linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
To: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: "linux-mm@kvack.org" <linux-mm@kvack.org>,
	"balbir@linux.vnet.ibm.com" <balbir@linux.vnet.ibm.com>,
	lizf@cn.fujitsu.com, a.p.zijlstra@chello.nl
Subject: Re: [-mm] [PATCH 0/4] memcg : radix-tree page_cgroup v2
Date: Thu, 27 Mar 2008 18:34:15 +0900	[thread overview]
Message-ID: <20080327183415.166db9ad.kamezawa.hiroyu@jp.fujitsu.com> (raw)
In-Reply-To: <20080327175654.C749.KOSAKI.MOTOHIRO@jp.fujitsu.com>

On Thu, 27 Mar 2008 18:12:42 +0900
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:

> Hi
> 
> >          TEST                                BASELINE     RESULT      INDEX
> > (1)      Execl Throughput                        43.0     2868.8      667.2
> > (2)      Execl Throughput                        43.0     2810.3      653.6
> > (3)      Execl Throughput                        43.0     2836.9      659.7
> > (4)      Execl Throughput                        43.0     2846.0      661.9
> > (5)      Execl Throughput                        43.0     2862.0      665.6
> > (6)      Execl Throughput                        43.0     3110.0      723.3
> > 
> > (1) .... rc5-mm1 + memory controller
> > (2) .... patch 1/4 is applied.      (use radix-tree always.)
> > (3) .... patch [1-3]/4 are applied. (caching by percpu)
> > (4) .... patch [1-4]/4 are applied. (uses prefetch)
> > (5) .... adjust sizeof(struct page) to be 64 bytes by padding.
> > (6) .... rc5-mm1 *without* memory controller
> 
> I am very surprised this result. 
> 723.3 -> 667.2 seems large performance impact.
> 
> Why do you need count resource usage when unlimited limit.
> Could you separate unlimited group to resource usage counting and no counting.
> I hope default cgroup keep no counting and no decrease performance.

At first, I'd like to reduce this overhead even under memory resource
controller's accounting ;)
We have boot-time-disable option now. But it doesn't seem what you want.

Considering workaround....
In current system, *unlimited* doesn't mean *no account*.
So, I think we have an option to add "no account" flag per cgroup.

Hmm..some interface to do
- allow "no account" -> "account"
- disallow "account" -> "no account"

Balbir-san, how do you think ?


Thanks,
-Kame





--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2008-03-27  9:34 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-03-27  8:44 KAMEZAWA Hiroyuki
2008-03-27  8:47 ` [-mm] [PATCH 1/4] memcg : radix-tree page_cgroup KAMEZAWA Hiroyuki
2008-03-28  0:55   ` Li Zefan
2008-03-28  1:13     ` KAMEZAWA Hiroyuki
2008-03-27  8:49 ` [-mm] [PATCH 2/4] memcg: boost by percpu KAMEZAWA Hiroyuki
2008-03-27  8:50 ` [-mm] [PATCH 3/4] memcg : shirink page cgroup KAMEZAWA Hiroyuki
2008-03-27  8:51 ` [-mm] [PATCH 4/4] memcg : radix-tree page_cgroup v2 KAMEZAWA Hiroyuki
2008-03-27  8:53   ` KAMEZAWA Hiroyuki
2008-03-27  9:12 ` [-mm] [PATCH 0/4] " KOSAKI Motohiro
2008-03-27  9:34   ` KAMEZAWA Hiroyuki [this message]
2008-03-27  9:53     ` Balbir Singh
2008-03-27 10:19       ` KAMEZAWA Hiroyuki

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=20080327183415.166db9ad.kamezawa.hiroyu@jp.fujitsu.com \
    --to=kamezawa.hiroyu@jp.fujitsu.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=balbir@linux.vnet.ibm.com \
    --cc=kosaki.motohiro@jp.fujitsu.com \
    --cc=linux-mm@kvack.org \
    --cc=lizf@cn.fujitsu.com \
    /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