Date: Thu, 24 Feb 2000 21:30:05 -0500 (EST) From: Donn Miller <dmmiller@cvzoom.net> To: David O'Brien <obrien@freebsd.org> Cc: User URANIA <urania@iss.com.tw>, freebsd-current@freebsd.org Subject: Re: Problem with IBM Netfinity 5000 Server Message-ID: <Pine.BSF.4.20.0002242126410.50965-100000@lcm208.cvzoom.net> In-Reply-To: <20000224145121.C90075@dragon.nuxi.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 24 Feb 2000, David O'Brien wrote: > On Wed, Feb 23, 2000 at 02:27:31PM +0800, User URANIA wrote: > > I also set these flags at my /etc/make.conf > > CFLAGS= -mpentiumpro -O6 -pipe -funroll-loops -fexpensive-optimizations > > COPTFLAGS= -mpentiumpro -O6 -pipe -funroll-loops > > -O6 (any -O above 3) is nonsense with the base GCC compiler. > I really have to wonder what people think they are getting by using such > hight -O values. I looked at the code for gcc a while back, and it looks like optimization levels of 3 and above are considered 3. So, -O6 is the same as -O3. I think gcc a long time ago had flags that went this high. I think pgcc uses -O6, though. Sometimes when using flags -mpentium -O3 -pipe, I'll get an "internal compiler error" unless I bump down the opt to -O2 or lower. This is good -- you want to have a compiler error saying the compiler can't do it, rather than having it creating the code anyway with bugs. - Donn To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.20.0002242126410.50965-100000>