From: Paul Turner <pjt@google.com>
To: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Ingo Molnar <mingo@elte.hu>,
linux-kernel@vger.kernel.org,
Bharata B Rao <bharata@linux.vnet.ibm.com>,
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>,
Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>,
Pavel Emelyanov <xemul@openvz.org>, Hu Tao <hutao@cn.fujitsu.com>
Subject: Re: [patch 00/17] CFS Bandwidth Control v7.1
Date: Fri, 8 Jul 2011 00:39:15 -0700 [thread overview]
Message-ID: <CAPM31RKH8m1CpVWmRVLi8bMNSgdQU+Qzc2sTaWbx4r1b2fHF7w@mail.gmail.com> (raw)
In-Reply-To: <1310049528.3282.583.camel@twins>
On Thu, Jul 7, 2011 at 7:38 AM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
> On Thu, 2011-07-07 at 13:23 +0200, Ingo Molnar wrote:
>>
>> The +1.5% increase in vanilla kernel context switching performance is
>> unfortunate - where does that overhead come from?
>
> Looking at the asm output, I think its partly because things like:
>
> @@ -602,6 +618,8 @@ static void update_curr(struct cfs_rq *c
> cpuacct_charge(curtask, delta_exec);
> account_group_exec_runtime(curtask, delta_exec);
> }
> +
> + account_cfs_rq_runtime(cfs_rq, delta_exec);
> }
>
>
> +static void account_cfs_rq_runtime(struct cfs_rq *cfs_rq,
> + unsigned long delta_exec)
> +{
> + if (!cfs_rq->runtime_enabled)
> + return;
> +
> + cfs_rq->runtime_remaining -= delta_exec;
> + if (cfs_rq->runtime_remaining > 0)
> + return;
> +
> + assign_cfs_rq_runtime(cfs_rq);
> +}
>
> generate a call, only to then take the first branch out, marking that
> function __always_inline would cure the call problem.
Indeed! I looked at this today, fixing this inlining recovers ~50% of
the cost; however, my numbers are not directly comparable to Hu's (~2%
originally, improving to ~1%).
> Going beyond that
> would be using static_branch() to track if there is any bandwidth
> tracking required at all.
>
I spent some time examining this option as well. Our toolchain
apparently is stuck on gcc-4.4 which left me scratching my head at the
supposed jump label assembly being omitted until I realized
CC_HAS_ASM_GOTO was missing. I will roll this up also and benchmark
tomorrow.
next prev parent reply other threads:[~2011-07-08 7:39 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-07-07 5:30 [patch 00/17] CFS Bandwidth Control v7.1 Paul Turner
2011-07-07 5:30 ` [patch 01/17] sched: (fixlet) dont update shares twice on on_rq parent Paul Turner
2011-07-21 18:28 ` [tip:sched/core] sched: Don't " tip-bot for Paul Turner
2011-07-07 5:30 ` [patch 02/17] sched: hierarchical task accounting for SCHED_OTHER Paul Turner
2011-07-07 5:30 ` [patch 03/17] sched: introduce primitives to account for CFS bandwidth tracking Paul Turner
2011-07-07 13:48 ` Peter Zijlstra
2011-07-07 21:30 ` Paul Turner
2011-07-07 5:30 ` [patch 04/17] sched: validate CFS quota hierarchies Paul Turner
2011-07-07 5:30 ` [patch 05/17] sched: accumulate per-cfs_rq cpu usage and charge against bandwidth Paul Turner
2011-07-07 5:30 ` [patch 06/17] sched: add a timer to handle CFS bandwidth refresh Paul Turner
2011-07-07 5:30 ` [patch 07/17] sched: expire invalid runtime Paul Turner
2011-07-07 5:30 ` [patch 08/17] sched: add support for throttling group entities Paul Turner
2011-07-07 5:30 ` [patch 09/17] sched: add support for unthrottling " Paul Turner
2011-07-07 5:30 ` [patch 10/17] sched: allow for positional tg_tree walks Paul Turner
2011-07-07 5:30 ` [patch 11/17] sched: prevent interactions with throttled entities Paul Turner
2011-07-07 5:30 ` [patch 12/17] sched: prevent buddy " Paul Turner
2011-07-07 5:30 ` [patch 13/17] sched: migrate throttled tasks on HOTPLUG Paul Turner
2011-07-07 5:30 ` [patch 14/17] sched: throttle entities exceeding their allowed bandwidth Paul Turner
2011-07-07 5:30 ` [patch 15/17] sched: add exports tracking cfs bandwidth control statistics Paul Turner
2011-07-07 5:30 ` [patch 16/17] sched: return unused runtime on group dequeue Paul Turner
2011-07-07 5:30 ` [patch 17/17] sched: add documentation for bandwidth control Paul Turner
2011-07-07 11:13 ` [patch 00/17] CFS Bandwidth Control v7.1 Peter Zijlstra
2011-07-11 1:22 ` Hu Tao
2011-07-07 11:23 ` Ingo Molnar
2011-07-07 11:28 ` Peter Zijlstra
2011-07-07 14:38 ` Peter Zijlstra
2011-07-07 14:51 ` Ingo Molnar
2011-07-07 14:54 ` Peter Zijlstra
2011-07-07 14:56 ` Ingo Molnar
2011-07-07 16:23 ` Jason Baron
2011-07-07 17:20 ` Peter Zijlstra
2011-07-07 18:15 ` Jason Baron
2011-07-07 20:36 ` jump_label defaults (was Re: [patch 00/17] CFS Bandwidth Control v7.1) Peter Zijlstra
2011-07-08 9:20 ` Peter Zijlstra
2011-07-08 15:47 ` Jason Baron
2011-07-07 16:52 ` [patch 00/17] CFS Bandwidth Control v7.1 Andi Kleen
2011-07-07 17:08 ` Peter Zijlstra
2011-07-07 17:59 ` Peter Zijlstra
2011-07-07 19:36 ` Jason Baron
2011-07-08 7:45 ` Paul Turner
2011-07-08 7:39 ` Paul Turner [this message]
2011-07-08 10:32 ` Peter Zijlstra
2011-07-09 7:34 ` Paul Turner
2011-07-10 18:12 ` Ingo Molnar
2011-07-07 14:06 ` Peter Zijlstra
2011-07-08 7:35 ` Paul Turner
2011-07-11 1:22 ` Hu Tao
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=CAPM31RKH8m1CpVWmRVLi8bMNSgdQU+Qzc2sTaWbx4r1b2fHF7w@mail.gmail.com \
--to=pjt@google.com \
--cc=a.p.zijlstra@chello.nl \
--cc=balbir@linux.vnet.ibm.com \
--cc=bharata@linux.vnet.ibm.com \
--cc=dhaval.giani@gmail.com \
--cc=hutao@cn.fujitsu.com \
--cc=kamalesh@linux.vnet.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=seto.hidetoshi@jp.fujitsu.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;
as well as URLs for NNTP newsgroup(s).