From: Jeff Garzik <jgarzik@pobox.com>
To: Andrew Theurer <habanero@us.ibm.com>
Cc: "David S. Miller" <davem@redhat.com>,
haveblue@us.ibm.com, mbligh@aracnet.com, wli@holomorphy.com,
arjanv@redhat.com, pbadari@us.ibm.com,
linux-kernel@vger.kernel.org, gh@us.ibm.com, johnstul@us.ibm.com,
jamesclv@us.ibm.com, akpm@digeo.com, mannthey@us.ibm.com
Subject: Re: userspace irq balancer
Date: Tue, 20 May 2003 10:21:58 -0400 [thread overview]
Message-ID: <20030520142158.GA27982@gtf.org> (raw)
In-Reply-To: <200305200907.41443.habanero@us.ibm.com>
On Tue, May 20, 2003 at 09:07:41AM -0500, Andrew Theurer wrote:
> On Tuesday 20 May 2003 01:40, David S. Miller wrote:
> > From: Dave Hansen <haveblue@us.ibm.com>
> > Date: 19 May 2003 23:36:23 -0700
> >
> > I don't even think we can do that. That code was being integrated
> > around the same time that our Specweb setup decided to go south on us
> > and start physically frying itself.
> >
> > This gets more amusing by the second. Let's kill this code
> > already. People who like the current algorithms can push
> > them into the userspace solution.
>
> Remember this all started with some idea of "fairness" among cpus and very
> little to do with performance. particularly on P4 with HT, where the first
> logical cpu got all the ints and tasks running on that cpu were slower than
> other cpus. This was in most cases the highest performing situation, -but-
> it was unfair to the tasks running on cpu0. irq_balance fixed this with a
> random target cpu that was in theory supposed to not change often enough to
> preserve cache warmth. In practice is the target cpus changed too often
> which thrashed cache and the HW overhead of changing the destination that
> often was way way to high.
You call that a fix? ;-) I call that working around a bug.
If tasks run slower on cpuX than cpuY because of a heavier int load,
that's the fault of the scheduler not the irqbalancer, be it in-kernel
or userspace. If there's a lesser-utilized cpu the task needs to be
migrated to that cpu from the irq-loaded one, when CPU accounting
notices the kernel interrupt handling having an impact.
Jeff
next prev parent reply other threads:[~2003-05-20 14:09 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <200305191314.06216.pbadari@us.ibm.com>
2003-05-19 22:07 ` userspace irq balancer Dave Hansen
2003-05-19 22:11 ` Arjan van de Ven
2003-05-19 22:22 ` Dave Hansen
2003-05-20 3:25 ` David S. Miller
2003-05-20 3:46 ` William Lee Irwin III
2003-05-20 5:03 ` Dave Hansen
2003-05-20 5:53 ` Martin J. Bligh
2003-05-20 6:13 ` David S. Miller
2003-05-20 6:36 ` Dave Hansen
2003-05-20 6:40 ` David S. Miller
2003-05-20 14:07 ` Andrew Theurer
2003-05-20 14:21 ` Jeff Garzik [this message]
2003-05-20 14:35 ` Andrew Theurer
[not found] ` <20030520.163833.104040023.davem@redhat.com>
2003-05-21 14:58 ` Martin J. Bligh
2003-05-21 22:55 ` David S. Miller
2003-05-21 11:00 ` Kai Bankett
2003-05-20 14:01 ` Martin J. Bligh
2003-05-20 9:00 ` Arjan van de Ven
2003-05-20 9:14 ` William Lee Irwin III
2003-05-20 9:17 ` Andrew Morton
[not found] ` <20030520.172230.102567463.davem@redhat.com>
2003-05-21 14:27 ` James Cleverdon
2003-05-20 15:41 Nakajima, Jun
2003-05-21 13:54 ` James Cleverdon
2003-05-21 22:56 ` Zwane Mwaikambo
2003-05-21 16:31 James Bottomley
2003-05-21 20:16 ` Arjan van de Ven
2003-05-21 21:43 Nakajima, Jun
2003-05-22 0:29 ` Gerrit Huizenga
2003-05-22 1:28 ` Martin J. Bligh
2003-05-22 1:44 ` Gerrit Huizenga
2003-05-22 2:03 ` William Lee Irwin III
2003-05-22 2:04 ` William Lee Irwin III
2003-05-22 2:12 ` Zwane Mwaikambo
2003-05-22 3:57 ` Martin J. Bligh
2003-05-22 17:24 ` Bill Davidsen
2003-05-22 22:44 ` David S. Miller
2003-05-26 22:24 ` Andrea Arcangeli
2003-05-26 23:26 ` Andrew Morton
2003-05-26 23:34 ` Andrea Arcangeli
2003-05-26 23:43 ` David S. Miller
[not found] ` <20030527000639.GA3767@dualathlon.random>
2003-05-27 0:15 ` David S. Miller
2003-05-27 0:41 ` Andrea Arcangeli
2003-05-27 0:48 ` David S. Miller
2003-05-27 1:09 ` Andrea Arcangeli
2003-05-27 1:13 ` David S. Miller
2003-05-27 1:26 ` Andrea Arcangeli
2003-05-27 6:11 ` David S. Miller
2003-05-27 11:53 ` Andrea Arcangeli
2003-05-27 22:04 ` David S. Miller
2003-05-27 22:27 ` Andrea Arcangeli
2003-05-27 23:55 ` David S. Miller
2003-06-13 6:22 ` David S. Miller
2003-06-13 18:23 ` Andrea Arcangeli
2003-05-27 1:16 ` Dave Jones
2003-05-27 1:17 ` David S. Miller
2003-05-27 9:07 ` Arjan van de Ven
2003-05-27 9:10 ` David S. Miller
2003-05-27 1:28 ` Andrea Arcangeli
2003-05-27 1:53 ` William Lee Irwin III
2003-05-27 1:59 ` Andrew Morton
2003-05-27 2:10 ` William Lee Irwin III
2003-05-27 2:15 ` Zwane Mwaikambo
2003-05-27 2:44 ` William Lee Irwin III
2003-05-27 2:45 ` Zwane Mwaikambo
2003-05-27 4:22 ` William Lee Irwin III
2003-05-27 2:15 ` Andrea Arcangeli
2003-05-27 2:14 ` Andrea Arcangeli
2003-05-27 2:26 ` William Lee Irwin III
2003-05-27 1:17 ` Andrea Arcangeli
2003-05-27 1:20 ` David S. Miller
2003-05-27 1:33 ` Andrea Arcangeli
2003-05-22 14:18 ` James Cleverdon
2003-05-22 14:43 ` William Lee Irwin III
2003-05-22 15:30 ` James Cleverdon
2003-05-22 15:45 ` William Lee Irwin III
2003-05-24 1:10 Nakajima, Jun
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=20030520142158.GA27982@gtf.org \
--to=jgarzik@pobox.com \
--cc=akpm@digeo.com \
--cc=arjanv@redhat.com \
--cc=davem@redhat.com \
--cc=gh@us.ibm.com \
--cc=habanero@us.ibm.com \
--cc=haveblue@us.ibm.com \
--cc=jamesclv@us.ibm.com \
--cc=johnstul@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mannthey@us.ibm.com \
--cc=mbligh@aracnet.com \
--cc=pbadari@us.ibm.com \
--cc=wli@holomorphy.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).