Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 08 Sep 2003 18:58:49 +0200
From:      "Poul-Henning Kamp" <phk@phk.freebsd.dk>
To:        David Schultz <das@FreeBSD.ORG>
Cc:        adrian@FreeBSD.ORG
Subject:   Re: Squid memory leaks in -stable using libc malloc 
Message-ID:  <10909.1063040329@critter.freebsd.dk>
In-Reply-To: Your message of "Sun, 07 Sep 2003 23:42:48 PDT." <20030908064248.GA7522@HAL9000.homeunix.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
In message <20030908064248.GA7522@HAL9000.homeunix.com>, David Schultz writes:
>
>Note also that there's a difference between heap size and physical
>memory usage.

I have not had much to add to this topic, but this one is so important
that I want to chime in.

You cannot infer _anything_ about the application from the heapsize
in a VM based system.

It is a totally irrelevant number, it has about the same information
contents as "the most recent block number allocated in a filesystem".

>A better
>metric for the memory pressure that an application is exerting on
>the system is its resident set size---'RES' in top(1).

Which I should add, tells nothing much unless the VM system is under
pressure.

>phkmalloc
>by default is not aggressive about returning memory to the system,
>but it usually exhibits very good behavior from the point of view
>of the VM system.

I can highly recommend the paper about phkmalloc in
src/share/doc/papers/malloc

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?10909.1063040329>