From: Paul Jackson <pj@sgi.com>
To: Linus Torvalds <torvalds@osdl.org>
Cc: mingo@elte.hu, andy@thermo.lanl.gov, mbligh@mbligh.org,
akpm@osdl.org, arjan@infradead.org, arjanv@infradead.org,
haveblue@us.ibm.com, kravetz@us.ibm.com,
lhms-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org,
linux-mm@kvack.org, mel@csn.ul.ie, nickpiggin@yahoo.com.au
Subject: Re: [Lhms-devel] [PATCH 0/7] Fragmentation Avoidance V19
Date: Sun, 6 Nov 2005 10:18:32 -0800 [thread overview]
Message-ID: <20051106101832.510b0245.pj@sgi.com> (raw)
In-Reply-To: <Pine.LNX.4.64.0511060746170.3316@g5.osdl.org>
Linus wrote:
> The thing is, if 99.8% of memory is cleanable, the 0.2% is still enough to
> make pretty much _every_ hugepage in the system pinned down.
Agreed.
I realized after writing this that I wasn't clear on something.
I wasn't focused the subject of this thread, adding hugetlb pages after
the system has been up a while.
I was focusing on a related subject - freeing up most of the ordinary
size pages on the dedicated application nodes between jobs on a large
system using
* a bootcpuset (for the classic Unix load) and
* dedicated nodes (for the HPC apps).
I am looking to provide the combination of:
1) specifying some hugetlb pages at system boot, plus
2) the ability to clean off most of the ordinary sized pages
from the application nodes between jobs.
Perhaps Andy or some of my HPC customers wish I was also looking
to provide:
3) the ability to add lots of hugetlb pages on the application
nodes after the system has run a while.
But if they are, then they have some more educatin' to do on me.
For now, I am sympathetic to your concerns with code and locking
complexity. Freeing up great globs of hugetlb sized contiguous chunks
of memory after a system has run a while would be hard.
We have to be careful which hard problems we decide to take on.
We can't take on too many, and we have to pick ones that will provide
a major long term advantage to Linux, over the forseeable changes in
system hardware and architecture.
Even if most of the processors that Andy has tested against would
benefit from dynamically added hugetlb pages, if we can anticipate
that this will not be a substained opportunity for Linux (and looking
at current x86 chips doesn't require much anticipating) then that
might not be the place to invest our precious core complexity dollars.
--
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <pj@sgi.com> 1.925.600.0401
next prev parent reply other threads:[~2005-11-06 18:19 UTC|newest]
Thread overview: 216+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-11-04 1:00 Andy Nelson
2005-11-04 1:16 ` Martin J. Bligh
2005-11-04 1:27 ` Nick Piggin
2005-11-04 5:14 ` Linus Torvalds
