memory size reported by ps

[ljm]

Hi,

Looking at memory usage by processes, I used a ps -efo "args,vsz". I would
expect to get the size of the process is real memory+pagingspace.
If I do a ps -lef I get a different memory size in the column "SZ". The
value in SZ is consistently ten times bigger than the vsz-value. That seems
starnge to me: I would expect the core-size to be smaller than the virtual
memory size.

Could anyone explain this please?

ljm


--
(c) ljm (bij xs4all) No part of this copyright message may be
reproduced, read or seen, dead or alive or by any means, including
but not limited to telepathy without the benevolence of the author.

[Nicholas Dronen]

ljm <me@here.there> wrote:
l> Hi,

l> Looking at memory usage by processes, I used a ps -efo "args,vsz". I would
l> expect to get the size of the process is real memory+pagingspace.
l> If I do a ps -lef I get a different memory size in the column "SZ". The
l> value in SZ is consistently ten times bigger than the vsz-value. That seems
l> starnge to me: I would expect the core-size to be smaller than the virtual
l> memory size.

l> Could anyone explain this please?

Using this

[url]http://www.opengroup.org/onlinepubs/007908799/xcu/ps.html[/url]

and this

[url]http://groups.google.com/groups?q=aix+sz+ps&selm=3onant%24kk6%40inet-nntp-gw-1.us.oracle.com&rnum=7[/url]

one might speculate that SZ is larger than vsz because SZ includes
"other process segments (such as text and shared memory). . . ."
That's speculation, and loose speculation at that. Someone who
can access the ps source can provide an authoritative answer.

Regards,

Nicholas

--
"Why shouldn't I top-post?" [url]http://www.aglami.com/tpfaq.html[/url]
"Meanings are another story." [url]http://www.ifas.org/wa/glossolalia.html[/url]