|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
Subject: Re: Interesting observations...
From: Russ Allbery (rra
stanford.edu)Date: Mon Aug 28 2000 - 17:17:16 CDT
- Next message: Jim Holthaus: "Re: "real" virtual (was :accepting mails for domains (aka relay))"
- Previous message: Deim Agoston: ""real" virtual (was :accepting mails for domains (aka relay))"
- Next in thread: Stephen_P_McInerney
sge.net: "Re: Interesting observations..."
- Maybe reply: Russ Allbery: "Re: Interesting observations..."
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Brad Knowles <blk
skynet.be> writes:
> At 4:26 PM +0100 2000/8/28, Lars Hecking wrote:
>> From the limited benchmarking with amavis-perl, I reckon that for
>> large files, perl allocates roughly 4x file size in memory before
>> doing anything else. Try a simple minimal perl program like
>>
>> use strict;
>> my
all = <STDIN>;
>> print
all;
>>
>> on a huge file, and check with top or proctool.
It shouldn't be quite *that* bad.
> If that's really going to allocate 4x500MB, I better find a
> bigger machine to run that on!
Perl is optimized for speed, not for memory consumption, and therefore can
eat memory pretty fast under some circumstances. Hashes are the worst
offenders; arrays shouldn't be too bad. Perl's allocating a lot of little
objects behind the scenes to keep track of things like string length,
numeric conversions, and so forth.
-- Russ Allbery (rrastanford.edu) <http://www.eyrie.org/~eagle/>
- Next message: Jim Holthaus: "Re: "real" virtual (was :accepting mails for domains (aka relay))"
- Previous message: Deim Agoston: ""real" virtual (was :accepting mails for domains (aka relay))"
- Next in thread: Stephen_P_McInerney
sge.net: "Re: Interesting observations..."
- Maybe reply: Russ Allbery: "Re: Interesting observations..."
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]