This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: sqrt via SSE2 registers

From: Jan Hubicka <jh at suse dot cz>
To: Tim Prince <tprince at computer dot org>
Cc: Jan Hubicka <jh at suse dot cz>, Tom Womack <tom at womack dot net>, gcc at gcc dot gnu dot org
Date: Tue, 19 Feb 2002 18:08:09 +0100
Subject: Re: sqrt via SSE2 registers
References: <001b01c1b6ff$f11fe6c0$5637f380@maths.nottingham.ac.uk> <20020217170516.GB14522@atrey.karlin.mff.cuni.cz> <E16dANH-0002P3-00@granger.mail.mindspring.net>

> >         pushl   %ebp
> >         movl    %esp, %ebp
> >         subl    $8, %esp
> >         sqrtsd  a, %xmm0
> >         andl    $-16, %esp
> >         movsd   %xmm0, a
> >         movl    %ebp, %esp
> >         popl    %ebp
> >         ret
> > .Lfe1:
> >         .size   main,.Lfe1-main
> >         .comm   a,8,8
> >         .ident  "GCC: (GNU) 3.1 20020217 (experimental)"
> > u-pl5:/tmp/egcs/build/gcc$
> So it seems we must remove the sqrt() code from 
> /usr/include/bits/mathinline.h or find some way to skip it when -mfpmath is 
> issued.

I believe it has been already removed in newer glibc releases.
Also gcc defines macro when SSE is available for such situations.

Honza

References:
- sqrt via SSE2 registers
  - From: Tom Womack
- Re: sqrt via SSE2 registers
  - From: Jan Hubicka
- Re: sqrt via SSE2 registers
  - From: Tim Prince

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]