From: ebiederm@xmission.com (Eric W. Biederman)
To: "David S. Miller" <davem@redhat.com>
Cc: mbligh@aracnet.com, hadi@cyberus.ca, tcw@tempest.prismnet.com,
linux-kernel@vger.kernel.org, netdev@oss.sgi.com, niv@us.ibm.com
Subject: Re: Early SPECWeb99 results on 2.5.33 with TSO on e1000
Date: 11 Sep 2002 09:31:53 -0600 [thread overview]
Message-ID: <m18z28o846.fsf@frodo.biederman.org> (raw)
In-Reply-To: <20020911.081521.103561835.davem@redhat.com>
"David S. Miller" <davem@redhat.com> writes:
> From: ebiederm@xmission.com (Eric W. Biederman)
> Date: 11 Sep 2002 09:06:36 -0600
>
> "Martin J. Bligh" <mbligh@aracnet.com> writes:
>
> > We can push about 420MB/s of IO out of this thing (out of that
> > theoretical 800Mb/s).
>
> Sounds about average for a P3. I have pushed the full 800MiB/s out of
> a P3 processor to memory but it was a very optimized loop.
>
> You pushed that over the PCI bus of your P3? Just to RAM
> doesn't count, lots of cpu's can do that.
>
> That's what makes his number interesting.
I agree. Getting 420MB/s to the pci bus is nice, especially with a P3.
The 800MB/s to memory was just the test I happened to conduct about 2 years
ago when I was still messing with slow P3 systems. It was a proof of
concept test to see if we could plug in an I/O card into a memory
slot.
On a current P4 system with the E7500 chipset this kind of thing is
easy. I have gotten roughly 450MB/s to a single myrinet card. And there
is enough theoretical bandwidth to do 4 times that. I haven't had a
chance to get it working in practice. When I attempted to run to gige
cards simultaneously I had some weird problem (probably interrupt
related) where adding additional pci cards did not deliver any extra
performance.
On a P3 to get writes from the cpu to hit 800MB/s you use the special
cpu instructions that bypass the cache.
My point was that I have tested the P3 bus in question and I achieved
a real world 800MB/s over it. So I expect that on the system in
question unless another bottleneck is hit, it should be possible to
achieve a real world 800MB/s of I/O. There are enough pci busses
to support that kind of traffic.
Unless the memory controller is carefully placed on the system though
doing 400+MB/s could easily eat up most of the available memory
bandwidth and reduce the system to doing some very slow cache line fills.
Eric
next prev parent reply other threads:[~2002-09-11 15:31 UTC|newest]
Thread overview: 101+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-09-05 18:30 Early SPECWeb99 results on 2.5.33 with TSO on e1000 Troy Wilson
2002-09-05 20:59 ` jamal
2002-09-05 22:11 ` Troy Wilson
2002-09-05 22:39 ` Nivedita Singhvi
2002-09-05 23:01 ` Dave Hansen
2002-09-05 22:48 ` Nivedita Singhvi
2002-09-06 1:47 ` jamal
2002-09-06 3:38 ` Nivedita Singhvi
2002-09-06 3:58 ` David S. Miller
2002-09-06 4:20 ` Nivedita Singhvi
2002-09-06 4:17 ` David S. Miller
2002-09-07 0:05 ` Troy Wilson
2002-09-06 3:56 ` David S. Miller
2002-09-06 3:47 ` David S. Miller
2002-09-06 6:48 ` Martin J. Bligh
2002-09-06 6:51 ` David S. Miller
2002-09-06 7:36 ` Andrew Morton
2002-09-06 7:22 ` David S. Miller
2002-09-06 9:54 ` jamal