2005-11-04 6:10 ` Paul Jackson
2005-11-04 6:38 ` Ingo Molnar
2005-11-04 7:26 ` Paul Jackson
2005-11-04 7:37 ` Ingo Molnar
2005-11-04 15:31 ` Linus Torvalds
2005-11-04 15:39 ` Martin J. Bligh
2005-11-04 15:53 ` Ingo Molnar
2005-11-06 7:34 ` Paul Jackson
2005-11-06 15:55 ` Linus Torvalds
2005-11-06 18:18 ` Paul Jackson [this message]
2005-11-06 8:44 ` Kyle Moffett
2005-11-06 16:12 ` Linus Torvalds
2005-11-06 17:00 ` Linus Torvalds
2005-11-07 1:28 ` Best CPU chipset for Linux? (was: [Lhms-devel] [PATCH 0/7] Fragmentation Avoidance V19) John Stoffel
2005-11-07 2:08 ` Linus Torvalds
2005-11-07 3:19 ` John Stoffel
2005-11-07 8:00 ` [Lhms-devel] [PATCH 0/7] Fragmentation Avoidance V19 Ingo Molnar
2005-11-07 11:00 ` Dave Hansen
2005-11-07 12:20 ` Ingo Molnar
2005-11-07 19:34 ` Steven Rostedt
2005-11-07 23:38 ` Joel Schopp
2005-11-13 2:30 ` Rob Landley
2005-11-14 1:58 ` Joel Schopp
2005-11-04 7:44 ` Eric Dumazet
2005-11-07 16:42 ` Adam Litke
2005-11-04 14:56 ` Andy Nelson
2005-11-04 15:18 ` Ingo Molnar
2005-11-04 15:39 ` Andy Nelson
2005-11-04 16:05 ` Ingo Molnar
2005-11-04 16:07 ` Linus Torvalds
2005-11-04 16:40 ` Ingo Molnar
2005-11-04 17:22 ` Linus Torvalds
2005-11-04 17:43 ` Andy Nelson
2005-11-04 16:00 ` Linus Torvalds
2005-11-04 16:13 ` Martin J. Bligh
2005-11-04 16:40 ` Linus Torvalds
2005-11-04 17:10 ` Martin J. Bligh
2005-11-04 16:14 ` Andy Nelson
2005-11-04 16:49 ` Linus Torvalds
-- strict thread matches above, loose matches on Subject: below --
2005-11-05 1:52 Seth, Rohit
2005-11-05 1:37 Seth, Rohit
2005-11-07 0:34 ` Andy Nelson
2005-11-07 18:58 ` Adam Litke
2005-11-07 20:51 ` Rohit Seth
2005-11-07 20:55 ` Andy Nelson
2005-11-07 20:58 ` Martin J. Bligh
2005-11-07 21:20 ` Rohit Seth
2005-11-07 21:33 ` Adam Litke
2005-11-08 2:12 ` David Gibson
2005-11-07 21:11 ` Adam Litke
2005-11-07 21:31 ` Rohit Seth
2005-11-04 21:51 Andy Nelson
2005-11-04 17:56 Andy Nelson
2005-11-04 17:03 Andy Nelson
2005-11-04 17:49 ` Linus Torvalds
2005-11-04 17:51 ` Andy Nelson
2005-11-04 20:12 ` Ingo Molnar
2005-11-04 21:04 ` Andy Nelson
2005-11-04 21:14 ` Ingo Molnar
2005-11-04 21:22 ` Linus Torvalds
2005-11-04 21:39 ` Linus Torvalds
2005-11-05 2:48 ` Rob Landley
2005-11-06 10:59 ` Paul Jackson
2005-11-04 21:31 ` Gregory Maxwell
2005-11-04 22:43 ` Andi Kleen
2005-11-05 0:07 ` Nick Piggin
2005-11-06 1:30 ` Zan Lynx
2005-11-06 2:25 ` Rob Landley
2005-11-04 15:19 Andy Nelson
2005-10-30 18:33 Mel Gorman
2005-10-31 5:57 ` [Lhms-devel] " Mike Kravetz
2005-10-31 6:37 ` Nick Piggin
2005-10-31 7:54 ` Andrew Morton
2005-10-31 7:11 ` Nick Piggin
2005-10-31 16:19 ` Mel Gorman
2005-10-31 23:54 ` Nick Piggin
2005-11-01 1:28 ` Mel Gorman
2005-11-01 1:42 ` Nick Piggin
