Best way to declare a double precision in Fortran?

Help
2

The following is more or less, my go-to module that I use it everywhere:

module mod_types
    use iso_fortran_env, only:  int8, int16, int32, int64, real32, real64, &
                                input_unit, output_unit, error_unit
    implicit none
    integer, parameter :: sp        = real32
    integer, parameter :: dp        = real64
    integer, parameter :: stdin     = input_unit
    integer, parameter :: stdout    = output_unit
    integer, parameter :: stderr    = error_unit
end module

An example use is:

module foo
use mod_types, only: wp=>dp 
implicit none
real(wp)::bar
end module

wp stands for working precision and if for some reason I want to change precision, I just change this wp=>sp so I do not need to change manually all real variables. I believe something in that line is considered the “standard way” now.

7 Likes