linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Michael Wang <wangyun@linux.vnet.ibm.com>
To: LKML <linux-kernel@vger.kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>,
	Peter Zijlstra <a.p.zijlstra@chello.nl>,
	Paul Turner <pjt@google.com>, Mike Galbraith <efault@gmx.de>,
	Andrew Morton <akpm@linux-foundation.org>,
	Tejun Heo <tj@kernel.org>,
	"Nikunj A. Dadhania" <nikunj@linux.vnet.ibm.com>,
	Namhyung Kim <namhyung@kernel.org>, Ram Pai <linuxram@us.ibm.com>
Subject: [RFC PATCH v2 0/3] sched: simplify the select_task_rq_fair()
Date: Thu, 17 Jan 2013 13:46:47 +0800	[thread overview]
Message-ID: <50F79047.1080907@linux.vnet.ibm.com> (raw)

v2 change log:
	Fix code style issues.
	Fix BUG of NULL affine_sd.
	Add the support for NUMA domain.
	Other small fix according to the old logical.
	Provide more test data with finer granularity.

This patch set is trying to simplify the select_task_rq_fair() with
schedule balance map.

After get rid of the complex code and reorganize the logical, pgbench show
the improvement, more the clients, bigger the improvement.

				Prev:		Post:

	| db_size | clients |	| tps  |	|  tps  | 
	+---------+---------+   +------+        +-------+
	| 22 MB   |       1 |   |10794 |        | 10862 |
	| 22 MB   |       2 |   |21567 |        | 21711 |
	| 22 MB   |       4 |   |41621 |        | 42609 |
	| 22 MB   |       8 |   |53883 |        | 58184 |	+7.98%
	| 22 MB   |      12 |   |50818 |        | 53464 |	+5.21%
	| 22 MB   |      16 |   |50463 |        | 55034 |	+9.06%
	| 22 MB   |      24 |   |46698 |        | 53702 |	+15.00%
	| 22 MB   |      32 |   |43404 |        | 54030 |	+24.48%
	| 7484 MB |       1 |   | 7974 |        |  8431 |
	| 7484 MB |       2 |   |19341 |        | 19509 |
	| 7484 MB |       4 |   |36808 |        | 38038 |
	| 7484 MB |       8 |   |47821 |        | 50377 |	+5.34%
	| 7484 MB |      12 |   |45913 |        | 49128 |	+7.00%
	| 7484 MB |      16 |   |46478 |        | 49670 |	+6.87%
	| 7484 MB |      24 |   |42793 |        | 48461 |	+13.25%
	| 7484 MB |      32 |   |36329 |        | 48436 |	+33.33%
	| 15 GB   |       1 |   | 7636 |        |  7852 |
	| 15 GB   |       2 |   |19195 |        | 19369 |
	| 15 GB   |       4 |   |35975 |        | 37323 |
	| 15 GB   |       8 |   |47919 |        | 50246 |	+4.86%
	| 15 GB   |      12 |   |45397 |        | 48608 |	+7.07%
	| 15 GB   |      16 |   |45926 |        | 49192 |	+7.11%
	| 15 GB   |      24 |   |42184 |        | 48007 |	+13.80%
	| 15 GB   |      32 |   |35983 |        | 47955 |	+33.27%

Please check the patch for more details about schedule balance map.

Support the NUMA domain but not tested.
Support the rebuild of domain but not tested.

Comments are very welcomed.

BTW:
	Compared with 3.7.0-rc6, 3.8.0-rc3 show big improvement with few
	clients, but get some damage with many clients, whatever, this patch
	set will help both of them to gain better performance.

Test with:
	12 cpu X86 server and linux-next 3.8.0-rc3.

Michael Wang (3):
	[RFC PATCH v2 1/3] sched: schedule balance map foundation
	[RFC PATCH v2 2/3] sched: build schedule balance map
	[RFC PATCH v2 3/3] sched: simplify select_task_rq_fair() with schedule balance map

Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 b/kernel/sched/core.c  |   44 ++++++++++++++++
 b/kernel/sched/fair.c  |  131 +++++++++++++++++++++++++------------------------
 b/kernel/sched/sched.h |   14 +++++
 kernel/sched/core.c    |   70 ++++++++++++++++++++++++++
 4 files changed, 197 insertions(+), 62 deletions(-)


             reply	other threads:[~2013-01-17  5:47 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-01-17  5:46 Michael Wang [this message]
2013-01-17  5:47 ` [RFC PATCH v2 1/3] sched: schedule balance map foundation Michael Wang
2013-01-17  5:48 ` [RFC PATCH v2 2/3] sched: build schedule balance map Michael Wang
2013-01-17  5:49 ` [RFC PATCH v2 3/3] sched: simplify select_task_rq_fair() with " Michael Wang

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=50F79047.1080907@linux.vnet.ibm.com \
    --to=wangyun@linux.vnet.ibm.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=akpm@linux-foundation.org \
    --cc=efault@gmx.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linuxram@us.ibm.com \
    --cc=mingo@redhat.com \
    --cc=namhyung@kernel.org \
    --cc=nikunj@linux.vnet.ibm.com \
    --cc=pjt@google.com \
    --cc=tj@kernel.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).