F90-related question

[Toon Moene]

On 12/11/2013 01:54 AM, Steve Kargl wrote:

< excellent explanation >

You are absolutely right that the option had the high potential of 
making programmers lazy.

The right approach to the "Cray-ism" problem is the one that I took in 
1992: Try if the code works on machines where REAL is a 32-bit floating 
point number and repair the places where it doesn't (in fact, our whole 
30,000+ line weather model at that time could run in 32-bit mode *in 
spite of the fact* that it had only been developed on Crays and Convexes 
with their Fortran compiler operating in "Cray mode").

> PS: Janus is looking for someone to write a section for the manual
> that details the differences and problems with these options.  You're
> more than welcome to use the code snippet that I've posted.
>
I will look into that (during the Christmas recess).  I think it would 
help if some historical context made clear why these options are as they 
are and what "sensible" use of them means.

Thanks !

-- 
Toon Moene - e-mail: toon@moene.org - phone: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
At home: http://moene.org/~toon/; weather: http://moene.org/~hirlam/
Progress of GNU Fortran: http://gcc.gnu.org/wiki/GFortran#news