From owner-freebsd-questions Mon Nov 27 16:44:50 2000 Delivered-To: freebsd-questions@freebsd.org Received: from backup.af.speednet.com.au (af.speednet.com.au [202.135.188.244]) by hub.freebsd.org (Postfix) with ESMTP id 0E4CA37B479 for ; Mon, 27 Nov 2000 16:44:45 -0800 (PST) Received: from backup.af.speednet.com.au (backup.af.speednet.com.au [172.22.2.4]) by backup.af.speednet.com.au (8.11.1/8.11.1) with ESMTP id eAS0iKF17540; Tue, 28 Nov 2000 11:44:22 +1100 (EST) (envelope-from andyf@speednet.com.au) Date: Tue, 28 Nov 2000 11:44:20 +1100 (EST) From: Andy Farkas X-Sender: andyf@backup.af.speednet.com.au To: Kenneth Culver Cc: Alfred Perlstein , freebsd-questions@FreeBSD.ORG Subject: Re: compiler optimizations In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG I built and installed a 4.2-RELEASE world with the following: CFLAGS= -m486 -march=i486 -O2 -pipe COPTFLAGS= -m486 -march=i486 -O2 -pipe ...mainly as an exercise in pain (it took about 26 hours :-) Its a 486DX-66 with 32MB and has been working fine for the last 4 days. I doubt I can see any performance increase - the thing is gonna be slow no matter what I do to it :-) On Mon, 27 Nov 2000, Kenneth Culver wrote: > Alright, I knew how to do it but thanks :-) Anyway, I compiled with -Os > -march=pentiumpro and my kernel started randomly panicing on heavy disk > activity (once starting netscape, and once syncing disks) so I think I'll > just stick with the defaults. > > > ================================================================= > | Kenneth Culver | FreeBSD: The best NT upgrade. | > | Unix Systems Administrator | ICQ #: 24767726 | > | and student at The | AIM: muythaibxr | > | The University of Maryland, | Website: (Under Construction) | > | College Park. | http://www.wam.umd.edu/~culverk/| > ================================================================= > > On Mon, 27 Nov 2000, Alfred Perlstein wrote: > > > * Kenneth Culver [001127 13:47] wrote: > > > I'm just curious about something. I noticed the other day that the entire > > > linux kernel is compiled with options -O2 -m486. I was wondering if this > > > really causes any performance increase and was also wondering if there is > > > anywhere on the web that can show this. I am wondering because I was going > > > to use these options with FreeBSD (or -march=pentiumpro or -mpentiumpro, > > > and -Os instead of -O2). > > > > Yes, the optimization works, however because we still want to support > > 386s and because of bad experiances with the gcc optimizer in the > > past we don't really like to debug problems with kernels that have > > high levels of optimization because sometimes it's gcc's fault (or > > at least it could be in the past). > > > > If you want to compile the kernel and world like that edit /etc/make.conf > > and add your options to CFLAGS: > > > > CFLAGS= -O -pipe > > > > -- > > -Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org] > > "I have the heart of a child; I keep it in a jar on my desk." > > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-questions" in the body of the message > -- :{ andyf@speednet.com.au Andy Farkas System Administrator Speednet Communications http://www.speednet.com.au/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message