SELECTED_REAL_KIND behavior when precision exceeded
Using the fortran compiler
f90: Sun Fortran 95 8.2 Patch 121020-05 2006/12/08
on an x86 computer running SunOS 5.10, the intrinsic function
SELECTED_REAL_KIND(24)
returns the value 16. Yet the compiler complains when I try to declare
REAL(16) :: x
^
"test4.f", Line = 2, Column = 12: ERROR: KIND value "16" is not supported for "REAL" type on x86 platforms.
This error message is clear enough, but I wonder about the SELECTED_REAL_KIND intrinsic. In "Fortran 95/2003 Explained" by Metcalf, Reid and Cohen, the description of the SELECTED_REAL_KIND intrinsic states (in part)
"If no corresponding kind value is available, the result is -1 if sufficient precision is unavailable, -2 if sufficient exponent range is unavailable, and -3 if both are unavailable."
How authoritative is this statement from Metcalf, Reid and Cohen?
