From: Dipankar Sarma <dipankar@in.ibm.com>
To: Paul Jackson <pj@sgi.com>
Cc: Andrew Morton <akpm@osdl.org>,
nagar@watson.ibm.com, ckrm-tech@lists.sourceforge.net,
efocht@hpce.nec.com, mbligh@aracnet.com,
lse-tech@lists.sourceforge.net, hch@infradead.org,
steiner@sgi.com, jbarnes@sgi.com, sylvain.jeaugey@bull.net,
djh@sgi.com, linux-kernel@vger.kernel.org, colpatch@us.ibm.com,
Simon.Derr@bull.net, ak@suse.de, sivanich@sgi.com
Subject: Re: [Lse-tech] [PATCH] cpusets - big numa cpu and memory placement
Date: Sat, 2 Oct 2004 20:25:21 +0530 [thread overview]
Message-ID: <20041002145521.GA8868@in.ibm.com> (raw)
In-Reply-To: <20041001230644.39b551af.pj@sgi.com>
On Fri, Oct 01, 2004 at 11:06:44PM -0700, Paul Jackson wrote:
> Cpuset Requirements
> ===================
>
> The three primary requirements that the SGI support engineers
> on our biggest configurations keep telling me are most important
> are:
> 1) isolation,
> 2) isolation, and
> 3) isolation.
> A big HPC job running on a dedicated set of CPUs and Memory Nodes
> should not lose any CPU cycles or Memory pages to outsiders.
>
....
>
> A job running in a cpuset should be able to use various configuration,
> resource management (CKRM for example), cpu and memory (numa) affinity
> tools, performance analysis and thread management facilities within a
> set, including pthreads and MPI, independently from what is happening
> on the rest of the system.
>
> One should be able to run a stock 3rd party app (Oracle is
> the canonical example) on a system side-by-side with a special
> customer app, each in their own set, neither interfering with
> the other, and the Oracle folks happy that their app is running
> in a supported environment.
One of the things we are working on is to provide exactly something
like this. Not just that, within the isolated partitions, we want
to be able to provide completely different environment. For example,
we need to be able to run or more realtime processes of an application
in one partition while the other partition runs the database portion
of the application. For this to succeed, they need to be completely
isolated.
It would be nice if someone explains a potential CKRM implementation for
this kind of complete isolation.
Thanks
Dipankar
next prev parent reply other threads:[~2004-10-02 14:53 UTC|newest]
Thread overview: 233+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-08-05 10:08 [PATCH] new bitmap list format (for cpusets) Paul Jackson
2004-08-05 10:10 ` [PATCH] cpusets - big numa cpu and memory placement Paul Jackson
2004-08-05 20:55 ` [Lse-tech] " Martin J. Bligh
2004-08-06 2:05 ` Paul Jackson
2004-08-06 3:24 ` Martin J. Bligh
2004-08-06 8:31 ` Paul Jackson
2004-08-06 15:30 ` Erich Focht
2004-08-06 15:35 ` Martin J. Bligh
2004-08-06 15:48 ` Hubertus Franke
