MPI_Type_get_contents(3)
MPI_Type_get_contents - Constructs new data types
As shipped in IRIX 6.5.30. Added in IRIX 6.5.19.
NAME MPI_Type_get_contents - Constructs new data types SYNOPSIS C: #include "mpi.h" int MPI_Type_get_contents(MPI_Datatype datatype, int max_integers, int max_addresses, int max_datatypes, int array_of_integers[], MPI_Aint array_of_addresses[], MPI_Datatype array_of_datatypes[]) Fortran: INCLUDE "mpif.h" (or USE MPI) INTEGER datatype, max_integers, max_addresses, max_datatypes, array_of_integers(*), array_of_datatypes(*), ierror INTEGER(KIND=MPI_ADDRESS_KIND) array_of_addresses(*) CALL MPI_TYPE_GET_CONTENTS(datatype, max_integers, max_addresses, max_datatypes, array_of_integers, array_of_addresses, array_of_datatypes, ierror) DESCRIPTION MPI_Type_get_contents constructs new data types. Arguments are as follows: datatype Specifies the data type to be accessed (handle). datatype must be a predefined unnamed data type or a derived data type. The call is erroneous if datatype is a predefined named data type. MPI_Type_get_contents can be invoked with a data type argument that was constructed by using MPI_TYPE_CREATE_F90_REAL, MPI_TYPE_CREATE_F90_INTEGER, or MPI_TYPE_CREATE_F90_COMPLEX (an unnamed predefined data type). In such a case, an empty array_of_datatypes is returned. The definition of data type equivalence implies that equivalent predefined data types are equal. By requiring the same handle for named predefined data types, it is possible to use the == or .EQ. comparison operator to determine the data type involved. max_integers Specifies the number of elements in array_of_integers (non-negative integer). max_addresses Specifies the number of elements in array_of_addresses (non-negative integer). max_datatypes Specifies the number of elements in array_of_datatypes (non-negative integer). array_of_integers Returns integer arguments used in constructing the data type (array of integers). array_of_addresses Returns address arguments used in constructing the data type (array of integers). array_of_datatypes Returns data type arguments used in constructing the data type (array of handles). ierror Specifies the return code value for successful completion, which is in MPI_SUCCESS. MPI_SUCCESS is defined in the mpif.h file. The values given for max_integers, max_addresses, and max_datatypes must be at least as large as the corresponding values returned in num_integers, num_addresses, and num_datatypes in the MPI_TYPE_GET_ENVELOPE call for the same data type argument. The data types returned in array_of_datatypes are handles to data type objects that are equivalent to the data types used in the original construction call. If these were derived data types, the returned data types are new data type objects, and the user is responsible for freeing these data types by using MPI_TYPE_FREE. If these were predefined data types, the returned data type is a constant equal to that predefined data type and cannot be freed. The committed state of returned derived data types is undefined (that is, the data types might or might not be committed). Furthermore, the content of attributes of returned data types is undefined. In the MPI-1 data type constructor calls, the address arguments in Fortran are of type INTEGER. In the new MPI-2 calls, the address arguments are of type INTEGER(KIND=MPI_ADDRESS_KIND). The MPI_Type_get_contents call returns all addresses in an argument of type INTEGER(KIND=MPI_ADDRESS_KIND). This is true even if the old MPI-1 calls were used. Thus, the location of returned values can be thought of as being returned by the C bindings. It can also be determined by examining the new MPI-2 calls for data type constructors for the deprecated MPI-1 calls that involve addresses. Since the Fortran verison of this function uses MPI_ADDRESS_KIND, which is supported only for ABI 64, the Fortran version of this function is also available only for ABI 64. By having all address arguments returned in the array_of_addresses argument, the result from a C and Fortran decoding of a data type gives the result in the same argument. It is assumed that an integer of type INTEGER(KIND=MPI_ADDRESS_KIND) will be at least as large as the integer argument used in data type construction with the old MPI-1 calls, so no loss of information will occur. EXAMPLES The following code defines what values are placed in each entry of the returned arrays depending on the data type constructor used for data type. It also specifies the size of the arrays needed, which are the values returned by MPI_TYPE_GET_ENVELOPE. Fortran calls: PARAMETER (LARGE = 1000) INTEGER TYPE, NI, NA, ND, COMBINER, I(LARGE), D(LARGE), IERROR INTEGER(KIND=MPI_ADDRESS_KIND) A(LARGE) ! CONSTRUCT DATATYPE TYPE (NOT SHOWN) CALL MPI_TYPE_GET_ENVELOPE(TYPE, NI, NA, ND, COMBINER, IERROR) IF ((NI .GT. LARGE) .OR. (NA .GT. LARGE) .OR. (ND .GT. LARGE)) THEN WRITE (*, *) "NI, NA, OR ND = ", NI, NA, ND, & " RETURNED BY MPI_TYPE_GET_ENVELOPE IS LARGER THAN LARGE = ", LARGE CALL MPI_ABORT(MPI_COMM_WORLD, 99) ENDIF CALL MPI_TYPE_GET_CONTENTS(TYPE, NI, NA, ND, I, A, D, IERROR) C calls: #define LARGE 1000 int ni, na, nd, combiner, i[LARGE]; MPI_Aint a[LARGE]; MPI_Datatype type, d[LARGE]; /* construct datatype type (not shown) */ MPI_Type_get_envelope(type, &ni, &na, &nd, &combiner); if ((ni > LARGE) || (na > LARGE) || (nd > LARGE)) { fprintf(stderr, "ni, na, or nd = %d %d %d returned by ", ni, na, nd); fprintf(stderr, "MPI_Type_get_envelope is larger than LARGE = %d0, LARGE); MPI_Abort(MPI_COMM_WORLD, 99); }; MPI_Type_get_contents(type, ni, na, nd, i, a, d); SEE ALSO MPI_File_open(3), MPI_Type_get_envelope(3)