This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH, gfortran testsuite]: Do not load denormals in gfortran.fortran_torture/intrinsic_nearest.f90

One of the many reasons that most traditional numerical analysts (including
me, though I am scarcely one) want nothing to do with dynamic modes is that
they cause havoc with optimisation.  For example, under some circumstances:

   PRINT *, 1.23*4.56
   CALL fred
   PRINT *, 1.23*4.56

will print two different values - according to IEEE 754, that is. According to Fortran, that's up to the implementor. 

Aargh!  Scratch that example - that's C99 I am describing.  I am trying to
juggle too many languages at once (I have courses on about 6) and am having
trouble keeping my specifications straight.  The Fortran example is:

   SUBROUTINE Fred
       < Various USEs >
       PRINT *, 1.23*4.56
   END SUBROUTINE Fred

   < Various USEs >
   CALL Fred
   < Set the mode >
   CALL Fred

Regards,
Nick Maclaren.



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]