From: Mike Galbraith <efault@gmx.de>
To: Peter Williams <pwil3058@bigpond.net.au>
Cc: Con Kolivas <kernel@kolivas.org>,
Linux Kernel <linux-kernel@vger.kernel.org>,
Kingsley Cheung <kingsley@aurema.com>,
Ingo Molnar <mingo@elte.hu>,
Rene Herman <rene.herman@keyaccess.nl>
Subject: Re: [RFC 0/5] sched: Add CPU rate caps
Date: Fri, 26 May 2006 10:04:20 +0200 [thread overview]
Message-ID: <1148630661.7589.65.camel@homer> (raw)
In-Reply-To: <20060526042021.2886.4957.sendpatchset@heathwren.pw.nest>
On Fri, 2006-05-26 at 14:20 +1000, Peter Williams wrote:
> These patches implement CPU usage rate limits for tasks.
>
> Although the rlimit mechanism already has a CPU usage limit (RLIMIT_CPU)
> it is a total usage limit and therefore (to my mind) not very useful.
> These patches provide an alternative whereby the (recent) average CPU
> usage rate of a task can be limited to a (per task) specified proportion
> of a single CPU's capacity.
The killer problem I see with this approach is that it doesn't address
the divide and conquer problem. If a task is capped, and forks off
workers, each worker inherits the total cap, effectively extending same.
IMHO, per task resource management is too severely limited in it's
usefulness, because jobs are what need managing, and they're seldom
single threaded. In order to use per task limits to manage any given
job, you have to both know the number of components, and manually
distribute resources to each component of the job. If a job has a
dynamic number of components, it becomes impossible to manage.
-Mike
next prev parent reply other threads:[~2006-05-26 8:02 UTC|newest]
Thread overview: 95+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-05-26 4:20 [RFC 0/5] sched: Add CPU rate caps Peter Williams
2006-05-26 4:20 ` [RFC 1/5] sched: Fix priority inheritence before CPU rate soft caps Peter Williams
2006-05-26 4:20 ` [RFC 2/5] sched: Add " Peter Williams
2006-05-26 10:48 ` Con Kolivas
2006-05-26 11:15 ` Mike Galbraith
2006-05-26 11:17 ` Con Kolivas
2006-05-26 11:30 ` Mike Galbraith
2006-05-26 13:55 ` Peter Williams
2006-05-27 6:31 ` Balbir Singh
2006-05-27 7:03 ` Peter Williams
2006-05-28 0:11 ` Peter Williams
2006-05-28 7:38 ` Balbir Singh
2006-05-28 13:35 ` Peter Williams
2006-05-28 14:42 ` Balbir Singh
2006-05-28 23:27 ` Peter Williams
2006-05-31 13:17 ` Kirill Korotaev
2006-05-31 23:39 ` Peter Williams
2006-06-01 8:09 ` Kirill Korotaev
2006-06-01 23:38 ` Peter Williams
2006-06-02 1:35 ` Peter Williams
2006-05-26 4:20 ` [RFC 3/5] sched: Add CPU rate hard caps Peter Williams
2006-05-26 6:58 ` Kari Hurtta
2006-05-27 1:00 ` Peter Williams
2006-05-26 11:00 ` Con Kolivas
2006-05-26 13:59 ` Peter Williams
2006-05-26 14:12 ` Con Kolivas
2006-05-26 14:23 ` Mike Galbraith
2006-05-27 0:16 ` Peter Williams
2006-05-27 9:28 ` Mike Galbraith
2006-05-28 2:09 ` Peter Williams
2006-05-27 6:48 ` Balbir Singh
2006-05-27 8:44 ` Peter Williams
2006-05-31 13:10 ` Kirill Korotaev
2006-05-31 15:59 ` Balbir Singh
2006-05-31 18:09 ` Mike Galbraith
2006-06-01 7:41 ` Kirill Korotaev
2006-06-01 8:34 ` Balbir Singh
2006-06-01 18:43 ` [ckrm-tech] " Chandra Seetharaman
2006-06-01 23:26 ` Peter Williams
2006-06-02 2:02 ` Chandra Seetharaman
2006-06-02 3:21 ` Peter Williams
2006-06-02 8:32 ` Balbir Singh
2006-06-02 13:30 ` Peter Williams
2006-06-02 18:58 ` Balbir Singh
2006-06-02 23:49 ` Peter Williams
2006-06-03 4:59 ` Balbir Singh
2006-06-02 19:06 ` Chandra Seetharaman
2006-06-03 0:04 ` Peter Williams
2006-06-02 0:36 ` Con Kolivas
2006-06-02 2:03 ` [ckrm-tech] " Chandra Seetharaman
2006-06-02 5:55 ` [ckrm-tech] [RFC 3/5] " Peter Williams
2006-06-02 7:47 ` Kirill Korotaev
2006-06-02 13:34 ` Peter Williams
2006-06-05 22:11 ` Sam Vilain
2006-06-06 8:24 ` Kirill Korotaev
2006-06-06 9:13 ` Con Kolivas
2006-06-06 9:28 ` Kirill Korotaev
2006-06-02 8:46 ` Mike Galbraith
2006-06-02 13:18 ` Peter Williams
2006-06-02 14:47 ` Mike Galbraith
2006-06-03 0:08 ` Peter Williams
2006-06-03 6:02 ` Mike Galbraith
2006-06-03 11:03 ` Peter Williams
2006-06-06 11:26 ` Srivatsa Vaddagiri
2006-06-02 7:34 ` Kirill Korotaev
2006-06-02 21:23 ` Shailabh Nagar
2006-06-01 23:47 ` Sam Vilain
2006-06-01 23:43 ` Peter Williams
2006-05-31 23:28 ` Peter Williams
2006-06-01 7:44 ` Kirill Korotaev
2006-06-01 23:21 ` Peter Williams
2006-05-26 4:21 ` [RFC 4/5] sched: Add procfs interface for CPU rate soft caps Peter Williams
2006-05-26 4:21 ` [RFC 5/5] sched: Add procfs interface for CPU rate hard caps Peter Williams
2006-05-26 8:04 ` Mike Galbraith [this message]
2006-05-26 16:11 ` [RFC 0/5] sched: Add CPU rate caps Björn Steinbrink
2006-05-28 22:46 ` Sam Vilain
2006-05-28 23:30 ` Peter Williams
2006-05-29 3:09 ` Sam Vilain
2006-05-29 3:41 ` Peter Williams
2006-05-29 21:16 ` Sam Vilain
2006-05-29 23:12 ` Peter Williams
2006-05-30 2:07 ` Sam Vilain
2006-05-30 2:45 ` Peter Williams
2006-05-30 22:05 ` Sam Vilain
2006-05-30 23:22 ` Peter Williams
2006-05-30 23:25 ` Peter Williams
2006-06-05 23:56 ` Peter Williams
2006-05-27 0:16 ` Peter Williams
2006-05-26 10:41 ` Con Kolivas
2006-05-27 1:28 ` Peter Williams
2006-05-27 1:42 ` Con Kolivas
2006-05-26 11:09 ` Con Kolivas
2006-05-26 14:00 ` Peter Williams
2006-05-26 11:29 ` Balbir Singh
2006-05-27 1:40 ` Peter Williams
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=1148630661.7589.65.camel@homer \
--to=efault@gmx.de \
--cc=kernel@kolivas.org \
--cc=kingsley@aurema.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=pwil3058@bigpond.net.au \
--cc=rene.herman@keyaccess.nl \
/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