[gfortran] I/O of large integer and real kinds

[Richard Henderson]

On Tue, Jun 07, 2005 at 09:38:32PM +0200, FX Coudert wrote:
> What are C types for real(10) and real(16) on ia64-hpux?

__float80 and __float128.  This applies to all ia64 and i386.
One of them will map to long double, but it's os dependant.

>   real(16) :: y
>         1
> Error: Kind 16 not supported for type REAL at (1)

Odd.

	__float80 x;
	__float128 y;

	int main ()
	{
	  x += 1;
	  y += 1;
	}

compiles, but doesn't link.  It looks like this stuff is incomplete.  :-/


r~