2004-08-07 6:30 ` Paul Jackson
2004-08-07 6:45 ` Paul Jackson
2004-08-06 15:49 ` Hubertus Franke
2004-08-06 15:52 ` Hubertus Franke
2004-08-06 15:55 ` Erich Focht
2004-08-07 6:10 ` Paul Jackson
2004-08-07 15:22 ` Erich Focht
2004-08-07 18:59 ` Paul Jackson
2004-08-08 3:17 ` Paul Jackson
2004-08-08 14:50 ` Martin J. Bligh
2004-08-11 0:43 ` Paul Jackson
2004-08-11 9:40 ` Erich Focht
2004-08-11 14:49 ` Martin J. Bligh
2004-08-11 17:50 ` Paul Jackson
2004-08-11 21:12 ` Shailabh Nagar
2004-08-12 7:15 ` Paul Jackson
2004-08-12 12:58 ` Jack Steiner
2004-08-12 14:50 ` Martin J. Bligh
2004-08-11 15:12 ` Shailabh Nagar
2004-08-08 20:22 ` Shailabh Nagar
2004-08-09 15:57 ` Hubertus Franke
2004-08-10 11:31 ` [ckrm-tech] " Paul Jackson
2004-08-10 22:38 ` Shailabh Nagar
2004-08-11 10:42 ` Erich Focht
2004-08-11 14:56 ` Shailabh Nagar
2004-08-14 8:51 ` Paul Jackson
2004-08-08 19:58 ` Shailabh Nagar
2004-10-01 23:41 ` Andrew Morton
2004-10-02 6:06 ` Paul Jackson
2004-10-02 14:55 ` Dipankar Sarma [this message]
2004-10-02 16:14 ` Hubertus Franke
2004-10-02 18:04 ` Paul Jackson
2004-10-02 23:21 ` Peter Williams
2004-10-02 23:44 ` Hubertus Franke
2004-10-03 0:00 ` Peter Williams
2004-10-03 3:44 ` Paul Jackson
2004-10-05 3:13 ` [ckrm-tech] " Matthew Helsley
2004-10-05 8:30 ` Hubertus Franke
2004-10-05 14:20 ` Paul Jackson
2004-10-03 2:59 ` Paul Jackson
2004-10-03 3:19 ` Paul Jackson
2004-10-03 3:53 ` Peter Williams
2004-10-03 4:47 ` Paul Jackson
2004-10-03 5:12 ` Peter Williams
2004-10-03 5:39 ` Paul Jackson
2004-10-03 4:02 ` Paul Jackson
2004-10-03 3:39 ` Paul Jackson
2004-10-03 14:36 ` Martin J. Bligh
2004-10-03 15:39 ` Paul Jackson
2004-10-03 23:53 ` Martin J. Bligh
2004-10-04 0:02 ` Martin J. Bligh
2004-10-04 0:53 ` Paul Jackson
2004-10-04 3:56 ` Martin J. Bligh
2004-10-04 4:24 ` Paul Jackson
2004-10-04 15:03 ` Martin J. Bligh
2004-10-04 15:53 ` [ckrm-tech] " Paul Jackson
2004-10-04 18:17 ` Martin J. Bligh
2004-10-04 20:25 ` Paul Jackson
2004-10-04 22:15 ` Martin J. Bligh
2004-10-05 9:17 ` Paul Jackson
2004-10-05 10:01 ` Paul Jackson
2004-10-05 22:24 ` Matthew Dobson
2004-10-05 9:26 ` Simon Derr
2004-10-05 9:58 ` Paul Jackson
2004-10-05 19:34 ` Martin J. Bligh
2004-10-06 0:28 ` Paul Jackson
2004-10-06 1:16 ` Martin J. Bligh
2004-10-06 2:08 ` Paul Jackson
2004-10-06 22:59 ` Matthew Dobson
2004-10-06 23:23 ` Peter Williams
2004-10-07 0:16 ` Rick Lindsley
2004-10-07 18:27 ` Paul Jackson
2004-10-07 8:51 ` Paul Jackson
2004-10-07 10:53 ` Rick Lindsley
2004-10-07 14:41 ` Martin J. Bligh
[not found] ` <20041007072842.2bafc320.pj@sgi.com>
2004-10-07 19:05 ` Rick Lindsley
2004-10-10 2:15 ` [ckrm-tech] " Paul Jackson
2004-10-11 22:06 ` Matthew Dobson
2004-10-11 22:58 ` Paul Jackson
2004-10-12 21:22 ` Matthew Dobson
2004-10-12 8:50 ` Simon Derr
2004-10-12 21:25 ` Matthew Dobson
