[Mpi-forum] [EXTERNAL] Re: An f2c question

Rolf Rabenseifner rabenseifner at hlrs.de
Tue Aug 24 16:26:26 CDT 2021


%IMPI_VAL was a typo. %MPI_VAL is correct, see MPI-4.0 page 18, line 20.

----- Original Message -----
> From: "Pritchard" <howardp at lanl.gov>
> To: "Rolf Rabenseifner" <rabenseifner at hlrs.de>, "Main MPI Forum mailing list" <mpi-forum at lists.mpi-forum.org>
> Sent: Tuesday, August 24, 2021 10:55:51 PM
> Subject: Re: [EXTERNAL] Re: [Mpi-forum] An f2c question

> Hi Ralph,
> 
> Okay, not quite right, looks like both mpich and open mpi use %MPI_VAL, so I'll
> go with that.
> 
> Thanks,
> 
> Howard
> 
>On 8/24/21, 2:28 PM, "Rolf Rabenseifner" <rabenseifner at hlrs.de> wrote:
> 
>    Dear Howard,
> 
>    SUBROUTINE MPI_TYPE_COMMIT(DATATYPE, IERR)
>    INTEGER :: IERR
>    ! some module USE to include the definition of TYPE(MPI_Datatype)
>    TYPE(MPI_Datatype) :: DATATYPE
>    CALL MPI_X_TYPE_COMMIT(DATATYPE%IMPI_VAL, IERR)
>    RETURN
>    END
> 
>    I do not know how to handle the optional ierr, but here is the way you should
>    handle the datatype handle.
> 
>    Best regards
>    Rolf
> 
>    ----- Original Message -----
>    > From: "Main MPI Forum mailing list" <mpi-forum at lists.mpi-forum.org>
>    > To: "Main MPI Forum mailing list" <mpi-forum at lists.mpi-forum.org>
>    > Cc: "Pritchard" <howardp at lanl.gov>
>    > Sent: Tuesday, August 24, 2021 10:19:27 PM
>    > Subject: [Mpi-forum] An f2c question
> 
>    > Hi All,
>    > 
>    > 
>    > 
>    > In porting an MPI application written mainly in Fortran and a little c to using
>    > mpi_f08 rather than mpi, I’ve hit an issue.
>    > 
>    > 
>    > 
>    > Namely, how would I write example 19.11 in the MPI 4 standard if the app uses
>    > type(MPI_Type) rather than integer?
>    > 
>    > Note I’m working with an application, so the MPI_Type structure is opaque, or at
>    > least I thought we’re supposed to treat it as opaque.
>    > 
>    > 
>    > 
>    > Here’s the example:
>    > 
>    > 
>    > 
>    > Example 19.11 The example below illustrates how the Fortran MPI function
>    > 
>    > MPI_TYPE_COMMIT can be implemented by wrapping the C MPI function
>    > 
>    > MPI_Type_commit with a C wrapper to do handle conversions. In this example a
>    > Fortran-C
>    > 
>    > interface is assumed where a Fortran function is all upper case when referred to
>    > from C and
>    > 
>    > arguments are passed by addresses.
>    > 
>    > ! FORTRAN PROCEDURE
>    > 
>    > SUBROUTINE MPI_TYPE_COMMIT(DATATYPE, IERR)
>    > 
>    > INTEGER :: DATATYPE, IERR
>    > 
>    > CALL MPI_X_TYPE_COMMIT(DATATYPE, IERR)
>    > 
>    > RETURN
>    > 
>    > END
>    > 
>    > /* C wrapper */
>    > 
>    > void MPI_X_TYPE_COMMIT(MPI_Fint *f_handle, MPI_Fint *ierr)
>    > 
>    > {
>    > 
>    > MPI_Datatype datatype;
>    > 
>    > datatype = MPI_Type_f2c(*f_handle);
>    > 
>    > *ierr = (MPI_Fint)MPI_Type_commit(&datatype);
>    > 
>    > *f_handle = MPI_Type_c2f(datatype);
>    > 
>    > return;
>    > 
>    > }
>    > 
>    > 
>    > 
>    > Now I suspect that most of the time if one has some old ‘c’ code assuming a
>    > pointer to a 4 byte fortran integer argument for the datatype is just going to
>    > work, but I’d rather not just try to be lucky.
>    > 
>    > Anyway, there are fortran gurus on this project and they’d be sure to want a
>    > better solution.
>    > 
>    > 
>    > 
>    > Thanks for any help/ideas,
>    > 
>    > 
>    > 
>    > Howard
>    > 
>    > 
>    > 
>    > 
>    > 
>    > —
>    > 
>    > 
>    > 
>    > 
>    >        
>    > 
>    > Howard Pritchard
>    > 
>    > Research Scientist
>    > 
>    > HPC-ENV
>    > 
>    > 
>    > 
>    > Los Alamos National Laboratory
>    > 
>    > howardp at lanl.gov
>    > 
>    > 
>    > 
>    > [ https://www.instagram.com/losalamosnatlab/ ] [
>    > https://twitter.com/LosAlamosNatLab ] [
>    > https://www.linkedin.com/company/los-alamos-national-laboratory/ ] [
>    > https://www.facebook.com/LosAlamosNationalLab/ ]
>    > 
>    > 
>    > 
>    > 
>    > 
>    > 
>    > _______________________________________________
>    > mpi-forum mailing list
>    > mpi-forum at lists.mpi-forum.org
>    > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum
> 
>    --
>    Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner at hlrs.de .
>    High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 .
>    University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 .
>    Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner .
>     Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) .

-- 
Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner at hlrs.de .
High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 .
University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 .
Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner .
Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) .


More information about the mpi-forum mailing list