From: Andrea Arcangeli <andrea@suse.de>
To: "Albert D. Cahalan" <acahalan@cs.uml.edu>
Cc: Theodore Tso <tytso@mit.edu>,
"David S. Miller" <davem@redhat.com>,
Miles Lane <miles@megapathdsl.net>,
linux-kernel@vger.kernel.org
Subject: Re: Kernel 2.5 Workshop RealVideo streams -- next time, please get better audio.
Date: Wed, 18 Apr 2001 17:07:02 +0200 [thread overview]
Message-ID: <20010418170702.B30770@athlon.random> (raw)
In-Reply-To: <20010417205722.A3626@think> <200104180246.f3I2kL1192784@saturn.cs.uml.edu>
In-Reply-To: <200104180246.f3I2kL1192784@saturn.cs.uml.edu>; from acahalan@cs.uml.edu on Tue, Apr 17, 2001 at 10:46:20PM -0400
On Tue, Apr 17, 2001 at 10:46:20PM -0400, Albert D. Cahalan wrote:
> support for NUMA hardware (it's not cache coherent) right now
btw, there are three kind of NUMA systems:
1) cc-numa first citizens (wildfire alpha, future chips)
2) cc-numa second citizens (origin2k)
3) non cache coherent numa machines
On the first class numa citizens NUMA means "heuristics for higher performance".
On those systems you don't need any NUMA change for correct operation of the
kernel (besides the fact you may need to use discontigmem to boot the kernel
if there can be huge physical holes in the physical layout of the ram but
that is true also for any other non numa machine with big holes in the physical
ram address space).
On the second and thrid class of NUMA systems NUMA means "required changes
for correct operations of the system". difference between 1 and 2 is that
category 2) needs also to put specialized PIO memory barriers to serialize the
I/O across different nodes. So it "only" additionaly requires total auditing of
the device drivers.
I think linux will need to optimize class 1 of systems and I assume SGI has the
PIO memory barriers patches for the device drivers to support class 2 as well.
Nobody ever considered the non cache coherent numa support so far AFIK and
I guess it will hardly end into mainline (personally I wouldn't be that
excited to deal with that additional complexity ;). If you can tell, what
system is it?
Andrea
next prev parent reply other threads:[~2001-04-18 14:54 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-04-17 0:45 Kernel 2.5 Workshop RealVideo streams -- next time, please get better audio Miles Lane
2001-04-17 0:53 ` David S. Miller
2001-04-17 1:37 ` Miles Lane
2001-04-18 0:57 ` Theodore Tso
2001-04-18 1:08 ` Miles Lane
2001-04-18 2:22 ` Joel Jaeggli
2001-04-18 2:46 ` Albert D. Cahalan
2001-04-18 11:58 ` Alan Cox
2001-04-18 15:07 ` Andrea Arcangeli [this message]
2001-04-17 1:53 ` Randolph Bentson
2001-04-17 1:56 ` Ben Ford
2001-04-17 3:43 ` Miles Lane
2001-04-17 4:47 ` Kernel 2.5 Workshop RealVideo streams -- next time, please get Albert D. Cahalan
2001-04-17 3:46 ` Kernel 2.5 Workshop RealVideo streams -- next time, please get better audio Miles Lane
2001-04-17 3:48 ` Larry McVoy
2001-04-17 3:54 ` Miles Lane
2001-04-17 3:56 ` Larry McVoy
2001-04-17 3:01 ` David Lang
2001-04-17 4:07 ` Miles Lane
2001-04-17 13:03 ` Mike A. Harris
2001-04-17 18:29 ` Kernel 2.5 Workshop RealVideo streams -- next time, please getbetter audio Miles Lane
2001-04-17 14:23 ` Kernel 2.5 Workshop RealVideo streams -- next time, please get better audio Jonathan Morton
2001-04-17 14:51 ` Kernel 2.5 Workshop RealVideo streams -- next time, please Alan Cox
2001-04-18 12:34 ` Kernel 2.5 Workshop RealVideo streams -- next time, please get better audio Eric W. Biederman
2001-04-18 12:44 ` Alan Cox
2001-04-18 18:29 ` Tim Wright
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=20010418170702.B30770@athlon.random \
--to=andrea@suse.de \
--cc=acahalan@cs.uml.edu \
--cc=davem@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=miles@megapathdsl.net \
--cc=tytso@mit.edu \
/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