From: Andrea Arcangeli <aarcange@redhat.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Minchan Kim <minchan.kim@gmail.com>, Mel Gorman <mel@csn.ul.ie>,
Johannes Weiner <jweiner@redhat.com>,
linux-mm@kvack.org
Subject: Re: [PATCH] remove compaction from kswapd
Date: Wed, 9 Mar 2011 18:00:41 +0100 [thread overview]
Message-ID: <20110309170041.GF2141@random.random> (raw)
In-Reply-To: <20110301164143.e44e5699.akpm@linux-foundation.org>
On Tue, Mar 01, 2011 at 04:41:43PM -0800, Andrew Morton wrote:
> help a lot! It would be better to have some good, solid quantitative
> justification for what is really an emergency patch.
I figured it should be ok to post at least differential results:
Req/Achieved Response Time (msec/Op)
NFS Ops
per sec. DIFF
-------------------------------------
2000 0.0
4000 0.1
6000 0.0
8000 0.2
10000 0.8
12000 3.2
14000 5.1
16000 4.0
18000 4.4
20000 4.5
22000 4.6
24000 3.6 (server resources nearly exhausted)
26000 0.8
28000 0.1
30000 0.0
That would be the difference between upstream and all patches posted
so far. Including only Mel's patch in Message-ID:
<20110302142542.GE14162@csn.ul.ie> is enough to achieve this (no need
of the lowlatency fixes as that patch makes compaction stop running in
a loop). If we only apply the other lowlatency fixes but not Me's
patch in the message-id above, the response time difference is smaller
but not as low as with the patch 20110302142542.GE14162@csn.ul.ie.
The numbers are very reproducible so it's no measurement error even if
it's only a few msec diff. Throughput isn't significantly affected the
real thing affected is latency (and as said just applying the
lowlatency fixes of compaction isn't enough to fix latency and
compaction still shows at top of profiling).
It's not measured on the raw upstream kernel but the compaction code
is mostly identical.
Hope this helps and I'd like to see Mel's patch included.
Thanks,
Andrea
--
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/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
prev parent reply other threads:[~2011-03-09 17:01 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-02-28 22:21 [PATCH] remove compaction from kswapd Andrea Arcangeli
2011-03-01 22:33 ` Minchan Kim
2011-03-01 22:39 ` Andrea Arcangeli
2011-03-01 23:10 ` Minchan Kim
2011-03-02 0:41 ` Andrew Morton
2011-03-02 4:38 ` Andrea Arcangeli
2011-03-02 4:39 ` Andrea Arcangeli
2011-03-02 4:53 ` Andrew Morton
2011-03-02 5:52 ` Andrea Arcangeli
2011-03-02 5:57 ` Andrew Morton
2011-03-02 17:21 ` Andrea Arcangeli
2011-03-02 14:25 ` Mel Gorman
2011-03-09 22:17 ` Andrew Morton
2011-03-09 23:50 ` Andrea Arcangeli
2011-03-10 10:11 ` Mel Gorman
2011-03-09 17:00 ` Andrea Arcangeli [this message]
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=20110309170041.GF2141@random.random \
--to=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=jweiner@redhat.com \
--cc=linux-mm@kvack.org \
--cc=mel@csn.ul.ie \
--cc=minchan.kim@gmail.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;
as well as URLs for NNTP newsgroup(s).