From: Vladislav Bolkhovitin <vst@vlnb.net>
To: Chris Worley <worleys@gmail.com>
Cc: "Pasi Kärkkäinen" <pasik@iki.fi>,
"Chetan Loke" <chetanloke@gmail.com>,
"Bart Van Assche" <bvanassche@acm.org>,
linux-scsi@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>,
"James Bottomley" <James.Bottomley@suse.de>,
scst-devel <scst-devel@lists.sourceforge.net>
Subject: Re: Linux I/O subsystem performance
Date: Wed, 25 Aug 2010 23:12:55 +0400 [thread overview]
Message-ID: <4C756B37.4000007@vlnb.net> (raw)
In-Reply-To: <AANLkTin85yqvykJPW=Zk46m2hA=n2wTRamYodRgMZCVM@mail.gmail.com>
Chris Worley, on 08/25/2010 12:31 AM wrote:
>> I also have an impression that Linux I/O subsystem has some performance
>> problems. For instance, in one recent SCST performance test only 8 Linux
>> initiators with fio as a load generator were able to saturate a single SCST
>> target with dual IB cards (SRP) on 4K AIO direct accesses over an SSD
>> backend. This rawly means that any initiator took several times (8?) more
>> processing time than the target.
>
> While I can't tell you where the bottlenecks are, I can share some
> performance numbers...
>
> 4 initiators can get>600K random 4KB IOPS off a single target...
Hmm, on the data you sent me only 8 initiators were capable to do so...
I'm glad to see an improvement here ;).
> which is ~150% of what the Emulex/Intel/Microsoft results show using 8
> targets at 4KB (their 1M IOPS was at 512 byte blocks, which is not a
> realistic test point
From my, a storage developer's, POV it isn't about if this test is
realistic or not. 512 bytes tests are good if you want to test how
processing effective your I/O stack, because they produce the max
possible CPU/memory/hardware interaction load. Since processing power
isn't unlimited, in case if it is a bottleneck, N IOPS on 512b < N IOPS
on 4K * 8 and system with more effective processing will have better
numbers.
Vlad
next prev parent reply other threads:[~2010-08-25 19:13 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-08-18 14:58 [Scst-devel] Fwd: Re: linuxcon 2010 Chetan Loke
2010-08-18 15:11 ` James Bottomley
[not found] ` <AANLkTimJGxn=5kEMH68XVWqFcYG3vpfLjLjZpFGqhG_4@mail.gmail.com>
2010-08-18 15:30 ` Bart Van Assche
2010-08-18 16:04 ` Chetan Loke
2010-08-18 16:18 ` James Bottomley
2010-08-18 17:50 ` Vladislav Bolkhovitin
2010-08-19 1:18 ` jack wang
2010-08-20 13:46 ` Ruben Laban
[not found] ` <8A96806D-6CD7-44AD-8A9D-143C098C95A4@uni-paderborn.de>
[not found] ` <1282256949.30453.278.camel@haakon2.linux-iscsi.org>
2010-08-21 18:42 ` Vladislav Bolkhovitin
2010-08-21 20:25 ` Nicholas A. Bellinger
2010-08-24 18:08 ` Vladislav Bolkhovitin
2010-08-21 20:43 ` James Bottomley
2010-08-22 7:39 ` Bart Van Assche
2010-08-22 20:29 ` James Bottomley
2010-08-23 13:47 ` Joe Landman
2010-08-23 15:12 ` Bart Van Assche
[not found] ` <AANLkTim-M6dfLvJQnbieFqZCGG33E+-i+u_soCq2p9f1@mail.gmail.com>
