Message-ID:	<19990725134331.A9005@tabor.ta.jcu.cz>
Date:	Sun, 25 Jul 1999 13:43:31 +0200
From:	Jan Hubicka <hubicka AT ta DOT jcu DOT cz>
To:	djgpp-workers AT delorie DOT com
Subject:	Re: Stack alignment
References:	<Pine DOT SUN DOT 3 DOT 91 DOT 990725110926 DOT 27488G-100000 AT is>
Mime-Version:	1.0
X-Mailer:	Mutt 0.93i
In-Reply-To:	<Pine.SUN.3.91.990725110926.27488G-100000@is>; from Eli Zaretskii on Sun, Jul 25, 1999 at 11:09:59AM +0300
Reply-To:	djgpp-workers AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Sun, Jul 25, 1999 at 11:09:59AM +0300, Eli Zaretskii wrote:
> 
> Jeffrey Law wrote on the gcc-bugs mailing list that gcc 2.95 attempts
> to keep the stack aligned to a 128-bit boundary.
> 
> Does this mean that we need to bump up the stack alignment to 16-byte
> boundary as well?  It's currently 8-byte aligned by the startup code.
This is good idea. Gcc has choosed 16 byte alignment for simd type
of PentiumPro (II/III) that isn't too important yet, but it seems
to be wise to prepare for it.
You may control the alignment using -mpreferred-stack-boundary=x
x is log 2 base for alignment. 

Also about binutils alignment.. I believe alignment is set to prety low number
there.. next (after 2.95) version will need 32 byte alignment for AMD-K6 code
(to fit well into cache lines)

Honza

- Raw text -