18.2.6. MPI_Abi_set_fortran_info

MPI_Abi_set_fortran_info — Sets Fortran ABI details using supplied info object

18.2.6.1. SYNTAX

18.2.6.1.1. C Syntax

int MPI_Abi_set_fortran_info(MPI_Info info)

18.2.6.1.2. Fortran Syntax

USE MPI
! or the older form: INCLUDE 'mpif.h'
MPI_ABI_SET_FORTRAN_INFO(INFO, IERROR)
    INTEGER INFO, IERROR

18.2.6.1.3. Fortran 2008 Syntax

USE mpi_f08
MPI_Abi_set_fortran_info(info, ierror)
    TYPE(MPI_Info), INTENT(IN) :: info
    INTEGER, OPTIONAL, INTENT(OUT) :: ierror

18.2.6.2. INPUT PARAMETERS

  • info: Fortran ABI details info object

18.2.6.3. OUTPUT PARAMETERS

  • ierror: Fortran only: Error status (integer).

18.2.6.4. DESCRIPTION

MPI_Abi_set_fortran_info allows the application to inform the implementation of the sizes of Fortran types and whether or not optional types are supported by the Fortran compiler. Before setting this information, the application should get this info object using MPI_Abi_get_fortran_info.

Open MPI uses its configured Fortran support when it is built with Fortran bindings. In that case, this routine returns MPI_ERR_ABI and does not replace Open MPI’s configured Fortran ABI information.

When Open MPI is built without Fortran bindings (--disable-mpi-fortran), the implementation does not know the Fortran compiler’s type properties, so MPI_Abi_get_fortran_info returns MPI_INFO_NULL and the application must supply them through info. This routine then configures the Fortran predefined datatypes from the following keys so that, for example, MPI_Type_size reports the application-specified size instead of MPI_UNDEFINED:

  • mpi_logical_size: size in bytes of Fortran default LOGICAL

  • mpi_integer_size: size in bytes of Fortran default INTEGER

  • mpi_real_size: size in bytes of Fortran default REAL; this also configures MPI_COMPLEX, which is a pair of default REAL values (2 * mpi_real_size bytes)

  • mpi_double_precision_size: size in bytes of Fortran DOUBLE PRECISION; this also configures MPI_DOUBLE_COMPLEX, which is a pair of DOUBLE PRECISION values (2 * mpi_double_precision_size bytes)

  • mpi_<type><n>_supported (boolean): whether the Fortran compiler supports the optional fixed-size type MPI_<TYPE><N> (for example mpi_integer8_supported for MPI_INTEGER8)

When a mpi_<type><n>_supported key is true, Open MPI makes the corresponding predefined datatype available with its inherent <n>-byte size when this build can represent a type of that size. A build that has no underlying type of the requested size (for example MPI_REAL16 on a platform whose widest floating-point type is 8 bytes) cannot provide that datatype; the key is accepted but the datatype remains unavailable. The keys therefore describe the Fortran compiler’s capabilities, not a guarantee of availability: as the standard specifies, an application determines whether a given optional datatype is actually available by calling MPI_Type_size and checking for MPI_UNDEFINED.

The info object must carry at least one key this routine recognizes (one of the size keys above or an mpi_<type><n>_supported key). An info that contains none of them conveys no Fortran type information and is rejected with MPI_ERR_INFO without changing any library state, so that the application’s single permitted call is not consumed by an empty request.

This routine may be called at most once; a subsequent call returns MPI_ERR_ABI. A duplicate of info is also stored for later retrieval via MPI_Abi_get_fortran_info. Note that, consistent with the above, a stored mpi_<type><n>_supported=true echoed back by MPI_Abi_get_fortran_info reflects what the application asserted about its Fortran compiler, not necessarily that MPI_<TYPE><N> is available in this build.

The composite Fortran datatypes (MPI_2INTEGER, MPI_2REAL, MPI_2DOUBLE_PRECISION, MPI_2COMPLEX, MPI_2DOUBLE_COMPLEX) are not described by these keys and remain unavailable in a --disable-mpi-fortran build.

18.2.6.5. NOTES

The application retains ownership of info and is responsible for freeing it via MPI_Info_free.

18.2.6.6. ERRORS

Almost all MPI routines return an error value; C routines as the return result of the function and Fortran routines in the last argument.

Before the error value is returned, the current MPI error handler associated with the communication object (e.g., communicator, window, file) is called. If no communication object is associated with the MPI call, then the call is considered attached to MPI_COMM_SELF and will call the associated MPI error handler. When MPI_COMM_SELF is not initialized (i.e., before MPI_Init/MPI_Init_thread, after MPI_Finalize, or when using the Sessions Model exclusively) the error raises the initial error handler. The initial error handler can be changed by calling MPI_Comm_set_errhandler on MPI_COMM_SELF when using the World model, or the mpi_initial_errhandler CLI argument to mpiexec or info key to MPI_Comm_spawn/MPI_Comm_spawn_multiple. If no other appropriate error handler has been set, then the MPI_ERRORS_RETURN error handler is called for MPI I/O functions and the MPI_ERRORS_ABORT error handler is called for all other MPI functions.

Open MPI includes three predefined error handlers that can be used:

  • MPI_ERRORS_ARE_FATAL Causes the program to abort all connected MPI processes.

  • MPI_ERRORS_ABORT An error handler that can be invoked on a communicator, window, file, or session. When called on a communicator, it acts as if MPI_Abort was called on that communicator. If called on a window or file, acts as if MPI_Abort was called on a communicator containing the group of processes in the corresponding window or file. If called on a session, aborts only the local process.

  • MPI_ERRORS_RETURN Returns an error code to the application.

MPI applications can also implement their own error handlers by calling:

Note that MPI does not guarantee that an MPI program can continue past an error.

See the MPI man page for a full list of MPI error codes.

See the Error Handling section of the MPI-5.0 standard for more information.

See also