From: Daniel Phillips <phillips@bonn-fries.net>
To: Rik van Riel <riel@conectiva.com.br>, Marco Colombo <marco@esi.it>
Cc: <mike_phillips@urscorp.com>, <linux-kernel@vger.kernel.org>
Subject: Re: VM Requirement Document - v0.0
Date: Tue, 3 Jul 2001 20:29:27 +0200 [thread overview]
Message-ID: <01070320292709.00338@starship> (raw)
In-Reply-To: <Pine.LNX.4.33L.0107021538030.14332-100000@imladris.rielhome.conectiva>
In-Reply-To: <Pine.LNX.4.33L.0107021538030.14332-100000@imladris.rielhome.conectiva>
On Monday 02 July 2001 20:42, Rik van Riel wrote:
> On Thu, 28 Jun 2001, Marco Colombo wrote:
> > I'm not sure that, in general, recent pages with only one access are
> > still better eviction candidates compared to 8 hours old pages. Here
> > we need either another way to detect one-shot activity (like the one
> > performed by updatedb),
>
> Fully agreed, but there is one problem with this idea.
> Suppose you have a maximum of 20% of your RAM for these
> "one-shot" things, now how are you going to be able to
> page in an application with a working set of, say, 25%
> the size of RAM ?
Easy. What's the definition of working set? Those pages that are frequently
referenced. So as the application starts up some of its pages will get
promoted from used-once to used-often. (On the other hand, the target
behavior here conflicts with the goal of grouping together several
temporally-related accesses to the same page together as one access, so
there's a subtle distinction to be made here, see below.)
The point here is that there are such things as run-once program pages, just
as there are use-once file pages. Both should get low priority and be
evicted early, regardless of the fact they were just loaded.
> If you don't have any special measures, the pages from
> this "new" application will always be treated as one-shot
> pages and the process will never be able to be cached in
> memory completely...
The self-balancing way of doing this is to promote pages from the old end of
the used-once list to the used-often (active) list at a rate corresponding to
the fault-in rate so we get more aggressive promotion of referenced-often
pages during program loading, and conversely, aggressive demotion of
referenced-once pages.
--
Daniel
--
Daniel
next prev parent reply other threads:[~2001-07-03 18:26 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-06-28 12:20 VM Requirement Document - v0.0 mike_phillips
2001-06-28 12:30 ` Alan Cox
2001-06-28 13:33 ` Tobias Ringstrom
2001-06-28 13:37 ` Alan Cox
2001-06-28 14:04 ` Tobias Ringstrom
2001-06-28 14:14 ` Alan Cox
2001-06-28 14:52 ` Daniel Phillips
2001-06-28 14:39 ` Daniel Phillips
2001-06-28 18:01 ` Marco Colombo
2001-07-02 18:42 ` Rik van Riel
2001-07-03 10:33 ` Marco Colombo
2001-07-03 15:04 ` Daniel Phillips
2001-07-03 18:24 ` Daniel Phillips
2001-07-04 8:12 ` Ari Heitner
2001-07-04 9:41 ` Marco Colombo
2001-07-04 15:03 ` Daniel Phillips
2001-07-03 18:29 ` Daniel Phillips [this message]
2001-07-04 8:32 ` Marco Colombo
2001-07-04 14:44 ` Daniel Phillips
2001-06-28 15:21 ` Jonathan Morton
2001-06-28 16:02 ` Daniel Phillips
-- strict thread matches above, loose matches on Subject: below --
2001-07-05 15:09 mike_phillips
2001-07-05 15:04 Daniel Phillips
[not found] ` <fa.jprli0v.qlofoc@ifi.uio.no>
[not found] ` <fa.e66agbv.hn0u1v@ifi.uio.no>
2001-07-05 1:49 ` Dan Maas
2001-07-05 13:02 ` Daniel Phillips
2001-07-05 14:00 ` Xavier Bestel
2001-07-05 14:51 ` Daniel Phillips
2001-07-05 15:00 ` Xavier Bestel
2001-07-05 15:12 ` Daniel Phillips
2001-07-05 15:12 ` Alan Shutko
[not found] ` <002501c104f4/mnt/sendme701a8c0@morph>
2001-07-09 12:17 ` Pavel Machek
2001-07-12 23:46 ` Daniel Phillips
2001-07-13 21:07 ` Pavel Machek
2001-07-06 19:09 ` Rik van Riel
2001-07-06 21:57 ` Daniel Phillips
2001-07-04 16:08 mike_phillips
[not found] <fa.oqkojpv.3hosb7@ifi.uio.no>
[not found] ` <fa.jpsks3v.1o2gag4@ifi.uio.no>
2001-06-27 0:43 ` Dan Maas
2001-06-27 0:45 ` Mike Castle
2001-06-27 10:50 ` Xavier Bestel
2001-06-27 8:53 Martin Knoblauch
2001-06-27 18:13 ` Rik van Riel
2001-06-28 6:59 ` Martin Knoblauch
2001-06-28 11:27 ` Helge Hafting
2001-06-28 11:54 ` Martin Knoblauch
2001-06-28 12:02 ` Tobias Ringstrom
2001-06-28 12:31 ` Xavier Bestel
2001-06-28 13:05 ` Tobias Ringstrom
2001-06-26 19:58 Jason McMullan
2001-06-26 21:21 ` Rik van Riel
2001-06-26 21:29 ` Jason McMullan
2001-06-26 21:33 ` John Stoffel
2001-06-26 21:42 ` Rik van Riel
2001-06-26 22:21 ` Stefan Hoffmeister
2001-06-26 22:48 ` Jeffrey W. Baker
2001-06-27 0:18 ` Mike Castle
2001-06-28 13:07 ` John Fremlin
2001-06-27 13:36 ` Marco Colombo
2001-06-27 3:55 ` Daniel Phillips
2001-06-27 14:09 ` Pozsar Balazs
2001-06-28 22:47 ` John Fremlin
2001-06-30 15:37 ` Pavel Machek
2001-07-10 10:34 ` David Woodhouse
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=01070320292709.00338@starship \
--to=phillips@bonn-fries.net \
--cc=linux-kernel@vger.kernel.org \
--cc=marco@esi.it \
--cc=mike_phillips@urscorp.com \
--cc=riel@conectiva.com.br \
/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).