linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Christoph Lameter <clameter@engr.sgi.com>
To: Rahul Iyer <rni@andrew.cmu.edu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>,
	linux-mm@kvack.org, Rik van Riel <riel@redhat.com>,
	Marcelo Tosatti <marcelo.tosatti@cyclades.com>
Subject: Re: Zoned CART
Date: Tue, 16 Aug 2005 13:49:34 -0700 (PDT)	[thread overview]
Message-ID: <Pine.LNX.4.62.0508161318420.7906@schroedinger.engr.sgi.com> (raw)
In-Reply-To: <43024435.90503@andrew.cmu.edu>

Hmm. I am a bit concerned about the proliferation of counters in CART 
because these may lead to bouncing cachelines.

The paper mentions some relationships between the different values. 

If we had a counter for the number of pages resident (nr_rpages) 
(|T1|+|T2|) then that counter would gradually approach c and then no 
longer change.

Then

|T2| = nr_rpages - |T1|

Similarly if we had a counter for the number of pages on the evicted 
list (nr_evicted) then that counter would also gradually approach c and 
then stay constant. nr_evicted would only increase if nr_rpages has 
already reached c which is another good thing to avoid bouncing 
cachelines.

Then also

|B2| = nr_evicted - |B1|

Thus we could reduce the frequency of counter increments on a fully 
loaded system (where nr_rpages = c and nr_eviced = c) by 
calculating some variables:

#define nr_inactive (nr_rpages - nr_active)
#define nr_evicted_longterm (nr_evicted - nr_evicted_shortterm)

There is also a relationship between |S| and |L| since these attributes 
are only used on resident pages.

|L| = nr_rpages - |S|

So

#define nr_longterm (nr_rpages - nr_shortterm)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2005-08-16 20:49 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-08-12 14:37 Peter Zijlstra
2005-08-12 15:42 ` Rahul Iyer
2005-08-12 15:52   ` Peter Zijlstra
2005-08-12 23:08   ` Marcelo Tosatti
2005-08-13 19:00     ` Rahul Iyer
2005-08-13 19:08       ` Marcelo Tosatti
2005-08-13 21:30         ` Rik van Riel
2005-08-14 18:31     ` Peter Zijlstra
2005-08-12 20:21 ` Marcelo Tosatti
2005-08-12 22:28   ` Marcelo Tosatti
2005-08-13 19:03   ` Rahul Iyer
2005-08-14 12:58 ` Peter Zijlstra
2005-08-15 21:31   ` Peter Zijlstra
2005-08-16 19:53     ` Rahul Iyer
2005-08-16 20:49       ` Christoph Lameter [this message]
2005-08-25 22:39         ` Peter Zijlstra
2005-08-26  0:01           ` Christoph Lameter
2005-08-26  3:59             ` Rahul Iyer
2005-08-26  7:09             ` Peter Zijlstra
2005-08-26 12:24               ` Rik van Riel
2005-08-26 21:03   ` Peter Zijlstra
2005-08-27 19:46     ` [RFC][PATCH] " Peter Zijlstra

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=Pine.LNX.4.62.0508161318420.7906@schroedinger.engr.sgi.com \
    --to=clameter@engr.sgi.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=linux-mm@kvack.org \
    --cc=marcelo.tosatti@cyclades.com \
    --cc=riel@redhat.com \
    --cc=rni@andrew.cmu.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