From: Andi Kleen <ak@suse.de>
To: jbarnes@sgi.com (Jesse Barnes)
Cc: thockin@hockin.org, linux-kernel@vger.kernel.org
Subject: Re: Memory mapped IO vs Port IO
Date: Fri, 12 Sep 2003 20:29:36 +0200 [thread overview]
Message-ID: <20030912202936.2abdaf85.ak@suse.de> (raw)
In-Reply-To: <20030912182430.GA1043@sgi.com>
On Fri, 12 Sep 2003 11:24:30 -0700
jbarnes@sgi.com (Jesse Barnes) wrote:
> On Fri, Sep 12, 2003 at 11:11:48AM -0700, Tim Hockin wrote:
> > On Fri, Sep 12, 2003 at 08:00:03PM +0200, Andi Kleen wrote:
> > > jbarnes@sgi.com (Jesse Barnes) writes:
> > >
> > > > Ok, Andi asked for benchmarks, so I ran some. Let this should be a
> > > > lesson on why you shouldn't use port I/O :) I ran these on an SGI Altix
> > > > w/900 MHz McKinley processors.
> > > >
> > > > Just straight calls to the routines (all of these are based on the
> > > > average of 100 iterations):
> > > > writeq(val, reg) time: 64 cycles
> > > > outl(val, reg) time: 2126 cycles
> > > ^^^^^
> > > >
> > > > A simple branch:
> > > > if (use_mmio)
> > > > writeq(val, reg) time: 132 cycles
> > > > else
> > > > outl(val, reg) time: 1990 cycles
> > > ^^^^^
> > > Something seems to be wrong in your numbers.
> > >
> > > Surely the outl in the if () cannot be faster than the pure outl() ?
> >
> > Also - a perhaps more useful test is a write followed by a read.
>
> Well, someone else will have to run that test. On Altix, a read() is
> freakishly expensive, and I'm not really interested in showing everyone
> how bad it is ;)
I guess the read will be very bad everywhere, and the PIO inl even worse.
-Andi
next prev parent reply other threads:[~2003-09-12 18:32 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20030911192550.7dfaf08c.ak@suse.de.suse.lists.linux.kernel>
[not found] ` <1063308053.4430.37.camel@huykhoi.suse.lists.linux.kernel>
[not found] ` <20030912162713.GA4852@sgi.com.suse.lists.linux.kernel>
[not found] ` <20030912174807.GA629@sgi.com.suse.lists.linux.kernel>
2003-09-12 18:00 ` Memory mapped IO vs Port IO Andi Kleen
2003-09-12 18:04 ` Jesse Barnes
2003-09-12 18:09 ` Andi Kleen
2003-09-12 18:11 ` Tim Hockin
2003-09-12 18:24 ` Jesse Barnes
2003-09-12 18:29 ` Andi Kleen [this message]
2003-09-12 18:58 ` Martin J. Bligh
2003-09-12 19:51 ` Jesse Barnes
2003-09-12 16:41 John Bradford
2003-09-12 21:52 ` Mike Fedyk
2003-09-13 4:01 ` Mike Fedyk
[not found] <20030911160116.GI21596@parcelfarce.linux.theplanet.co.uk.suse.lists.linux.kernel>
2003-09-11 16:17 ` Andi Kleen
2003-09-11 16:25 ` Matthew Wilcox
2003-09-11 16:31 ` Andi Kleen
2003-09-11 16:42 ` Matthew Wilcox
2003-09-11 17:12 ` Jamie Lokier
2003-09-11 17:25 ` Andi Kleen
2003-09-12 1:39 ` jw schultz
2003-09-12 16:10 ` Anthony Dominic Truong
2003-09-12 16:22 ` Jamie Lokier
2003-09-12 16:27 ` Jesse Barnes
2003-09-12 17:48 ` Jesse Barnes
2003-09-11 17:13 ` Jamie Lokier
[not found] ` <20030911161450.GA23536@sgi.com.suse.lists.linux.kernel>
2003-09-11 16:20 ` Andi Kleen
-- strict thread matches above, loose matches on Subject: below --
2003-09-11 16:01 Matthew Wilcox
2003-09-11 16:14 ` Måns Rullgård
2003-09-11 16:14 ` Jesse Barnes
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=20030912202936.2abdaf85.ak@suse.de \
--to=ak@suse.de \
--cc=jbarnes@sgi.com \
--cc=linux-kernel@vger.kernel.org \
--cc=thockin@hockin.org \
/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).