From: Peter Zaitsev <peter@mysql.com>
To: "Martin J. Bligh" <mbligh@aracnet.com>
Cc: Andrew Morton <akpm@osdl.org>,
andrea@suse.de, riel@redhat.com, linux-kernel@vger.kernel.org
Subject: Re: 2.4.23aa2 (bugfixes and important VM improvements for the high end)
Date: Thu, 04 Mar 2004 12:21:36 -0800 [thread overview]
Message-ID: <1078431695.2770.498.camel@abyss.local> (raw)
In-Reply-To: <132310000.1078421713@flay>
On Thu, 2004-03-04 at 09:35, Martin J. Bligh wrote:
>
> 2. If it's gettimeofday hammering it (which it probably is, from previous
> comments by others, and my own experience), then vsyscall gettimeofday
> (John's patch) may well fix it up.
Well, as I wrote MySQL does not use a lot of gettimeofday. It rather
has 2-3 calls to time() per query, but it is very small number compared
to othet syscalls.
>
> 3. Are you using the extra user address space? Otherwise yes, it'll be
> all downside. And 4/4 vs 3/1 isn't really a fair comparison ... 4/4 is
> designed for bigboxen, so 4/4 vs 2/2 would be better, IMHO. People have
> said before that DB performance can increase linearly with shared area
> sizes (for some workloads), so that'd bring you a 100% or so increase
> in performance for 4/4 to counter the loss.
I do not really understand this :)
I know 4/4 was designed for BigBoxes, however we're more interested in
side effect we have - having 4G per user process instead of 3G in 3G/1G
split. As MySQL is designed as single process this is what rather
important for us.
I was not using extra address space in this test, as the idea was to see
how much slowdown 4G/4G split gives you with all other being the same.
Based on other benchmarks I know extra performance extra 1Gb used as
buffers can give.
Bringing this numbers together I shall conclude what 4G/4G does not make
sense for most MySQL loads, as 1Gb used for internal buffers (vs 1Gb
used for file cache) will not give high enough performance to cover such
major speed loss.
There are exceptions of course, for example the case where your full
workload will fit in 3G cache while will not fit in 2G (very edge one),
or in case you need 4G just to manage 10000+ connections with
reasonable buffers etc, which is also far from most typical scenario.
For "Big Boxes" I just would not advice having 32bit configuration at
all - happily nowadays you can get 64bit pretty cheap.
--
Peter Zaitsev, Senior Support Engineer
MySQL AB, www.mysql.com
Meet the MySQL Team at User Conference 2004! (April 14-16, Orlando,FL)
http://www.mysql.com/uc2004/
next prev parent reply other threads:[~2004-03-04 20:22 UTC|newest]
Thread overview: 100+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-02-27 1:33 Andrea Arcangeli
2004-02-27 4:38 ` Rik van Riel
2004-02-27 17:32 ` Andrea Arcangeli
2004-02-27 19:08 ` Rik van Riel
2004-02-27 20:29 ` Andrew Morton
2004-02-27 20:49 ` Rik van Riel
2004-02-27 20:55 ` Andrew Morton
2004-02-27 21:28 ` Andrea Arcangeli
2004-02-27 21:37 ` Andrea Arcangeli
2004-02-28 3:22 ` Andrea Arcangeli
2004-03-01 11:10 ` Nikita Danilov
2004-02-27 21:15 ` Andrea Arcangeli
2004-02-27 22:03 ` Martin J. Bligh
2004-02-27 22:23 ` Andrew Morton
2004-02-28 2:32 ` Andrea Arcangeli
2004-02-28 4:57 ` Wim Coekaerts
2004-02-28 6:18 ` Andrea Arcangeli
2004-02-28 6:45 ` Martin J. Bligh
2004-02-28 7:05 ` Andrea Arcangeli
2004-02-28 9:19 ` Dave Hansen
2004-03-18 2:44 ` Andrea Arcangeli
[not found] ` <20040228061838.GO8834@dualathlon.random.suse.lists.linux.kernel>
2004-02-28 12:46 ` Andi Kleen
2004-02-29 1:39 ` Andrea Arcangeli
2004-02-29 2:29 ` Andi Kleen
2004-02-29 16:34 ` Andrea Arcangeli
2004-02-28 6:10 ` Martin J. Bligh
2004-02-28 6:43 ` Andrea Arcangeli
2004-02-28 7:00 ` Martin J. Bligh