2010-08-23 16:07 ` Chetan Loke
2010-08-23 18:03 ` Chetan Loke
2010-08-24 7:25 ` Pasi Kärkkäinen
2010-08-24 14:43 ` Linux I/O subsystem performance (was: linuxcon 2010...) Vladislav Bolkhovitin
2010-08-24 14:55 ` Matthew Wilcox
2010-08-24 17:51 ` Linux I/O subsystem performance Vladislav Bolkhovitin
2010-08-24 20:40 ` Matthew Wilcox
2010-08-24 14:55 ` [Scst-devel] Fwd: Re: linuxcon 2010 Chetan Loke
[not found] ` <4C7404C4.4040704@vlnb.net>
2010-08-24 20:31 ` Linux I/O subsystem performance (was: linuxcon 2010...) Chris Worley
2010-08-25 19:12 ` Vladislav Bolkhovitin [this message]
2010-09-16 15:05 ` Chris Worley
2010-08-23 19:41 ` [Scst-devel] Fwd: Re: linuxcon 2010 Vladislav Bolkhovitin
2010-08-24 14:41 ` Vladislav Bolkhovitin
2010-08-24 14:51 ` Chris Weiss
2010-08-24 14:56 ` Matthew Wilcox
2010-08-25 22:20 ` Konrad Rzeszutek Wilk
2010-08-25 22:45 ` Ted Ts'o
2010-08-24 14:57 ` James Bottomley
2010-08-24 19:48 ` Vladislav Bolkhovitin
2010-08-24 21:23 ` Nicholas A. Bellinger
2010-08-26 20:11 ` Vladislav Bolkhovitin
2010-08-26 21:23 ` Nicholas A. Bellinger
2010-08-28 17:32 ` Vladislav Bolkhovitin
2010-08-28 20:47 ` Nicholas A. Bellinger
2010-08-30 20:47 ` Vladislav Bolkhovitin
2010-08-30 21:46 ` Nicholas A. Bellinger
2010-09-02 19:38 ` Vladislav Bolkhovitin
2010-09-02 20:25 ` Nicholas A. Bellinger
2010-09-05 20:18 ` Dmitry Torokhov
2010-09-05 21:50 ` Nicholas A. Bellinger
2010-09-05 23:13 ` Mark Deneen
2010-09-06 0:12 ` Nicholas A. Bellinger
2010-09-06 0:58 ` Mark Deneen
2010-09-06 1:34 ` Nicholas A. Bellinger
2010-09-06 5:04 ` Dmitry Torokhov
2010-09-05 23:41 ` Dmitry Torokhov
2010-09-05 23:59 ` Nicholas A. Bellinger
2010-09-06 4:56 ` Dmitry Torokhov
2010-09-06 10:39 ` James Bottomley
2010-09-06 11:02 ` Bart Van Assche
2010-09-06 11:27 ` James Bottomley
2010-09-06 15:26 ` Vladislav Bolkhovitin
2010-09-06 21:47 ` Vladislav Bolkhovitin
2010-09-06 21:55 ` Nicholas A. Bellinger
2010-09-06 22:14 ` david
2010-09-07 0:44 ` Dmitry Torokhov
2010-09-07 3:45 ` Chetan Loke
2010-09-07 6:15 ` Bart Van Assche
2010-09-07 6:08 ` Bart Van Assche
2010-09-07 6:26 ` Dmitry Torokhov
2010-09-07 20:14 ` Vladislav Bolkhovitin
2010-09-07 20:14 ` Vladislav Bolkhovitin
2010-09-06 21:16 ` Greg KH
2010-09-06 17:28 ` Chetan Loke
2010-09-06 21:52 ` Vladislav Bolkhovitin
2010-08-18 17:51 ` Chetan Loke
2010-08-18 16:19 ` Bart Van Assche
2010-08-18 16:28 ` Joe Landman
2010-08-18 17:52 ` Vladislav Bolkhovitin
2010-08-18 15:12 ` Chetan Loke
2010-08-18 17:52 ` Vladislav Bolkhovitin
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=4C756B37.4000007@vlnb.net \
--to=vst@vlnb.net \
--cc=James.Bottomley@suse.de \
--cc=bvanassche@acm.org \
--cc=chetanloke@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=pasik@iki.fi \
--cc=scst-devel@lists.sourceforge.net \
--cc=worleys@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