[not found] ` <27700000.1130769270@[10.10.2.4]>
[not found] ` <20051031112409.153e7048.akpm@osdl.org>
[not found] ` <3660000.1130787652@flay>
2005-10-31 23:59 ` Nick Piggin
2005-11-01 1:36 ` Mel Gorman
[not found] ` <4366A8D1.7020507@yahoo.com.au>
[not found] ` <Pine.LNX.4.58.0510312333240.29390@skynet>
[not found] ` <4366C559.5090504@yahoo.com.au>
2005-11-01 15:25 ` Martin J. Bligh
2005-11-01 15:33 ` Dave Hansen
2005-11-01 16:57 ` Mel Gorman
2005-11-01 17:00 ` Mel Gorman
2005-11-01 18:58 ` Rob Landley
[not found] ` <Pine.LNX.4.58.0511010137020.29390@skynet>
[not found] ` <4366D469.2010202@yahoo.com.au>
[not found] ` <Pine.LNX.4.58.0511011014060.14884@skynet>
2005-11-01 13:56 ` Ingo Molnar
2005-11-01 14:10 ` Dave Hansen
2005-11-01 14:29 ` Ingo Molnar
2005-11-01 14:49 ` Dave Hansen
2005-11-01 15:01 ` Ingo Molnar
2005-11-01 15:22 ` Dave Hansen
[not found] ` <20051102084946.GA3930@elte.hu>
[not found] ` <436880B8.1050207@yahoo.com.au>
2005-11-02 9:32 ` Dave Hansen
2005-11-02 9:48 ` Nick Piggin
2005-11-02 10:54 ` Dave Hansen
2005-11-02 15:02 ` Martin J. Bligh
2005-11-03 3:21 ` Nick Piggin
2005-11-03 15:36 ` Martin J. Bligh
2005-11-03 15:40 ` Arjan van de Ven
2005-11-03 15:51 ` Linus Torvalds
2005-11-03 15:57 ` Martin J. Bligh
2005-11-03 16:20 ` Arjan van de Ven
2005-11-03 16:27 ` Mel Gorman
2005-11-03 16:46 ` Linus Torvalds
2005-11-03 16:52 ` Martin J. Bligh
2005-11-03 17:19 ` Linus Torvalds
2005-11-03 17:48 ` Dave Hansen
2005-11-03 17:51 ` Martin J. Bligh
2005-11-03 17:59 ` Arjan van de Ven
2005-11-03 18:08 ` Linus Torvalds
2005-11-03 18:17 ` Martin J. Bligh
2005-11-03 18:44 ` Linus Torvalds
2005-11-03 18:51 ` Martin J. Bligh
2005-11-03 19:35 ` Linus Torvalds
2005-11-03 22:40 ` Martin J. Bligh
2005-11-03 22:56 ` Linus Torvalds
2005-11-03 23:01 ` Martin J. Bligh
2005-11-04 0:58 ` Nick Piggin
2005-11-04 1:06 ` Linus Torvalds
2005-11-04 1:20 ` Paul Mackerras
2005-11-04 1:22 ` Nick Piggin
2005-11-04 1:48 ` Mel Gorman
2005-11-04 1:59 ` Nick Piggin
2005-11-04 2:35 ` Mel Gorman
2005-11-04 1:26 ` Mel Gorman
2005-11-03 21:11 ` Mel Gorman
2005-11-03 18:03 ` Linus Torvalds
2005-11-03 20:00 ` Paul Jackson
2005-11-03 20:46 ` Mel Gorman
2005-11-03 18:48 ` Martin J. Bligh
2005-11-03 19:08 ` Linus Torvalds
2005-11-03 22:37 ` Martin J. Bligh
2005-11-03 23:16 ` Linus Torvalds
2005-11-03 23:39 ` Martin J. Bligh
2005-11-04 0:42 ` Nick Piggin
2005-11-04 4:39 ` Andrew Morton
2005-11-04 16:22 ` Mel Gorman
2005-11-03 15:53 ` Martin J. Bligh
2005-11-01 16:48 ` Kamezawa Hiroyuki
2005-11-01 16:59 ` Kamezawa Hiroyuki
2005-11-01 17:19 ` Mel Gorman
2005-11-02 0:32 ` KAMEZAWA Hiroyuki
2005-11-02 11:22 ` Mel Gorman
2005-11-01 18:06 ` linux-os (Dick Johnson)
2005-11-02 7:19 ` Ingo Molnar
