From: Matt Fleming <matt@codeblueprint.co.uk>
To: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@kernel.org>,
linux-kernel <linux-kernel@vger.kernel.org>,
Mike Galbraith <umgwanakikbuti@gmail.com>,
Yuyang Du <yuyang.du@intel.com>
Subject: Re: [PATCH] sched/fair: Do not decay new task load on first enqueue
Date: Tue, 4 Oct 2016 21:16:07 +0100 [thread overview]
Message-ID: <20161004201607.GQ16071@codeblueprint.co.uk> (raw)
In-Reply-To: <CAKfTPtC9d8iE_jAJ9-bej886imqN2WNjsqGPP3t-CVAtiyxC-Q@mail.gmail.com>
On Wed, 28 Sep, at 04:46:06AM, Vincent Guittot wrote:
>
> ok so i'm a bit confused there
> my understand of your explanation above is that now we left a small
> amount of load in runnable_load_avg after the dequeue so another cpu
> will be chosen. But this explanation seems to be the opposite of what
> Matt said in a previous email that:
> "The performance drop comes from the fact that enqueueing/dequeueing a
> task with load 1002 during fork() results in a zero runnable_load_avg,
> which signals to the load balancer that the CPU is idle, so the next
> time we fork() we'll pick the same CPU to enqueue on -- and the cycle
> continues."
Right, we want to avoid the performance drop, which we can do by
leaving a small amount of load in runnable_load_avg. I think Dietmar
and me are saying the same thing.
next prev parent reply other threads:[~2016-10-04 20:16 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-23 11:58 [PATCH] sched/fair: Do not decay new task load on first enqueue Matt Fleming
2016-09-23 14:30 ` Vincent Guittot
2016-09-27 13:48 ` Dietmar Eggemann
2016-09-27 19:24 ` Matt Fleming
2016-09-27 19:21 ` Matt Fleming
2016-09-28 10:14 ` Peter Zijlstra
2016-09-28 11:06 ` Dietmar Eggemann
2016-09-28 11:19 ` Peter Zijlstra
2016-09-28 11:31 ` Dietmar Eggemann
2016-09-28 11:46 ` Vincent Guittot
2016-09-28 12:00 ` Vincent Guittot
2016-10-04 21:25 ` Matt Fleming
2016-10-04 20:16 ` Matt Fleming [this message]
2016-09-28 12:27 ` Vincent Guittot
2016-09-28 13:13 ` Vincent Guittot
2016-09-29 16:15 ` Dietmar Eggemann
2016-10-03 13:05 ` Vincent Guittot
2016-09-28 17:59 ` Dietmar Eggemann
2016-09-28 19:37 ` Matt Fleming
2016-09-30 20:30 ` Matt Fleming
2016-10-09 3:39 ` Wanpeng Li
2016-10-10 10:01 ` Matt Fleming
2016-10-10 10:09 ` Wanpeng Li
2016-10-11 10:27 ` Matt Fleming
2016-10-10 12:29 ` Vincent Guittot
2016-10-10 13:54 ` Dietmar Eggemann
2016-10-10 18:29 ` Vincent Guittot
2016-10-11 9:44 ` Dietmar Eggemann
2016-10-11 10:39 ` Matt Fleming
2016-10-18 10:11 ` Matt Fleming
2016-10-10 17:34 ` Vincent Guittot
2016-10-11 10:24 ` Matt Fleming
2016-10-11 13:14 ` Vincent Guittot
2016-10-11 18:57 ` Matt Fleming
2016-10-12 7:41 ` Vincent Guittot
2016-10-18 11:09 ` Peter Zijlstra
2016-10-18 15:19 ` Vincent Guittot
2016-10-18 10:29 ` Matt Fleming
2016-10-18 11:10 ` Peter Zijlstra
2016-10-18 11:29 ` Matt Fleming
2016-10-18 12:15 ` Peter Zijlstra
2016-10-19 6:38 ` Vincent Guittot
2016-10-19 9:53 ` Peter Zijlstra
2016-11-09 16:53 ` Vincent Guittot
2016-10-04 20:11 ` Matt Fleming
2016-10-09 5:57 ` [lkp] [sched/fair] f54c5d4e28: hackbench.throughput 10.6% improvement kernel test robot
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=20161004201607.GQ16071@codeblueprint.co.uk \
--to=matt@codeblueprint.co.uk \
--cc=dietmar.eggemann@arm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=peterz@infradead.org \
--cc=umgwanakikbuti@gmail.com \
--cc=vincent.guittot@linaro.org \
--cc=yuyang.du@intel.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).