2004-10-10 2:28 ` Paul Jackson
2004-10-09 0:06 ` Matthew Dobson
[not found] ` <4165A31E.4070905@watson.ibm.com>
2004-10-08 13:14 ` Paul Jackson
2004-10-08 15:42 ` Hubertus Franke
2004-10-08 18:23 ` Paul Jackson
2004-10-09 1:00 ` Matthew Dobson
2004-10-09 20:08 ` [Lse-tech] " Paul Jackson
2004-10-11 22:16 ` Matthew Dobson
2004-10-11 22:42 ` Paul Jackson
2004-10-10 0:05 ` Paul Jackson
2004-10-11 22:18 ` Matthew Dobson
2004-10-11 22:39 ` Paul Jackson
2004-10-09 0:51 ` Matthew Dobson
2004-10-10 0:50 ` [Lse-tech] " Paul Jackson
2004-10-10 0:59 ` Paul Jackson
2004-10-09 0:22 ` Matthew Dobson
2004-10-12 22:24 ` [Lse-tech] " Hanna Linder
2004-10-13 20:56 ` Matthew Dobson
2004-10-07 12:47 ` [Lse-tech] " Simon Derr
2004-10-07 14:49 ` Martin J. Bligh
2004-10-07 17:54 ` Paul Jackson
2004-10-07 18:13 ` Martin J. Bligh
2004-10-08 9:23 ` Erich Focht
2004-10-08 9:50 ` Andrew Morton
2004-10-08 10:40 ` Erich Focht
2004-10-08 14:26 ` Martin J. Bligh
2004-10-08 9:53 ` Nick Piggin
2004-10-08 11:40 ` Erich Focht
2004-10-08 14:24 ` Martin J. Bligh
2004-10-08 22:37 ` Erich Focht
2004-10-14 10:35 ` Eric W. Biederman
2004-10-14 11:22 ` Erich Focht
2004-10-14 11:23 ` Paul Jackson
2004-10-14 19:39 ` Paul Jackson
2004-10-14 22:38 ` Hubertus Franke
2004-10-15 1:26 ` Paul Jackson
2004-10-07 18:25 ` Andrew Morton
2004-10-07 19:52 ` Paul Jackson
2004-10-07 21:04 ` [ckrm-tech] " Matthew Helsley
2004-10-10 3:22 ` Paul Jackson
2004-10-07 19:16 ` Rick Lindsley
2004-10-10 2:35 ` Paul Jackson
2004-10-10 5:12 ` [ckrm-tech] " Paul Jackson
2004-10-08 23:48 ` Matthew Dobson
2004-10-09 0:18 ` Nick Piggin
2004-10-11 23:00 ` Matthew Dobson
2004-10-11 23:09 ` Nick Piggin
2004-10-05 22:33 ` Matthew Dobson
2004-10-06 3:01 ` Paul Jackson
2004-10-06 23:12 ` Matthew Dobson
2004-10-07 8:59 ` [ckrm-tech] " Paul Jackson
2004-10-04 0:45 ` Paul Jackson
2004-10-04 11:44 ` Rick Lindsley
2004-10-04 22:46 ` [ckrm-tech] " Paul Jackson
2004-10-05 22:19 ` Matthew Dobson
2004-10-06 2:39 ` Paul Jackson
2004-10-06 23:21 ` Matthew Dobson
2004-10-07 9:41 ` [ckrm-tech] " Paul Jackson
2004-10-06 2:47 ` Paul Jackson
2004-10-06 9:43 ` Simon Derr
2004-10-06 13:27 ` Paul Jackson
2004-10-06 21:55 ` Peter Williams
2004-10-06 22:49 ` Paul Jackson
2004-10-06 8:02 ` Simon Derr
2005-02-07 23:59 ` Matthew Dobson
2005-02-08 0:20 ` Andrew Morton
2005-02-08 0:34 ` Paul Jackson
2005-02-08 9:54 ` Dinakar Guniguntala
2005-02-08 9:49 ` Nick Piggin
2005-02-08 16:13 ` Martin J. Bligh
2005-02-08 23:26 ` Nick Piggin
2005-02-09 4:23 ` Paul Jackson
2005-02-08 19:32 ` Matthew Dobson
2005-02-09 2:53 ` Nick Piggin
2005-02-08 19:00 ` Matthew Dobson
2005-02-08 20:42 ` Paul Jackson
2005-02-08 22:14 ` Matthew Dobson
2005-02-08 23:58 ` Shailabh Nagar
2005-02-09 0:27 ` Paul Jackson
2005-02-09 0:24 ` Paul Jackson
2005-02-09 17:59 ` [ckrm-tech] " Chandra Seetharaman
