X-pop3-spooler: POP3MAIL 2.1.0 b 3 961213 -bs-
Delivered-To: pcg AT goof DOT com
Message-ID: <19980318174324.07833@cerebro.laendle>
Date: Wed, 18 Mar 1998 17:43:24 +0100
From: Marc Lehmann <pcg AT goof DOT com>
To: Thomas Koehler <T DOT Koehler AT pfh DOT research DOT philips DOT com>
Cc: beastium <beastium-list AT Desk DOT nl>
Subject: Re: paranoia & extra precision
References: <dhl9JCA2SHD1EwEN AT foobar DOT co DOT uk> <350E2A3F DOT 328BCA24 AT pfh DOT research DOT philips DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.88
In-Reply-To: <350E2A3F.328BCA24@pfh.research.philips.com>; from Thomas Koehler on Tue, Mar 17, 1998 at 08:46:07AM +0100
X-Operating-System: Linux version 2.1.85 (root AT cerebro) (gcc version pgcc-2.91.06 980129 (gcc-2.8.0 release)) 
Status: RO
Content-Length: 1315
Lines: 27

On Tue, Mar 17, 1998 at 08:46:07AM +0100, Thomas Koehler wrote:
> Paul Shirley wrote:
> 
> > even when optimising. Is a C compiler really allowed to make this change
> > without explicit permission? (ie not just a general -On flag)
> 
> I agree, why else is there an explicit compiler switch 
>  -ffast-math
>               This option allows GCC to violate some ANSI or IEEE
>               rules/specifications  in the interest of optimizing
>               code for speed. 

Correct. But it doesn't say gcc adheres to IEEE fp. On many platforms, this
is impossible (for gcc, e.g. because you can choose which fpu emulation you
want or the platform standard simply isn't ieee)). Switching (strict) IEEE
on by default doesn't make sense:

Find a (useful!) program that doesn't work with the current gcc semantics
and explain why sacrificing speed for the rest should be the default.

      -----==-                                              |
      ----==-- _                                            |
      ---==---(_)__  __ ____  __       Marc Lehmann       +--
      --==---/ / _ \/ // /\ \/ /       pcg AT goof DOT com       |e|
      -=====/_/_//_/\_,_/ /_/\_\                          --+
    The choice of a GNU generation                        |
                                                          |