Date: Fri, 08 Apr 2005 15:16:19 +0800 From: David Xu <davidxu@freebsd.org> To: Francois Tigeot <ftigeot@wolfpond.org> Cc: ports@freebsd.org Subject: Re: FreeBSD Port: subversion-1.1.3 Message-ID: <42562FC3.8040704@freebsd.org> In-Reply-To: <20050408064122.GC45371@aoi.wolfpond.org> References: <425564D9.5050504@LogicX.us> <4255DC88.6060608@freebsd.org> <4255E04B.6020601@LogicX.us> <4255E43C.8010804@freebsd.org> <20050408064122.GC45371@aoi.wolfpond.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Francois Tigeot wrote:
>On Fri, Apr 08, 2005 at 09:54:04AM +0800, David Xu wrote:
>
>
>>I heard only -O works, -O2 does not work, -Os sometimes is a higher
>>optimization
>>level than -O2.
>>
>>
>
>AFAIK, -Os is only a subset of -O2.
>
>-Os only includes the -O2 optimisations which don't increase code size;
>it gives generally better performance than -O2 due to a better cache
>footprint.
>
>I have found this particularly true when using VIA C3 processors.
>
>
>
According to gcc manual:
-Os Optimize for size. This enables all -O2 optimizations that do
not typically increase code size. It also performs
further op-
timizations designed to reduce code size.
So the options are overlapped, and also each option has its own
optimizations, it should
be treated that both -Os and -O2 have dangerous optimizations.
David Xu
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42562FC3.8040704>