Date: Fri, 24 Feb 2006 14:55:59 -0800 From: Jason Evans <jasone@freebsd.org> To: Poul-Henning Kamp <phk@phk.freebsd.dk> Cc: Mike Jakubik <mikej@rogers.com>, freebsd-current@freebsd.org, Peter Fraser <petros.fraser@gmail.com>, Mark Kirkwood <markir@paradise.net.nz> Subject: Re: Virtual memory consumption (both user and kernel) in modern CURRENT Message-ID: <45B20008-CAC0-41FC-9F35-CEFE59611265@freebsd.org> In-Reply-To: <1743.1140821039@critter.freebsd.dk> References: <1743.1140821039@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Feb 24, 2006, at 2:43 PM, Poul-Henning Kamp wrote: > In message <06424916-2FC9-4A31-A109-654682BFC1C4@freebsd.org>, > Jason Evans writes: >> I did some quick experiments last night, where I booted a -current >> system with phkmalloc, then with jemalloc. The total increase in >> resident memory was negligible, on the order of kilobytes. > > Running which applications ? This was immediately after boot. I used ssh to log in, then ran top to get a snapshot at ~60 seconds after boot. Things like getty, zsh, sshd, and sendmail were running (total of 23 processes). Admittedly, this is a rather lame experimental design, but it was simple to perform, and I wanted to know if there were huge differences in system memory usage, as a number of people have repeatedly implied. This experiment convinced me that there are not any consistently substantial differences in memory usage, though I fully expect there to be variation, depending on the application(s) being run. Jason
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?45B20008-CAC0-41FC-9F35-CEFE59611265>