From: dino@in.ibm.com
To: Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@elte.hu>,
Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: linux-kernel@vger.kernel.org, linux-rt-users@vger.kernel.org,
John Stultz <johnstul@us.ibm.com>,
Darren Hart <dvhltc@us.ibm.com>, John Kacur <jkacur@redhat.com>
Subject: [patch -rt 00/17] [patch -rt] Sched load balance backport
Date: Thu, 22 Oct 2009 18:07:43 +0530 [thread overview]
Message-ID: <20091022123743.506956796@spinlock.in.ibm.com> (raw)
Problem: The current -rt, 2.6.31.4-rt14, has several load balancing issues.
It can easily be seen with dbench running with the same number of processes
as the number of cpus on a SMP system. The dbench threads are running as
SCHED_OTHER processes. Several threads end up running on the same cpu even
though other cpus are idling. This results in severe throughput regression.
Peter Zijlstra posted several load balance patches meant for mainline a
while ago. The following patch series is a backport of the same to
2.6.31.4-rt14. Peter's original patches can be found at
http://marc.info/?l=linux-kernel&m=125198436208787&w=2
Patches 1 through 14 are backports of the original patches except
patch 10 which has the original code + bug fixes.
Patches 15 - 17 are various relevant fixes that have gone on top
of the original patches.
With these patches, load balancing improves considerably on -rt.
However it does not completely resolve the problem. This is still
under investigation.
This has been stress tested on x86_64 and i686
-Dinakar
--
next reply other threads:[~2009-10-22 12:42 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-10-22 12:37 dino [this message]
2009-10-22 12:37 ` [patch -rt 01/17] sched: restore __cpu_power to a straight sum of power dino
2009-10-22 12:37 ` [patch -rt 02/17] sched: SD_PREFER_SIBLING dino
2009-10-22 12:37 ` [patch -rt 03/17] sched: update the cpu_power sum during load-balance dino
2009-10-22 12:37 ` [patch -rt 04/17] sched: add smt_gain dino
2009-10-22 12:37 ` [patch -rt 05/17] sched: dynamic cpu_power dino
2009-10-22 12:37 ` [patch -rt 06/17] sched: scale down cpu_power due to RT tasks dino
2009-10-22 12:37 ` [patch -rt 07/17] sched: try to deal with low capacity dino
2009-10-22 12:37 ` [patch -rt 08/17] sched: remove reciprocal for cpu_power dino
2009-10-22 12:37 ` [patch -rt 09/17] x86: move APERF/MPERF into a X86_FEATURE dino
2009-10-22 12:37 ` [patch -rt 10/17] x86: Add generic aperf/mperf code dino
2009-10-22 12:37 ` [patch -rt 11/17] Provide an arch specific hook for cpufreq based scaling of cpu_power dino
2009-10-22 12:37 ` [patch -rt 12/17] x86: sched: provide arch implementations using aperf/mperf dino
2009-10-22 12:37 ` [patch -rt 13/17] sched: cleanup wake_idle power saving dino
2009-10-22 12:37 ` [patch -rt 14/17] sched: cleanup wake_idle dino
2009-10-22 12:37 ` [patch -rt 15/17] sched: Add a missing = dino
2009-10-22 12:37 ` [patch -rt 16/17] sched: Deal with low-load in wake_affine() dino
2009-10-22 12:38 ` [patch -rt 17/17] sched: Fix dynamic power-balancing crash dino
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=20091022123743.506956796@spinlock.in.ibm.com \
--to=dino@in.ibm.com \
--cc=a.p.zijlstra@chello.nl \
--cc=dvhltc@us.ibm.com \
--cc=jkacur@redhat.com \
--cc=johnstul@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-rt-users@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=tglx@linutronix.de \
/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