From: Peter Zijlstra <peterz@infradead.org>
To: bharata@linux.vnet.ibm.com
Cc: linux-kernel@vger.kernel.org,
Dhaval Giani <dhaval.giani@gmail.com>,
Balbir Singh <balbir@linux.vnet.ibm.com>,
Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>,
Srivatsa Vaddagiri <vatsa@in.ibm.com>,
Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>,
Ingo Molnar <mingo@elte.hu>, Pavel Emelyanov <xemul@openvz.org>,
Herbert Poetzl <herbert@13thfloor.at>,
Avi Kivity <avi@redhat.com>, Chris Friesen <cfriesen@nortel.com>,
Paul Menage <menage@google.com>,
Mike Waychison <mikew@google.com>, Paul Turner <pjt@google.com>,
Nikhil Rao <ncrao@google.com>
Subject: Re: [PATCH v3 2/7] sched: accumulate per-cfs_rq cpu usage
Date: Thu, 14 Oct 2010 11:19:56 +0200 [thread overview]
Message-ID: <1287047996.29097.173.camel@twins> (raw)
In-Reply-To: <20101012075109.GC9893@in.ibm.com>
On Tue, 2010-10-12 at 13:21 +0530, Bharata B Rao wrote:
> +#ifdef CONFIG_CFS_BANDWIDTH
> + {
> + .procname = "sched_cfs_bandwidth_slice_us",
> + .data = &sysctl_sched_cfs_bandwidth_slice,
> + .maxlen = sizeof(unsigned int),
> + .mode = 0644,
> + .proc_handler = proc_dointvec_minmax,
> + .extra1 = &one,
> + },
> +#endif
So this is basically your scalability knob.. the larger this value less
less frequent we have to access global state, but the less parallelism
is possible due to fewer CPUs depleting the total quota, leaving nothing
for the others.
I guess one could go try and play load-balancer games to try and
mitigate this by pulling this group's tasks to the CPU(s) that have move
bandwidth for that group, but balancing that against the regular
load-balancer goal of well balancing load, will undoubtedly be
'interesting'...
next prev parent reply other threads:[~2010-10-14 9:20 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-10-12 7:49 [PATCH v3 0/7] CFS Bandwidth Control Bharata B Rao
2010-10-12 7:50 ` [PATCH v3 1/7] sched: introduce primitives to account for CFS bandwidth tracking Bharata B Rao
2010-10-13 13:00 ` Balbir Singh
2010-10-14 5:14 ` Bharata B Rao
2010-10-14 7:52 ` Peter Zijlstra
2010-10-14 12:38 ` Balbir Singh
2010-10-14 13:24 ` Peter Zijlstra
2010-12-06 9:02 ` Bharata B Rao
2010-10-12 7:51 ` [PATCH v3 2/7] sched: accumulate per-cfs_rq cpu usage Bharata B Rao
2010-10-13 13:30 ` Balbir Singh
2010-10-13 13:46 ` Nikhil Rao
2010-10-13 13:59 ` Balbir Singh
2010-10-13 14:41 ` Nikhil Rao
2010-10-14 5:39 ` Balbir Singh
2010-10-14 8:57 ` Peter Zijlstra
2010-10-14 9:07 ` Paul Turner
2010-10-14 9:13 ` Peter Zijlstra
2010-10-14 9:01 ` Peter Zijlstra
2010-10-14 9:14 ` Paul Turner
2010-10-14 9:27 ` Peter Zijlstra
2010-10-14 9:53 ` Paul Turner
2010-10-14 9:19 ` Peter Zijlstra [this message]
2010-10-14 9:27 ` Paul Turner
2010-10-14 9:40 ` Bharata B Rao
2010-10-12 7:52 ` [PATCH v3 3/7] sched: throttle cfs_rq entities which exceed their local quota Bharata B Rao
2010-10-13 6:34 ` KAMEZAWA Hiroyuki
2010-10-13 6:44 ` Paul Turner
2010-10-13 6:47 ` Bharata B Rao
2010-10-13 6:52 ` Paul Turner
2010-10-13 7:00 ` KAMEZAWA Hiroyuki
2010-10-13 7:13 ` Paul Turner
2010-10-14 9:12 ` Peter Zijlstra
2010-10-14 9:50 ` KAMEZAWA Hiroyuki
2010-10-14 9:59 ` Peter Zijlstra
2010-10-14 10:08 ` KAMEZAWA Hiroyuki
2010-10-14 10:25 ` Paul Turner
2010-10-14 10:41 ` Peter Zijlstra
2010-10-14 23:30 ` KAMEZAWA Hiroyuki
2010-10-14 10:37 ` Peter Zijlstra
2010-10-14 9:58 ` Paul Turner
2010-10-12 7:52 ` [PATCH v3 4/7] sched: unthrottle cfs_rq(s) who ran out of quota at period refresh Bharata B Rao
2010-10-15 4:45 ` Balbir Singh
2010-12-07 13:13 ` Bharata B Rao
2010-10-12 7:53 ` [PATCH v3 5/7] sched: add exports tracking cfs bandwidth control statistics Bharata B Rao
2010-10-12 7:54 ` [PATCH v3 6/7] sched: hierarchical task accounting for FAIR_GROUP_SCHED Bharata B Rao
2010-10-12 7:55 ` [PATCH v3 7/7] sched: Return/expire slack quota using generation counters Bharata B Rao
2010-10-13 5:14 ` [PATCH v3 0/7] CFS Bandwidth Control KAMEZAWA Hiroyuki
2010-10-13 5:44 ` Herbert Poetzl
2010-10-13 6:26 ` Paul Turner
2010-11-17 8:32 ` Lai Jiangshan
2010-11-19 3:24 ` Bharata B Rao
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=1287047996.29097.173.camel@twins \
--to=peterz@infradead.org \
--cc=avi@redhat.com \
--cc=balbir@linux.vnet.ibm.com \
--cc=bharata@linux.vnet.ibm.com \
--cc=cfriesen@nortel.com \
--cc=dhaval.giani@gmail.com \
--cc=herbert@13thfloor.at \
--cc=kamalesh@linux.vnet.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=menage@google.com \
--cc=mikew@google.com \
--cc=mingo@elte.hu \
--cc=ncrao@google.com \
--cc=pjt@google.com \
--cc=svaidy@linux.vnet.ibm.com \
--cc=vatsa@in.ibm.com \
--cc=xemul@openvz.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