Aleksandar Donev wrote:
> Toon Moene wrote:
>
>>Extra integer registers to hold array offset addresses
>
> Let me ask another question: Is there a reason to keep a program in 32-bit
> mode if it can be ported easily to 64-bit mode (i.e. assume it is portable,
> as Richard's)?
Not that I know - see below ...
> My programs are usually FLOP or memory bandwidth bound and do
> not really need 64 bit integers (but all the reals are 64 bits).
Most 64-bit architectures use the I32LP64 model, meaning that integers
(both Fortran and C ints) are 32 bits and long and pointers are 64 bits.
> As I am
> reading, the reason behind the 32-bit mode is for backward compatibility and
> ease of migration, and the real thing to do is run in 64-bit mode once you
> port the code successfully.
That would be the way I would proceed (but for the fact that as a
compiler writer and tester, I much prefer 64-bit *big-endian*
architectures - but that's a whole different story ...)
Hope this helps,
--
Toon Moene - mailto:[log in to unmask] - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)
|