From owner-freebsd-sparc64@FreeBSD.ORG Fri Sep 8 17:11:55 2006 Return-Path: X-Original-To: sparc64@freebsd.org Delivered-To: freebsd-sparc64@FreeBSD.ORG Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 77A0816A4DA for ; Fri, 8 Sep 2006 17:11:55 +0000 (UTC) (envelope-from shannon@widomaker.com) Received: from wilma.widomaker.com (wilma.widomaker.com [204.17.220.5]) by mx1.FreeBSD.org (Postfix) with ESMTP id CC5BF43D62 for ; Fri, 8 Sep 2006 17:11:51 +0000 (GMT) (envelope-from shannon@widomaker.com) Received: from [69.72.99.166] (helo=escape.goid.lan) by wilma.widomaker.com with esmtp (Exim 3.36 #1) id 1GLjtN-0000TW-00 for sparc64@freebsd.org; Fri, 08 Sep 2006 13:11:49 -0400 Received: from daydream.goid.lan (IDENT:0@daydream.goid.lan [192.168.1.10]) by escape.goid.lan (8.11.6p3/8.11.6) with ESMTP id k88H9gV04929 for ; Fri, 8 Sep 2006 13:09:42 -0400 (EDT) Received: from daydream.goid.lan (localhost [127.0.0.1]) by daydream.goid.lan (8.13.6/8.13.4) with ESMTP id k88H9gbQ030425 for ; Fri, 8 Sep 2006 13:09:42 -0400 Received: (from shannon@localhost) by daydream.goid.lan (8.13.6/8.13.4/Submit) id k88H9gTA030424 for sparc64@freebsd.org; Fri, 8 Sep 2006 13:09:42 -0400 Date: Fri, 8 Sep 2006 13:09:41 -0400 From: Charles Shannon Hendrix To: sparc64@freebsd.org Message-ID: <20060908170941.GC29348@widomaker.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.1i Cc: Subject: memory usage X-BeenThere: freebsd-sparc64@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting FreeBSD to the Sparc List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 08 Sep 2006 17:11:55 -0000 I have a question about memory usage on FreeBSD-sparc64. I have a process table which looks like this on a FreeBSD 6.1 server, a Sun Ultra 2/300: PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 52906 www 1 4 0 83888K 9792K accept 1 0:00 0.00% httpd 52907 www 1 4 0 83888K 9792K accept 1 0:00 0.00% httpd 52905 root 1 96 0 83840K 9752K select 0 0:01 0.00% httpd 52964 mysql 1 96 0 56424K 29104K select 0 0:02 0.00% mysqld 52849 shannon 1 96 0 29864K 3944K select 0 0:00 0.00% sshd 52814 shannon 1 96 0 29864K 3944K select 1 0:00 0.00% sshd 52847 root 1 4 0 29864K 3872K sbwait 1 0:00 0.00% sshd 52812 root 1 4 0 29864K 3872K sbwait 0 0:00 0.00% sshd 4317 root 1 96 0 21768K 2920K select 0 0:03 0.00% svnserve 35260 root 1 96 0 18856K 3016K select 0 0:16 0.00% sshd 8830 root 1 96 0 18520K 7808K select 0 5:00 0.00% perl5.8.8 24861 root 1 8 0 17472K 1728K wait 0 0:00 0.00% login 52860 root 1 8 0 17440K 1656K wait 1 0:00 0.00% su 52850 shannon 1 20 0 10144K 3616K pause 1 0:00 0.00% tcsh 52815 shannon 1 20 0 10144K 3608K pause 1 0:00 0.00% tcsh 52861 root 1 5 0 10016K 3464K ttyin 0 0:00 0.00% csh 25425 shannon 1 5 0 9744K 3536K ttyin 1 0:01 0.00% tcsh To me those numbers are a bit high, but this is my first time running FreeBSD instead of Solaris on 64-bit SPARC, so I didn't know what to expect. Memory usage for roughly the same processes wasn't this high under Solaris 10, in some cases nowhere near as high. The machine runs great, I just wanted to know if was normal or not. The apache build is 1.x with PHP and about 5 other modules, nothing fancy. Still, 30MB for sshd seems pretty high even though that is just the virtual size. Is it something about the sparc64 libraries that pushes the virtual sizes much higher than with 32-bit sparc? Could I make any significant savings by doing some critial builds with a newer version of gcc? I'm not worried right now, mostly just curious. -- shannon "AT" widomaker.com -- ["People should have access to the data which you have about them. There should be a process for them to challenge any inaccuracies." -- Arthur Miller]