Xref:	news2.mv.net comp.lang.c:54819 comp.lang.c++:70367 comp.os.msdos.djgpp:740 gnu.g++.help:4477 gnu.gcc.help:5678
From:	axl AT zedat DOT fu-berlin DOT de (Axel Thimm)
Newsgroups:	comp.lang.c,comp.lang.c++,gnu.gcc.help,gnu.g++.help,comp.os.msdos.djgpp
Subject:	Re: float != float and floats as return types
Date:	Fri, 02 Feb 1996 03:31:08 GMT
Organization:	FU Berlin
Lines:	26
Message-ID:	<4es0ep$is2@fu-berlin.de>
References:	<DM19oy DOT KsJ AT granite DOT mv DOT net>
Reply-To:	axl AT zedat DOT fu-berlin DOT de
NNTP-Posting-Host:	axl.dialup.fu-berlin.de (160.45.218.93)
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> wrote:
[...]
>> -ffloat-store

>Just to make the picture complete: this switch should only be used in 
>programs which break without it, because holding intermediate results in 
>the most precise representation makes FP computations more accurate.  
[...]
>Another case of justified use of that switch is a program which explores 
>the intricacies of the floating-point representation of a machine.

The reason I "expored" this behaviour was an optimization by hand of a
simulation program. A lattices site value is very often used in
quadratic terms, so introducing a second lattice and storing theese
values seemed like just speeding up the process. I was amazed when the
output changed (the results were still within the stat. error).
So if someone is writing programs that can do exactly the same thing
with two different implementations, then he would very much like to turn
such optimizations temporarily off, to get the same results. Later he
would turn them on again.

===
Axel Thimm <axl AT zedat DOT fu-berlin DOT de>
Freie Universitaet Berlin
===

- Raw text -