2005-11-02 7:46 ` Gerrit Huizenga
2005-11-02 8:50 ` Nick Piggin
2005-11-02 9:12 ` Gerrit Huizenga
2005-11-02 9:37 ` Nick Piggin
2005-11-02 10:17 ` Gerrit Huizenga
2005-11-02 23:47 ` Rob Landley
2005-11-03 4:43 ` Nick Piggin
2005-11-03 6:07 ` Rob Landley
2005-11-03 7:34 ` Nick Piggin
2005-11-03 17:54 ` Rob Landley
2005-11-03 20:13 ` Jeff Dike
2005-11-03 16:35 ` Jeff Dike
2005-11-03 16:23 ` Badari Pulavarty
2005-11-03 18:27 ` Jeff Dike
2005-11-03 18:49 ` Rob Landley
2005-11-04 4:52 ` Andrew Morton
2005-11-04 5:35 ` Paul Jackson
2005-11-04 5:48 ` Andrew Morton
2005-11-04 6:42 ` Paul Jackson
2005-11-04 7:10 ` Andrew Morton
2005-11-04 7:45 ` Paul Jackson
2005-11-04 8:02 ` Andrew Morton
2005-11-04 9:52 ` Paul Jackson
2005-11-04 15:27 ` Martin J. Bligh
2005-11-04 15:19 ` Martin J. Bligh
2005-11-04 17:38 ` Andrew Morton
2005-11-04 6:16 ` Bron Nelson
2005-11-02 10:41 ` Ingo Molnar
2005-11-02 11:04 ` Gerrit Huizenga
2005-11-02 12:00 ` Ingo Molnar
2005-11-02 12:42 ` Dave Hansen
2005-11-02 15:02 ` Gerrit Huizenga
2005-11-03 0:10 ` Rob Landley
2005-11-02 7:57 ` Nick Piggin
2005-11-02 0:51 ` Nick Piggin
2005-11-02 7:42 ` Dave Hansen
2005-11-02 8:24 ` Nick Piggin
2005-11-02 8:33 ` Yasunori Goto
2005-11-02 8:43 ` Nick Piggin
2005-11-02 14:51 ` Martin J. Bligh
2005-11-02 23:28 ` Rob Landley
2005-11-03 5:26 ` Jeff Dike
2005-11-03 5:41 ` Rob Landley
2005-11-01 14:41 ` Mel Gorman
2005-11-01 14:46 ` Ingo Molnar
2005-11-01 15:23 ` Mel Gorman
2005-11-01 18:33 ` Rob Landley
2005-11-01 19:02 ` Ingo Molnar
2005-11-01 14:50 ` Dave Hansen
2005-11-01 15:24 ` Mel Gorman
2005-11-02 5:11 ` Andrew Morton
2005-11-01 18:23 ` Rob Landley
2005-11-01 20:31 ` Joel Schopp
2005-11-01 20:59 ` Joel Schopp
2005-11-02 1:06 ` Nick Piggin
2005-11-02 1:41 ` Martin J. Bligh
2005-11-02 2:03 ` Nick Piggin
2005-11-02 2:24 ` Martin J. Bligh
2005-11-02 2:49 ` Nick Piggin
2005-11-02 4:39 ` Martin J. Bligh
2005-11-02 5:09 ` Nick Piggin
2005-11-02 5:14 ` Martin J. Bligh
2005-11-02 6:23 ` KAMEZAWA Hiroyuki
2005-11-02 10:15 ` Nick Piggin
2005-11-02 7:19 ` Yasunori Goto
2005-11-02 11:48 ` Mel Gorman
2005-11-02 11:41 ` Mel Gorman
2005-11-02 11:37 ` Mel Gorman
2005-11-02 15:11 ` Mel Gorman
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=20051106101832.510b0245.pj@sgi.com \
--to=pj@sgi.com \
--cc=akpm@osdl.org \
--cc=andy@thermo.lanl.gov \
--cc=arjan@infradead.org \
--cc=arjanv@infradead.org \
--cc=haveblue@us.ibm.com \
--cc=kravetz@us.ibm.com \
--cc=lhms-devel@lists.sourceforge.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mbligh@mbligh.org \
--cc=mel@csn.ul.ie \
--cc=mingo@elte.hu \
--cc=nickpiggin@yahoo.com.au \
--cc=torvalds@osdl.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
Powered by JetHome