linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Richard Kennedy <richard@rsk.demon.co.uk>
To: Christoph Lameter <cl@linux-foundation.org>
Cc: penberg@cs.helsinki.fi, linux-mm <linux-mm@kvack.org>
Subject: Re: [PATCH][RFC] slub: increasing order reduces memory usage of some key caches
Date: Thu, 17 Jul 2008 11:09:08 +0100	[thread overview]
Message-ID: <1216289348.3061.16.camel@castor.localdomain> (raw)
In-Reply-To: <487E1ACF.3030603@linux-foundation.org>

On Wed, 2008-07-16 at 10:59 -0500, Christoph Lameter wrote:
> Patch to do this the right way in slub:
> 
> Index: linux-2.6/mm/slub.c
> ===================================================================
> --- linux-2.6.orig/mm/slub.c	2008-07-16 10:42:07.000000000 -0500
> +++ linux-2.6/mm/slub.c	2008-07-16 10:53:36.000000000 -0500
> @@ -1860,6 +1860,10 @@
>  
>  		rem = slab_size % size;
>  
> +		/* Never waste more than half of the size of an object*/
> +		if (rem > size / 2)
> +			continue;
> +
>  		if (rem <= slab_size / fract_leftover)
>  			break;

Thanks, I'll give that a try.

Do we need to limit the number of times this applies though?

for example, 216 byte structures will give

order:objs/slab:waste
0 :  18 :208
1 :  37 :200
2 :  75 :184
3 : 151 :152
4 : 303 : 88

I'm not sure where the balance point between efficient memory usage &
fragmentation pressure lies, but my gut feeling is that order 4 is just
too big for a structure this small.  

Richard

--
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:[~2008-07-17 10:09 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-16 12:29 Richard Kennedy
2008-07-16 13:21 ` Christoph Lameter
2008-07-16 13:58   ` Richard Kennedy
2008-07-16 14:03     ` Christoph Lameter
2008-07-16 14:30       ` Richard Kennedy
2008-07-16 14:33       ` Vegard Nossum
2008-07-16 14:46         ` Christoph Lameter
2008-07-18  9:57       ` Richard Kennedy
2008-07-18 14:17         ` Christoph Lameter
2008-07-18 14:20         ` Christoph Lameter
2008-07-18 14:42           ` Richard Kennedy
2008-07-18 14:55             ` Christoph Lameter
2008-07-18 15:22               ` Richard Kennedy
2008-07-18 17:13                 ` Christoph Lameter
2008-07-16 19:52   ` Mel Gorman
2008-07-17  9:48     ` Richard Kennedy
2008-07-16 15:52 ` Christoph Lameter
2008-07-16 15:59 ` Christoph Lameter
2008-07-17 10:09   ` Richard Kennedy [this message]
2008-07-17 16:56     ` Christoph Lameter
2008-07-18 10:17       ` Richard Kennedy
2008-07-18 14:21         ` Christoph Lameter

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=1216289348.3061.16.camel@castor.localdomain \
    --to=richard@rsk.demon.co.uk \
    --cc=cl@linux-foundation.org \
    --cc=linux-mm@kvack.org \
    --cc=penberg@cs.helsinki.fi \
    /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