On Thu, 18 May 2006, Tim Prince wrote:
> I'm certain the standard doesn't require the existence of an option like
> -r8.
> ...
> As 0.4 involves an inexact conversion between decimal and binary
> representation, some sort of constant propagation, in other than target
> platform precision, or extra precision evaluation, is needed to produce
> the result you appear to advocate.
I changed my code by adding the statement
real(8) :: a, b
and converting a=4.0 and b=0.4 into a=4.0_8 and b=0.4_8. Then
the printout of a and b gives
4.00000000000000000000 4.00000000000000000000
and still the comparison returns 'T'. How can this happen?
Roland
+----------------------------------------------------------------------+
| Roland Schilling Home-Office: +49(89)32929-670 |
| Max-Planck-Institut für Gravitationsphysik |
| -- Albert-Einstein-Institut -- |
| D-85748 Garching E-mail: [log in to unmask] |
| Germany http://www.geo600.uni-hannover.de |
+----------------------------------------------------------------------+
|