Re: getting same output as gfortran, long_float

[robin.vowels@gmail.com]

On Friday, May 1, 2015 at 5:45:23 PM UTC+10, Dmitry A. Kazakov wrote:
> On Fri, 01 May 2015 02:32:27 -0500, Nasser M. Abbasi wrote:
> 
> > Actually, it is much more complicated. The way I wrote it
> > above is not the optimal way. One is supposed to call
> > SELECTED_REAL_KIND(n,e) requesting n significant digits
> > and e number of digits in the exponent (e is optional).
> 
> Huh, they finally learned something after half of a century! (:-))
> 
> > i.e. one is supposed to write
> > 
> > SELECTED_REAL_KIND(8,3)  specifies real type of
> > (+-) 0.xxxxxxxx * 10 ^(+-)xxx
> > 
> > If the compiler does not support this, then the compile
> > will fail. This is in a way similar to Ada's
> > 
> >     type my_type is digits n;
> 
> Yes, though Ada also mandates that for the precision specified, the
> implementation must guarantee certain accuracy of operations. Maybe in the
> following 50 years FORTRAN will get that idea as well.
> 
> > And it is supposed to be portable way of doing things, vs.
> > using Real*16.
> 
> Actually REAL*16 is exactly portable.

No it's not.
Some compilers treat that is an error.
That form is non-standard.

SELECTED_REAL_KIND or a similar modern form is how
precision may be specified in a portable manner.