How does the Fortran90 function floor not have a main type?

[Simfish]

/home/disk/p/atms380/xx/October-Runs/timeManMod/SourceMods/time_manager.F90(664):
error #6404: This name does not have a type, and must have an explicit
type. [FLOOR]
tmd = day_earth/PLANET_DAY_RATIO - floor(day_earth/PLANET_DAY_RATIO)

I'm running Fortran 90 with the ifort compiler. As far as I can tell, floor is a function introduced in Fortran 90

Using Fortran compiler: ifort -O -I/home/disk/eos11/bitz/cam3.1/cam1/models/utils/esmf -I/home/disk/eos11/bitz/cam3.1/cam1/models/utils/esmf/src/include -I/home/disk/eos11/bitz/cam3.1/cam1/models/utils/esmf/build/linux_intel -I/home/disk/eos11/bitz/cam3.1/cam1/models/utils/esmf/include -I/home/disk/eos11/bitz/cam3.1/cam1/models/utils/esmf/src/Infrastructure/mpiuni
Fortran Compiler version:
Intel(R) Fortran Intel(R) 64 Compiler Professional for applications running on Intel(R) 64, Version 11.1 Build 20091130 Package ID: l_cprof_p_11.1.064

 

[Mark44]

The compiler seems to be complaining that the PLANET_DAY_RATIO thing in the call to the floor intrinsic doesn't have a type. A workaround might be to define it as a parameter.

 

[uart]

That's strange, it should recognize "floor()"

Can you test if int() and nint() work?

 

[Simfish]

I see - thanks very much for the reply!

Here is the subroutine. I defined PLANET_DAY_RATIO within the subroutine. Am I doing it wrong?

subroutine get_curr_date(yr, mon, day, tod, offset)

! Return date components valid at end of current timestep with an optional
! offset (positive or negative) in seconds.

   implicit none
   
! Arguments
   integer, intent(out) ::&
      yr,    &! year
      mon,   &! month
      day,   &! day of month
      tod     ! time of day (seconds past 0Z)

   integer, optional, intent(in) :: offset  ! Offset from current time in seconds.
                                            ! Positive for future times, negative 
                                            ! for previous times.

! Local variables
   character(len=*), parameter :: sub = 'get_curr_date'
   integer :: rc
   type(esmf_date) :: date
   type(esmf_time) :: off
   integer :: ymd
   integer :: leap_days
   integer :: yZero
   integer :: day_earth
   float :: PLANET_DAY_RATIO

(stuff)

   yr = ymd/10000
   mon = mod(ymd, 10000) / 100
   day = mod(ymd, 100)
	PLANET_DAY_RATIO = 0.5 !0.5 is for spinning twice as fast, or 43200 seconds
	yZero = start_ymd/10000
	leap_days = (yr -yZero)/4
   day_earth = day_earth + 365*(yr -yZero) + leap_days
	tmd = day_earth/PLANET_DAY_RATIO - floor(day_earth / PLANET_DAY_RATIO)
   
end subroutine get_curr_date

 

[Simfish]

That's strange, it should recognize "floor()"

Can you test if int() and nint() work?

mod() definitely works, so I'm pretty sure that floor() should work too

 

[uart]

Hi simfish. I'm not real big on Fortran but when I have used it I've generally had problems whenever I've tried to mix integers and reals in the same expression. Honestly I've given up on trying to figure out the whys and where-fores and now I simply assign all integers to a temporary real before using them in a mixed expression.

Something like

integer :: day_earth
real :: day_earth_tmp, PLANET_DAY_RATIO

...
day_earth_tmp = day_earth
tmd = day_earth_tmp/PLANET_DAY_RATIO - floor(day_earth_tmp / PLANET_DAY_RATIO)

 

[Simfish]

Ah okay - thanks for the suggestion! I'll try to see if that works tomorrow.