Re: check if is NaN

Re: check if is NaN

[Thomas.Koenig]

(Cc: to gcc-bugs@gcc.gnu.org)
In article <8mr7bo$bf1$1@nnrp1.deja.com>,  <ezio_riva@my-deja.com> wrote:

>a) I 'm using g77 but I need to use the option -ffast-math.

I was quite surprised that

      external isnand
      logical isnand
      double precision a
      a = acos(1.2)
      print *,a,isnand(a)
      end
      logical function isnand(a)   
      double precision a
      isnand = a .ne. a
      end

produced

  NAN F

with GNU Fortran 0.5.25 19991030 (prerelease) on Linux i386 and
compilation with -O4 -ffast-math.  Inspection of the generated assembly
code showed

.globl isnand_
	.type	 isnand_,@function
isnand_:
	pushl %ebp
	movl %esp,%ebp
	xorl %eax,%eax
	leave
	ret

which shows that the test was optimized away.

I'm not sure wether to classify this as a bug or not.  The documentation
warns that -ffast-math makes programs non IEEE-conforming, but still...
What I would really like, I suppose, would be an ISNAN intrinsic
function.

BTW,

      isnand = a.ge. 0. .or. a .ne. 0.

works, but a very clever optimizer could also optimize this away...
>From aoliva@redhat.com Wed Aug 09 09:01:00 2000
From: Alexandre Oliva <aoliva@redhat.com>
To: "Tim Prince" <tprince@computer.org>
Cc: "Markus Werle" <markus@lufmech.rwth-aachen.de>, <gcc-bugs@gcc.gnu.org>
Subject: Re: Funny warning message: greater than minimum
Date: Wed, 09 Aug 2000 09:01:00 -0000
Message-id: <orpunil96f.fsf@guarana.lsd.ic.unicamp.br>
References: <399160F4.86730B9B@lufmech.rwth-aachen.de> <000901c0020d$d18c7f10$0100000a@TIMYX18EWDT6RQ>
X-SW-Source: 2000-08/msg00186.html
Content-length: 431

On Aug  9, 2000, "Tim Prince" <tprince@computer.org> wrote:

> Perhaps this warning belongs under -pedwarn or whatever.

Indeed, it does :-)

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                  aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist    *Please* write to mailing lists, not to me

Re: check if is NaN

[Toon Moene]

Thomas.Koenig@cologne.de wrote:

> >a) I 'm using g77 but I need to use the option -ffast-math.
> 
> I was quite surprised that
> 
>       external isnand
>       logical isnand
>       double precision a
>       a = acos(1.2)
>       print *,a,isnand(a)
>       end
>       logical function isnand(a)
>       double precision a
>       isnand = a .ne. a
>       end
> 
> produced
> 
>   NAN F
> 
> with GNU Fortran 0.5.25 19991030 (prerelease) on Linux i386 and
> compilation with -O4 -ffast-math.  Inspection of the generated assembly
> code showed
> 
> .globl isnand_
>         .type    isnand_,@function
> isnand_:
>         pushl %ebp
>         movl %esp,%ebp
>         xorl %eax,%eax
>         leave
>         ret
> 
> which shows that the test was optimized away.
> 
> I'm not sure wether to classify this as a bug or not.  The documentation
> warns that -ffast-math makes programs non IEEE-conforming, but still...

Exactly  - moreover, I have warned about this behaviour in the past on
comp.lang.fortran.

I'm sorry - the whole purpose of -ffast-math is to cut corners,
especially where following the IEEE standard zealously would incur
significant overhead.

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://g95.sourceforge.net/ (under construction)
>From mschwendt@web.de Wed Aug 09 14:57:00 2000
From: Michael Schwendt <mschwendt@web.de>
To: gcc-bugs@gcc.gnu.org
Subject: Re: Member initialization of arrays of constants
Date: Wed, 09 Aug 2000 14:57:00 -0000
Message-id: <Pine.LNX.4.21.0008092343210.1338-100000@faldor.intranet>
References: <Pine.LNX.4.21.0007240337290.955-100000@faldor.intranet>
X-SW-Source: 2000-08/msg00195.html
Content-length: 825

Hi!

After a discussion on comp.std.c++ and quite some time spent on
reading the Standard and Bjarne Stroustrup's book, I feel sort of
brain-washed and tend to believe that this code must not compile
without errors because struct A contains uninitialized const members
that either require a constructor or cannot be initialized
explicitly:

struct A
{
	const char c[4];  // ill-formed
	const int i;      // requires constructor
};

class C
{
    static const A a;
};
	
const A C::a = { 'T','E','S','T', 2 };  // error! struct A is bad


In Stroustrup's "C++ Programming Language" 3rd Edition, see:

  10.4.2
  10.4.6.1
  10.4.6.2

In the C++ Standard see:

  8.5/9
  9/4
  9.2/4
  9.2/5

Conclusively,

Regards,
Mike

-- 
"I don't need luck. I got a system. I play all the odd red numbers."
- Danny Wilde (The Persuaders!)