linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Lukasz Luba <lukasz.luba@arm.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org,
	"Rafael J. Wysocki" <rafael@kernel.org>,
	Ingo Molnar <mingo@kernel.org>,
	Dietmar Eggemann <dietmar.eggemann@arm.com>,
	Vincent Guittot <vincent.guittot@linaro.org>,
	Viresh Kumar <viresh.kumar@linaro.org>,
	Qais Yousef <qyousef@layalina.io>
Subject: Re: [RFC PATCH 0/7] sched: cpufreq: Remove magic margins
Date: Thu, 7 Sep 2023 15:42:13 +0100	[thread overview]
Message-ID: <cf5c628a-e047-b5e0-b2a0-f2b280015d02@arm.com> (raw)
In-Reply-To: <20230907142923.GJ10955@noisy.programming.kicks-ass.net>



On 9/7/23 15:29, Peter Zijlstra wrote:
> On Thu, Sep 07, 2023 at 02:57:26PM +0100, Lukasz Luba wrote:
>>
>>
>> On 9/7/23 14:26, Peter Zijlstra wrote:
>>> On Wed, Sep 06, 2023 at 10:18:50PM +0100, Qais Yousef wrote:
>>>
>>>> This is probably controversial statement. But I am not in favour of util_est.
>>>> I need to collect the data, but I think we're better with 16ms PELT HALFLIFE as
>>>> default instead. But I will need to do a separate investigation on that.
>>>
>>> I think util_est makes perfect sense, where PELT has to fundamentally
>>> decay non-running / non-runnable tasks in order to provide a temporal
>>> average, DVFS might be best served with a termporal max filter.
>>>
>>>
>>
>> Since we are here...
>> Would you allow to have a configuration for
>> the util_est shifter: UTIL_EST_WEIGHT_SHIFT ?
>>
>> I've found other values than '2' better in some scenarios. That helps
>> to prevent a big task to 'down' migrate from a Big CPU (1024) to some
>> Mid CPU (~500-700 capacity) or even Little (~120-300).
> 
> Larger values, I'm thinking you're after? Those would cause the new
> contribution to weight less, making the function more smooth, right?

Yes, more smooth, because we only use the 'ewma' goodness for decaying
part (not the raising [1]).

> 
> What task characteristic is tied to this? That is, this seems trivial to
> modify per-task.

In particular Speedometer test and the main browser task, which reaches
~900util, but sometimes vanish and waits for other background tasks
to do something. In the meantime it can decay and wake-up on
Mid/Little (which can cause a penalty to score up to 5-10% vs. if
we pin the task to big CPUs). So, a longer util_est helps to avoid
at least very bad down migration to Littles...

[1] https://elixir.bootlin.com/linux/v6.5.1/source/kernel/sched/fair.c#L4442

  reply	other threads:[~2023-09-07 15:53 UTC|newest]

Thread overview: 64+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-27 23:31 [RFC PATCH 0/7] sched: cpufreq: Remove magic margins Qais Yousef
2023-08-27 23:31 ` [RFC PATCH 1/7] sched/pelt: Add a new function to approximate the future util_avg value Qais Yousef
2023-09-06 12:56   ` Dietmar Eggemann
2023-09-06 21:19     ` Qais Yousef
2023-09-07 11:12       ` Dietmar Eggemann
2023-09-10 19:58         ` Qais Yousef
2023-09-13 17:22           ` Dietmar Eggemann
2023-09-16 19:49             ` Qais Yousef
2023-09-16 19:52               ` Qais Yousef
2023-08-27 23:31 ` [RFC PATCH 2/7] sched/pelt: Add a new function to approximate runtime to reach given util Qais Yousef
2023-09-06 12:56   ` Dietmar Eggemann
2023-09-06 20:44     ` Dietmar Eggemann
2023-09-06 21:38       ` Qais Yousef
2023-09-15  9:15   ` Hongyan Xia
2023-09-16 19:56     ` Qais Yousef
2023-08-27 23:31 ` [RFC PATCH 3/7] sched/fair: Remove magic margin in fits_capacity() Qais Yousef
2023-09-06 14:38   ` Dietmar Eggemann
2023-09-06 21:45     ` Qais Yousef
2023-08-27 23:32 ` [RFC PATCH 4/7] sched: cpufreq: Remove magic 1.25 headroom from apply_dvfs_headroom() Qais Yousef
2023-09-07 11:34   ` Peter Zijlstra
2023-09-10 19:23     ` Qais Yousef
2023-08-27 23:32 ` [RFC PATCH 5/7] sched/schedutil: Add a new tunable to dictate response time Qais Yousef
2023-09-06 21:13   ` Dietmar Eggemann
2023-09-06 21:52     ` Qais Yousef
2023-09-07 11:44   ` Peter Zijlstra
2023-09-10 19:25     ` Qais Yousef
2023-08-27 23:32 ` [RFC PATCH 6/7] sched/pelt: Introduce PELT multiplier Qais Yousef
2023-08-27 23:32 ` [RFC PATCH 7/7] cpufreq: Change default transition delay to 2ms Qais Yousef
2023-09-06  9:18 ` [RFC PATCH 0/7] sched: cpufreq: Remove magic margins Lukasz Luba
2023-09-06 21:18   ` Qais Yousef
2023-09-07  7:48     ` Lukasz Luba
2023-09-07 11:53       ` Peter Zijlstra
2023-09-07 13:06         ` Lukasz Luba
2023-09-07 13:29           ` Peter Zijlstra
2023-09-07 13:33             ` Lukasz Luba
2023-09-07 13:38               ` Peter Zijlstra
2023-09-07 13:45                 ` Lukasz Luba
2023-09-08 12:51                   ` Daniel Bristot de Oliveira
2023-09-12 11:57                     ` Lukasz Luba
2023-09-10 18:20         ` Qais Yousef
2023-09-10 18:14       ` Qais Yousef
2023-09-07 13:26     ` Peter Zijlstra
2023-09-07 13:57       ` Lukasz Luba
2023-09-07 14:29         ` Peter Zijlstra
2023-09-07 14:42           ` Lukasz Luba [this message]
2023-09-07 20:16             ` Peter Zijlstra
2023-09-12 11:51               ` Lukasz Luba
2023-09-12 14:01                 ` Vincent Guittot
2023-09-13  9:53                   ` Lukasz Luba
2023-09-12 14:28                 ` Peter Zijlstra
2023-09-10 19:06             ` Qais Yousef
2023-09-10 18:46       ` Qais Yousef
2023-09-07 13:08 ` Peter Zijlstra
2023-09-08  0:17   ` Qais Yousef
2023-09-08  7:40     ` Dietmar Eggemann
2023-09-08 14:07       ` Qais Yousef
2023-09-12 17:18         ` Dietmar Eggemann
2023-09-16 19:38           ` Qais Yousef
2023-09-08 10:25     ` Peter Zijlstra
2023-09-08 13:33       ` Qais Yousef
2023-09-08 13:58         ` Peter Zijlstra
2023-09-08 13:59         ` Peter Zijlstra
2023-09-08 14:11           ` Qais Yousef
2023-09-10 21:17         ` Qais Yousef

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=cf5c628a-e047-b5e0-b2a0-f2b280015d02@arm.com \
    --to=lukasz.luba@arm.com \
    --cc=dietmar.eggemann@arm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=peterz@infradead.org \
    --cc=qyousef@layalina.io \
    --cc=rafael@kernel.org \
    --cc=vincent.guittot@linaro.org \
    --cc=viresh.kumar@linaro.org \
    /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).