2013-05-06  7:20       ` Michael Wang

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20130506111041.GA15446@dyad.programming.kicks-ass.net \
    --to=a.p.zijlstra@chello.nl \
    --cc=akpm@linux-foundation.org \
    --cc=alex.shi@intel.com \
    --cc=bp@alien8.de \
    --cc=efault@gmx.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mgorman@suse.de \
    --cc=mingo@redhat.com \
    --cc=morten.rasmussen@arm.com \
    --cc=namhyung@kernel.org \
    --cc=pjt@google.com \
    --cc=preeti@linux.vnet.ibm.com \
    --cc=riel@redhat.com \
    --cc=tglx@linutronix.de \
    --cc=vincent.guittot@linaro.org \
    --cc=viresh.kumar@linaro.org \
    --cc=wangyun@linux.vnet.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.