X-pop3-spooler:	POP3MAIL 2.1.0 b 3 961213 -bs-
Delivered-To:	pcg AT goof DOT com
Message-ID:	<19980314022303.40050@cerebro.laendle>
Date:	Sat, 14 Mar 1998 02:23:03 +0100
From:	Marc Lehmann <pcg AT goof DOT com>
To:	beastium <beastium-list AT Desk DOT nl>
Subject:	Re: paranoia & extra precision [was -fno-float-store in pgcc]
References:	<19980312234235 DOT 53559 AT cerebro DOT laendle> <Pine DOT SOL DOT 3 DOT 96 DOT 980313134433 DOT 249B-100000 AT stekt10>
Mime-Version:	1.0
X-Mailer:	Mutt 0.88
In-Reply-To:	<Pine.SOL.3.96.980313134433.249B-100000@stekt10>; from Tuukka Toivonen on Fri, Mar 13, 1998 at 01:59:01PM +0200
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
Lines:	44

On Fri, Mar 13, 1998 at 01:59:01PM +0200, Tuukka Toivonen wrote:
> On Thu, 12 Mar 1998, Marc Lehmann wrote:
> 
> >yeah, it's even worse.. so you have sth. between 53 and 80 bits of precision.
> 
> - If you want extended precision -> store values in memory with 10 bytes.
>   You'll get always 80 bits precision.
> - If you want to use `double' format with some extra precision,
>   just set high precision on and use doubles.
> - If you want IEEE-compliance, set FPU to low precision and everything is
>   right again.

but imho, the very very large speed penalty of this approach renders
it useless. the fact that the compiler can workaround these problems
and "make the hardware iee compliant" doesn't make it iee compliant.

my c64 is as ieee compliant as the intel chips, then (it could also
do ieee calculations)

> The hardware makes all of this possible. Don't blame FPU if C compiler can
> not do it (again, I don't say it couldn't; I don't know).

it can, but nobody wants to slowdown their programs more than necessary.
gcc already has to workaround some problems, where just ignoring
the current mode creates much bigger errors than the double <> long double
issue.

> >pragma = nono ;) until c9x adds the _Pragma syntax, #pragma's are just
> >useless.
> 
> I think that I read somewhere why #pragma's shouldn't be use. I think it
> would be nice way to 'hint' the compiler... I don't understand what's the
> problem with them? Everyone except GCC uses them. (Well... at least
> Borland/Watcom :)

sure, watcom also supports "inline" asm... of course, not in macros ;)

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

- Raw text -