Re: Portable isfinite for gfortran 4.0.3 and ifort 8.1/9.0

In article <1152959999.935400.224490@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"Niels L Ellegaard" <niels.ellegaard@xxxxxxxxx> writes:
I am looking for an implementation of isfinite that works on the
following three compilers

gfortran 4.0.3, ifort 8.1, and ifort 9.0

You probably want to upgrade to at least gfortran 4.1.1.

I do not use the -fastmath flag when compiling, but I do use the
optimization flags. what is the simplest failproof way to do the test?

See Joost's post for a function that may work. Some
compilers may optimize 'isfinite = (a-a).EQ.0' to TRUE for
all values of 'a'.

--
Steve
http://troutmask.apl.washington.edu/~kargl/
.

Relevant Pages

  • Re: Symbols in function return specifiers
    ... END FUNCTION test ... I tested the example with 3 compilers (g95, ... g95 and gfortran compile the two cases without any warning, ... ifort compiles only when the type of the function is declared ...
    (comp.lang.fortran)
  • Re: Zero-size arrays
    ... were also found in ifort, gfortran, pathscale and some more which I do ... and commercial compilers in this regard; ... Fortran 90/95, but with configure check also things like Fortran ...
    (comp.lang.fortran)
  • Re: Zero-size arrays
    ... were also found in ifort, gfortran, pathscale and some more which I do ... and commercial compilers in this regard; ... Fortran 90/95, but with configure check also things like Fortran ...
    (comp.lang.fortran)
  • Portable isfinite for gfortran 4.0.3 and ifort 8.1/9.0
    ... following three compilers ... gfortran 4.0.3, ifort 8.1, and ifort 9.0 ... optimization flags. ...
    (comp.lang.fortran)
  • Re: Summary of F2003 features supported in each compiler
    ... And even for a given vendor, ... both: ifort 9.1.040 ... Only move_alloc: gfortran 4.2/4.3 ... 'complete' F2003 compilers? ...
    (comp.lang.fortran)