Why gfortran give my mpi module various warnings while intel fortran just fine

[CRquantum]

@themos

You shouldn’t be including mpif.h AND using mpi module. Just “use mpi”.

Thank you very much. The include mpif.h and use mpi seems the same.

About interfaces, take mpi_bcast for example,

MPI_Bcast(3) man page (version 4.1.6)

It says,

Fortran Syntax

USE MPI
! or the older form: INCLUDE ’mpif.h’
MPI_BCAST(BUFFER, COUNT, DATATYPE, ROOT, COMM, IERROR)
    <type>    BUFFER(*)
    INTEGER    COUNT, DATATYPE, ROOT, COMM, IERROR

I can be wrong, but t seems mpi_bcast should be able to accept different datatype and buffer already, right?

I think if in the code I just directly use the intrinsic mpi_bcast for different datatype it may not have warning. But perhaps because in my mpi module, I define my subroutine bcast with overloading for different datatype, see

interface bcast ! broadcast from process 0
   module procedure bcasti1,bcasti1d,bcasti2d,bcasti3d,bcasti4d
   module procedure bcasti1_8,bcasti1d_8
   module procedure bcastr1,bcastr1d,bcastr2d,bcastr3d,bcastr4d
   module procedure bcastlogi
   module procedure bcastchar
end interface bcast

and inside each my-defined subroutine like

   subroutine bcasti1(i)
   integer(kind=i4) :: i
   integer :: ierror
   call mpi_bcast(i,1,mpii4,0,mpi_comm_world,ierror)
   return
   end subroutine bcasti1

   subroutine bcasti1d(i)
   integer(kind=i4) :: i(:)
   integer :: ierror
   call mpi_bcast(i,size(i),mpii4,0,mpi_comm_world,ierror)
   return
   end subroutine bcasti1d

As explained by @RonShepard

Mpif90 gfortran strange warning for mpi subroutines: 'Rank mismatch between actual argument...'

I think what is happening is that the compiler is generating interfaces to the called subrouitine internally, and using that interface to compare multiple calls to the subroutine. So whichever call it sees first effectively defines the interface, and then any subsequent calls are compared to that. In your case, one call has a scalar argument and the other call has an array argument. If only one call exists, then no problem, it is only when multiple calls are there that the inconsistency is detec…

In subroutine bcasti1, i is a number with integer(kind=i4) :: i, in In subroutine bcasti1d, i is a 1d array with integer(kind=i4) :: i(:). The interface of integer and interger array is different, perhaps that make gfortran’s intrinsic mpi_bcast a little unhappy so it gives me an warning. But Intel Fortran seems OK. In reality, when the program run, it seems always fine for some years on supercomputers.

If you may please provide a concrete minimal example as to how to overcome this warning in gfortran, based on my module, it would be very appreciated.