From: Jens Axboe <axboe@suse.de>
To: Kurt Garloff <garloff@suse.de>
Cc: Ian Pratt <m+Ian.Pratt@cl.cam.ac.uk>,
Philip R Auld <pauld@egenera.com>,
Xen development list <xen-devel@lists.xensource.com>,
Vincent Hanquez <tab@snarc.org>,
Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Subject: Re: poor domU VBD performance.
Date: Thu, 31 Mar 2005 17:41:30 +0200 [thread overview]
Message-ID: <20050331154130.GQ9204@suse.de> (raw)
In-Reply-To: <20050331153925.GP9204@suse.de>
On Thu, Mar 31 2005, Jens Axboe wrote:
> On Thu, Mar 31 2005, Kurt Garloff wrote:
> > Hi,
> >
> > On Thu, Mar 31, 2005 at 09:33:12AM -0500, Philip R Auld wrote:
> > > This effects merging though, right? I don't think the the front
> > > end has done any merging.
> >
> > The noop elevator does front and back merging.
> > My understanding is that it's used in the frontend driver.
> >
> > Otherwise, unplugging on every block would indeed be quite bad ...
>
> Not necessarily - either your io rate is not fast enough to sustain a
> substantial queue depth, in that case you get plugging on basically
> every io anyways. If on the other hand the io rate is high enough to
> maintain a queue depth of > 1, then the plugging will never take place
> because the queue never empties.
>
> So all in all, I don't think the temporary work-around will be such a
> bad idea. I would still rather implement the queue tracking though, it
> should not be more than a few lines of code.
There are still cases where it will be suboptimal of course, I didn't
intend to claim it will always be as fast as queue tracking! If you are
unlucky enough that the first request will reach the target device and
get started before the next one, you will have a small and a large part
of any given request executed. This isn't good for performance,
naturally. But queueing is so fast, I would be surprised if this
happened much in the real world.
--
Jens Axboe
next prev parent reply other threads:[~2005-03-31 15:41 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-03-30 11:16 RE: RE: poor domU VBD performance Ian Pratt
2005-03-30 17:01 ` peter bier
2005-03-30 18:05 ` Andrew Theurer
2005-03-31 7:05 ` RE: " Jens Axboe
2005-03-31 7:10 ` Jens Axboe
2005-03-31 8:17 ` Keir Fraser
2005-03-31 8:19 ` Jens Axboe
2005-03-31 14:33 ` Philip R Auld
2005-03-31 15:34 ` Kurt Garloff
2005-03-31 15:39 ` Jens Axboe
2005-03-31 15:41 ` Jens Axboe [this message]
2005-03-31 16:27 ` Nivedita Singhvi
2005-03-31 17:43 ` Jens Axboe
2005-03-31 18:27 ` Kurt Garloff
2005-03-31 21:59 ` Nivedita Singhvi
2005-03-31 15:49 ` Keir Fraser
2005-03-31 16:02 ` Andrew Theurer
2005-03-31 17:44 ` Jens Axboe
2005-03-31 16:55 ` Philip R Auld
2005-03-31 16:53 ` Philip R Auld
2005-03-31 18:01 ` Jens Axboe
2005-03-31 18:43 ` Philip R Auld
2005-03-31 19:07 ` Keir Fraser
2005-03-31 19:10 ` Keir Fraser
2005-03-31 19:20 ` Jens Axboe
2005-03-31 19:21 ` Jens Axboe
-- strict thread matches above, loose matches on Subject: below --
2005-04-04 19:36 Ian Pratt
2005-04-04 22:35 ` Nicholas Lee
2005-04-12 10:29 ` peter bier
2005-04-02 10:56 Ian Pratt
2005-04-02 12:10 ` Cédric Schieli
2005-04-01 23:22 Ian Pratt
2005-04-02 10:36 ` Cédric Schieli
2005-04-02 19:54 ` peter bier
2005-03-31 22:36 Ian Pratt
2005-03-31 23:05 ` Andrew Theurer
2005-04-01 21:40 ` Cédric Schieli
2005-03-31 21:32 Ian Pratt
2005-03-31 22:13 ` Andrew Theurer
2005-03-31 17:55 Ian Pratt
2005-03-31 18:04 ` Jens Axboe
2005-03-31 18:57 ` Keir Fraser
2005-03-31 19:22 ` Jens Axboe
2005-03-31 20:49 ` Andrew Theurer
2005-03-31 21:15 ` Keir Fraser
2005-03-31 21:27 ` Andrew Theurer
2005-04-01 5:43 ` Jens Axboe
2005-04-01 16:36 ` peter bier
[not found] <A95E2296287EAD4EB592B5DEEFCE0E9D1E3905@liverpoolst.ad.cl.cam.ac.uk>
2005-03-29 22:45 ` RE: " Kurt Garloff
2005-03-29 22:59 ` Andrew Theurer
2005-03-29 23:19 ` Kurt Garloff
2005-03-29 23:26 ` Andrew Theurer
2005-03-29 14:19 Ian Pratt
2005-03-29 15:27 ` peter bier
2005-03-28 22:17 Ian Pratt
2005-03-29 8:44 ` peter bier
2005-03-28 20:14 Ian Pratt
2005-03-28 20:18 ` Andrew Theurer
2005-03-28 21:48 ` Andrew Theurer
2005-03-28 23:38 ` Peter Bier
2005-03-29 0:27 ` Andrew Theurer
2005-03-29 11:39 ` peter bier
2005-03-28 18:55 Ian Pratt
2005-03-28 19:33 ` Andrew Theurer
2005-03-27 17:41 Ian Pratt
2005-03-28 8:48 ` peter bier
2005-03-28 12:44 ` peter bier
2005-03-29 6:20 ` Pasi Kärkkäinen
2005-03-26 18:14 Peter Bier
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=20050331154130.GQ9204@suse.de \
--to=axboe@suse.de \
--cc=Christian.Limpach@cl.cam.ac.uk \
--cc=garloff@suse.de \
--cc=m+Ian.Pratt@cl.cam.ac.uk \
--cc=pauld@egenera.com \
--cc=tab@snarc.org \
--cc=xen-devel@lists.xensource.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.