2005-02-11 2:46 ` Chandra Seetharaman
2005-02-11 9:21 ` Paul Jackson
2005-02-12 1:37 ` Chandra Seetharaman
2005-02-12 6:16 ` Paul Jackson
2005-02-11 16:54 ` Jesse Barnes
2005-02-11 18:42 ` Chandra Seetharaman
2005-02-11 18:50 ` Jesse Barnes
2005-02-08 16:15 ` Martin J. Bligh
2005-02-08 22:17 ` Matthew Dobson
2004-10-03 16:02 ` Paul Jackson
2004-10-03 23:47 ` Martin J. Bligh
2004-10-04 3:33 ` Paul Jackson
2004-10-03 20:10 ` Tim Hockin
2004-10-04 1:56 ` Paul Jackson
2004-10-03 3:35 ` Paul Jackson
2004-10-03 20:21 ` Erich Focht
2004-10-03 20:48 ` Andrew Morton
2004-10-04 14:05 ` Erich Focht
2004-10-04 14:57 ` Martin J. Bligh
2004-10-04 15:30 ` Paul Jackson
2004-10-04 15:41 ` Martin J. Bligh
2004-10-04 16:02 ` Paul Jackson
2004-10-04 18:19 ` Martin J. Bligh
2004-10-04 18:29 ` Paul Jackson
2004-10-04 15:38 ` Paul Jackson
2004-10-04 16:46 ` Paul Jackson
2004-10-04 3:41 ` Paul Jackson
2004-10-04 13:58 ` Hubertus Franke
2004-10-04 14:13 ` Simon Derr
2004-10-04 14:15 ` Erich Focht
2004-10-04 15:23 ` Paul Jackson
2004-10-04 14:37 ` Paul Jackson
2004-10-02 15:46 ` [ckrm-tech] " Marc E. Fiuczynski
2004-10-02 16:17 ` Hubertus Franke
2004-10-02 17:53 ` Paul Jackson
2004-10-02 18:16 ` Hubertus Franke
2004-10-02 19:14 ` Paul Jackson
2004-10-02 23:29 ` Peter Williams
2004-10-02 23:51 ` Hubertus Franke
2004-10-02 20:40 ` Andrew Morton
2004-10-02 23:08 ` Hubertus Franke
2004-10-02 22:26 ` Alan Cox
2004-10-03 2:49 ` Paul Jackson
2004-10-03 12:19 ` Hubertus Franke
2004-10-03 3:25 ` Paul Jackson
2004-10-03 2:26 ` Paul Jackson
2004-10-03 14:11 ` Paul Jackson
2004-10-02 17:47 ` Paul Jackson
2004-08-05 20:47 ` [Lse-tech] [PATCH] new bitmap list format (for cpusets) Martin J. Bligh
2004-08-05 21:45 ` Paul Jackson
[not found] ` <Pine.A41.4.53.0408060930100.20680@isabelle.frec.bull.fr>
2004-08-06 10:14 ` Paul Jackson
2004-08-09 8:01 ` Paul Jackson
2004-08-09 14:49 ` Martin J. Bligh
2004-08-10 23:43 ` Paul Jackson
2004-08-11 13:11 ` Dinakar Guniguntala
2004-08-11 16:17 ` Paul Jackson
2004-08-11 18:05 ` Dinakar Guniguntala
2004-08-11 20:40 ` Paul Jackson
2004-08-12 9:48 ` Dinakar Guniguntala
2004-08-12 10:11 ` Paul Jackson
2004-08-12 12:34 ` Dinakar Guniguntala
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=20041002145521.GA8868@in.ibm.com \
--to=dipankar@in.ibm.com \
--cc=Simon.Derr@bull.net \
--cc=ak@suse.de \
--cc=akpm@osdl.org \
--cc=ckrm-tech@lists.sourceforge.net \
--cc=colpatch@us.ibm.com \
--cc=djh@sgi.com \
--cc=efocht@hpce.nec.com \
--cc=hch@infradead.org \
--cc=jbarnes@sgi.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lse-tech@lists.sourceforge.net \
--cc=mbligh@aracnet.com \
--cc=nagar@watson.ibm.com \
--cc=pj@sgi.com \
--cc=sivanich@sgi.com \
--cc=steiner@sgi.com \
--cc=sylvain.jeaugey@bull.net \
/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