This is the mail archive of the gcc-patches@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: [RFC PATCH]: Reciprocal sqrt (rsqrt) conversion pass

From: Michael Matz <matz at suse dot de>
To: Uros Bizjak <ubizjak at gmail dot com>
Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Richard Guenther <rguenther at suse dot de>, Andrew Pinski <pinskia at gmail dot com>
Date: Thu, 14 Jun 2007 16:45:57 +0200 (CEST)
Subject: Re: [RFC PATCH]: Reciprocal sqrt (rsqrt) conversion pass
References: <5787cf470706130335k103552c8hcd58931a87b07328@mail.gmail.com> <Pine.LNX.4.64.0706141604540.8724@wotan.suse.de> <5787cf470706140713w5036730eq3997d860b987a34c@mail.gmail.com>

Hi,

On Thu, 14 Jun 2007, Uros Bizjak wrote:

> > Um, no.  At least the SSE rsqrt instructions have only results 
> > accurate to 12bits, they are not meant to be used by themself.  I 
> > would be against such transformation _even_ with unsafe-math-optims.
> 
> This is referenced to an early RFC patch. Final version has NR 
> enhancement step and works OK with adequate precision.

Yeah, saw those mails too late, objection withdrawn.


Ciao,
Michael.

References:
- [RFC PATCH]: Reciprocal sqrt (rsqrt) conversion pass
  - From: Uros Bizjak
- Re: [RFC PATCH]: Reciprocal sqrt (rsqrt) conversion pass
  - From: Michael Matz
- Re: [RFC PATCH]: Reciprocal sqrt (rsqrt) conversion pass
  - From: Uros Bizjak

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