2002-09-06 14:29 ` Martin J. Bligh
2002-09-06 15:38 ` Dave Hansen
2002-09-06 16:11 ` Martin J. Bligh
2002-09-06 16:21 ` Nivedita Singhvi
2002-09-06 15:29 ` Dave Hansen
2002-09-06 16:29 ` Martin J. Bligh
2002-09-06 17:36 ` Dave Hansen
2002-09-06 18:26 ` Andi Kleen
2002-09-06 18:33 ` Dave Hansen
2002-09-06 18:36 ` David S. Miller
2002-09-06 18:45 ` Martin J. Bligh
2002-09-06 18:43 ` David S. Miller
2002-09-06 19:19 ` Nivedita Singhvi
2002-09-06 19:21 ` David S. Miller
2002-09-06 19:45 ` Nivedita Singhvi
2002-09-06 19:26 ` Andi Kleen
2002-09-06 19:24 ` David S. Miller
2002-09-06 19:45 ` Martin J. Bligh
2002-09-06 17:26 ` Gerrit Huizenga
2002-09-06 17:37 ` David S. Miller
2002-09-06 18:19 ` Gerrit Huizenga
2002-09-06 18:26 ` Martin J. Bligh
2002-09-06 18:36 ` David S. Miller
2002-09-06 18:51 ` Martin J. Bligh
2002-09-06 18:48 ` David S. Miller
2002-09-06 19:05 ` Gerrit Huizenga
2002-09-06 19:01 ` David S. Miller
2002-09-06 20:29 ` Alan Cox
2002-09-06 18:34 ` David S. Miller
2002-09-06 18:57 ` Gerrit Huizenga
2002-09-06 18:58 ` David S. Miller
2002-09-06 19:52 ` Gerrit Huizenga
2002-09-06 19:49 ` David S. Miller
2002-09-06 20:03 ` Gerrit Huizenga
2002-09-06 23:48 ` Troy Wilson
2002-09-11 9:11 ` Eric W. Biederman
2002-09-11 14:10 ` Martin J. Bligh
2002-09-11 15:06 ` Eric W. Biederman
2002-09-11 15:15 ` David S. Miller
2002-09-11 15:31 ` Eric W. Biederman [this message]
2002-09-11 15:27 ` Martin J. Bligh
2002-09-12 7:28 ` Todd Underwood
2002-09-12 12:30 ` jamal
2002-09-12 13:57 ` Todd Underwood
2002-09-12 14:11 ` Alan Cox
2002-09-12 14:41 ` todd-lkml
2002-09-12 23:12 ` David S. Miller
2002-09-13 21:59 ` todd-lkml
2002-09-13 22:04 ` David S. Miller
2002-09-15 20:16 ` jamal
2002-09-16 4:23 ` David S. Miller
2002-09-16 14:16 ` todd-lkml
2002-09-16 19:52 ` David S. Miller
2002-09-16 21:32 ` todd-lkml
2002-09-16 21:29 ` David S. Miller
2002-09-16 22:53 ` David Woodhouse
2002-09-16 22:46 ` David S. Miller
2002-09-16 23:03 ` David Woodhouse
2002-09-16 23:08 ` Jeff Garzik
2002-09-16 23:02 ` David S. Miller
2002-09-16 23:48 ` Jeff Garzik
2002-09-16 23:43 ` David S. Miller
2002-09-17 0:01 ` Jeff Garzik
2002-09-17 10:31 ` jamal
2002-09-13 22:12 ` Nivedita Singhvi
2002-09-12 17:18 ` Nivedita Singhvi
2002-09-06 23:56 ` Troy Wilson
2002-09-06 23:52 ` David S. Miller
2002-09-07 0:18 ` Nivedita Singhvi
2002-09-07 0:27 ` Troy Wilson
-- strict thread matches above, loose matches on Subject: below --
2002-09-05 20:47 Feldman, Scott
2002-09-06 11:44 Robert Olsson
2002-09-06 14:37 ` Martin J. Bligh
2002-09-06 15:38 ` Robert Olsson
2002-09-06 18:35 Manfred Spraul
2002-09-06 18:38 ` David S. Miller
2002-09-06 19:40 ` Manfred Spraul
2002-09-06 19:34 ` David S. Miller
2002-09-10 12:02 ` Robert Olsson
2002-09-10 16:55 ` Manfred Spraul
2002-09-11 7:46 ` Robert Olsson
2002-09-10 14:59 Mala Anand
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=m18z28o846.fsf@frodo.biederman.org \
--to=ebiederm@xmission.com \
--cc=davem@redhat.com \
--cc=hadi@cyberus.ca \
--cc=linux-kernel@vger.kernel.org \
--cc=mbligh@aracnet.com \
--cc=netdev@oss.sgi.com \
--cc=niv@us.ibm.com \
--cc=tcw@tempest.prismnet.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).