2004-02-28 7:29 ` Andrea Arcangeli
2004-02-28 14:55 ` Rik van Riel
2004-02-28 15:06 ` Arjan van de Ven
2004-02-29 1:43 ` Andrea Arcangeli
[not found] ` < 1078370073.3403.759.camel@abyss.local>
2004-03-04 3:14 ` Peter Zaitsev
2004-03-04 3:33 ` Andrew Morton
2004-03-04 3:44 ` Peter Zaitsev
2004-03-04 4:07 ` Andrew Morton
2004-03-04 4:44 ` Peter Zaitsev
2004-03-04 4:52 ` Andrea Arcangeli
2004-03-04 5:10 ` Andrew Morton
2004-03-04 5:27 ` Andrea Arcangeli
2004-03-04 5:38 ` Andrew Morton
2004-03-05 20:19 ` Jamie Lokier
2004-03-05 20:33 ` Andrea Arcangeli
2004-03-05 21:44 ` Jamie Lokier
2004-03-04 12:12 ` Rik van Riel
2004-03-04 16:21 ` Peter Zaitsev
2004-03-04 18:13 ` Andrea Arcangeli
2004-03-04 17:35 ` Martin J. Bligh
2004-03-04 18:16 ` Andrea Arcangeli
2004-03-04 19:31 ` Martin J. Bligh
2004-03-04 20:21 ` Peter Zaitsev [this message]
2004-03-05 10:33 ` Ingo Molnar
2004-03-05 14:15 ` Andrea Arcangeli
2004-03-05 14:32 ` Ingo Molnar
2004-03-05 14:58 ` Andrea Arcangeli
2004-03-05 15:26 ` Ingo Molnar
2004-03-05 15:53 ` Andrea Arcangeli
2004-03-07 8:41 ` Ingo Molnar
2004-03-07 10:29 ` Nick Piggin
2004-03-07 17:33 ` Andrea Arcangeli
2004-03-08 5:15 ` Nick Piggin
2004-03-07 17:24 ` Andrea Arcangeli
2004-03-05 21:28 ` Martin J. Bligh
2004-03-05 18:42 ` Martin J. Bligh
2004-03-05 19:13 ` Andrea Arcangeli
2004-03-05 19:55 ` Martin J. Bligh
2004-03-05 20:29 ` Andrea Arcangeli
2004-03-05 20:41 ` Andrew Morton
2004-03-05 21:07 ` Andrea Arcangeli
2004-03-05 22:12 ` Andrew Morton
2004-03-05 14:34 ` Ingo Molnar
2004-03-05 14:59 ` Andrea Arcangeli
2004-03-05 15:02 ` Ingo Molnar
[not found] ` <20040305150225.GA13237@elte.hu.suse.lists.linux.kernel>
2004-03-05 15:51 ` Andi Kleen
2004-03-05 16:23 ` Ingo Molnar
2004-03-05 16:39 ` Andrea Arcangeli
2004-03-07 8:16 ` Ingo Molnar
2004-03-10 13:21 ` Andi Kleen
2004-03-05 16:42 ` Andrea Arcangeli
2004-03-05 16:49 ` Ingo Molnar
2004-03-05 16:58 ` Andrea Arcangeli
2004-03-05 20:11 ` Jamie Lokier
2004-03-06 5:12 ` Jamie Lokier
2004-03-06 12:56 ` Magnus Naeslund(t)
2004-03-06 13:13 ` Magnus Naeslund(t)
2004-03-07 11:55 ` Ingo Molnar
2004-03-07 6:50 ` Peter Zaitsev
2004-03-02 9:10 ` Kurt Garloff
2004-03-02 15:32 ` Martin J. Bligh
2004-02-27 21:42 ` Hugh Dickins
2004-02-27 23:18 ` Marcelo Tosatti
2004-02-27 22:39 ` Andrew Morton
2004-02-27 20:31 ` Andrea Arcangeli
2004-02-29 6:34 ` Mike Fedyk
[not found] <20040304175821.GO4922@dualathlon.random>
2004-03-04 22:14 ` Rik van Riel
2004-03-04 23:24 ` Andrea Arcangeli
2004-03-05 3:43 ` Rik van Riel
[not found] <1u7eQ-6Bz-1@gated-at.bofh.it>
[not found] ` <1ue6M-45w-11@gated-at.bofh.it>
[not found] ` <1uofN-4Rh-25@gated-at.bofh.it>
[not found] ` <1vRz3-5p2-11@gated-at.bofh.it>
[not found] ` <1vRSn-5Fc-11@gated-at.bofh.it>
[not found] ` <1vS26-5On-21@gated-at.bofh.it>
[not found] ` <1wkUr-3QW-11@gated-at.bofh.it>
[not found] ` <1wolx-7ET-31@gated-at.bofh.it>
[not found] ` <1woEM-7Yx-41@gated-at.bofh.it>
[not found] ` <1wp8b-7x-3@gated-at.bofh.it>
[not found] ` <1wp8l-7x-25@gated-at.bofh.it>
[not found] ` <1x0qG-Dr-3@gated-at.bofh.it>
2004-03-12 21:15 ` Andi Kleen
2004-03-18 19:50 ` Peter Zaitsev
[not found] ` <1woEJ-7Yx-25@gated-at.bofh.it>
[not found] ` <1wp8c-7x-5@gated-at.bofh.it>
[not found] ` <1wprd-qI-21@gated-at.bofh.it>
[not found] ` <1wpUz-Tw-21@gated-at.bofh.it>
[not found] ` <1x293-2nT-7@gated-at.bofh.it>
2004-03-12 21:25 ` Andi Kleen
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=1078431695.2770.498.camel@abyss.local \
--to=peter@mysql.com \
--cc=akpm@osdl.org \
--cc=andrea@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mbligh@aracnet.com \
--cc=riel@redhat.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
all inboxes | Powered by JetHome®