This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [patch, libgfortran] PR32446 F0.n output format fails with large numbers

From: Jerry DeLisle <jvdelisle at verizon dot net>
To: Tobias Burnus <burnus at net-b dot de>
Cc: Fortran List <fortran at gcc dot gnu dot org>, gcc-patches <gcc-patches at gcc dot gnu dot org>
Date: Sun, 24 Jun 2007 11:30:16 -0700
Subject: Re: [patch, libgfortran] PR32446 F0.n output format fails with large numbers
References: <467C8D6F.6090305@verizon.net> <467E75E8.6000107@net-b.de>

Tobias Burnus wrote:

Hi Jerry,

Jerry DeLisle wrote:

Regression tested on x86-64-pc-Gnu/Linux.
Test case included.
OK for trunk?

Ok.

Tobias

Post scriptum:

I find the following quite puzzling. For "print '(F0.1)',huge(1.0)",
gfortran prints:
  340282346638528859811704200000000000000.0
whereas ifort prints
  340282346638528859811704183484516925400.0
and NAG, Sun, g95 print:
  340282346638528859811704183484516925440.0
which matches the exquation/tree-dump:
huge = ((1-(1/2)^24)2^128) = 340282346638528859811704183484516925440

Reading any of these numbers with any of these compilers gives huge(r)
thus one should probably not worry much about it.

I noticed this too. It is sort of absurd to carry all those digits for a single precision number. Maybe the other compilers are using a double internally.

Thanks for review.

Jerry

References:
- [patch, libgfortran] PR32446 F0.n output format fails with large numbers
  - From: Jerry DeLisle
- Re: [patch, libgfortran] PR32446 F0.n output format fails with large numbers
  - From: Tobias Burnus

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]