Date: Wed, 7 Jul 1999 01:34:10 +0200
To: pgcc AT delorie DOT com
Subject: Re: -O2 versus -O1
Message-ID: <19990707013410.H25412@cerebro.laendle>
Mail-Followup-To: pgcc AT delorie DOT com
References: <19990705163430 DOT C3137 AT cerebro DOT laendle> <Pine DOT LNX DOT 4 DOT 10 DOT 9907061159420 DOT 21233-100000 AT hq DOT alert DOT sk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <Pine.LNX.4.10.9907061159420.21233-100000@hq.alert.sk>; from Niteshadow on Tue, Jul 06, 1999 at 12:11:45PM +0200
X-Operating-System: Linux version 2.2.10 (root AT cerebro) (gcc driver version egcs-2.91.66 19990314 (egcs-1.1.2 release) executing gcc version 2.7.2.3) 
From: Marc Lehmann <pcg AT goof DOT com>
Reply-To: pgcc AT delorie DOT com
X-Mailing-List: pgcc AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

On Tue, Jul 06, 1999 at 12:11:45PM +0200, Niteshadow wrote:
> I am walking on thin ice here.. I do not think, that specifying -O2 should
> introduce rounding errors

Rounding errors are a fundamental property of floating point operations
and are usually not introduced by the compiler. some optimization can
cause minor changes in the generated values (usually the results are more
exact), especially on the x86 which doesn't implement ieee in hardware.

If the resulting streams play correct I wouldn't care about it (you could
use -ffloat-store to get rid of them, at a high speed penalty). If they
don't then its most probably a compiler bug.

--  
      -----==-                                             |
      ----==-- _                                           |
      ---==---(_)__  __ ____  __       Marc Lehmann      +--
      --==---/ / _ \/ // /\ \/ /       pcg AT goof DOT com      |e|
      -=====/_/_//_/\_,_/ /_/\_\       XX11-RIPE         --+
    The choice of a GNU generation                       |
                                                         |