NAME

Constants - The meaning of MPI constants

DATATYPES

Special types:

MPI_DATATYPE_NULL - Null, invalid type

MPI_UB - Sets upper bound of a structure

MPI_LB - Sets lower bound of a structure

MPI_PACKED - Packed type, created by MPI_Pack() and can be unpacked with MPI_Unpack()

MPI_BYTE - unsigned char

C types:

MPI_CHAR - char

MPI_SHORT - short

MPI_INT - int

MPI_LONG - long

MPI_LONG_LONG_INT - long long

MPI_SIGNED_CHAR - signed char

MPI_UNSIGNED_CHAR - unsigned char

MPI_WCHAR - wchar_t

MPI_UNSIGNED_SHORT - unsigned short

MPI_UNSIGNED - unsigned int

MPI_UNSIGNED_LONG - unsigned long

MPI_UNSIGNED_LONG_LONG - unsigned long long

MPI_INT8_T - int8_t

MPI_INT16_T - int16_t

MPI_INT32_T - int32_t

MPI_INT64_T - int64_t

MPI_INT128_T - __int128_t

MPI_UINT8_T - yint8_t

MPI_UINT16_T - yint16_t

MPI_UINT32_T - uint32_t

MPI_UINT64_T - uint64_t

MPI_UINT128_T - __uint128_t

MPI_FLOAT - float

MPI_DOUBLE - double

MPI_LONG_DOUBLE - long double

MPI_FLOAT128 - __float128

MPI_C_FLOAT_COMPLEX - complex float

MPI_C_COMPLEX - complex float

MPI_C_DOUBLE_COMPLEX - complex double

MPI_C_LONG_DOUBLE_COMPLEX - complex long double

MPI_C_FLOAT128_COMPLEX - typedef _Complex float complex_float128 __attribute__((mode(TC)))   complex_float128

MPI_C_BOOL - _Bool

FORTRAN types:

MPI_CHARACTER - CHARACTER

MPI_LOGICAL - LOGICAL

MPI_INTEGER - INTEGER

MPI_LOGICAL1 - LOGICAL*1

MPI_LOGICAL2 - LOGICAL*2

MPI_LOGICAL4 - LOGICAL*4

MPI_LOGICAL8 - LOGICAL*8

MPI_INTEGER1 - INTEGER*1

MPI_INTEGER2 - INTEGER*2

MPI_INTEGER4 - INTEGER*4

MPI_INTEGER8 - INTEGER*8

MPI_INTEGER16 - INTEGER*16

MPI_REAL - REAL

MPI_DOUBLE_PRECISION - DOUBLE PRECISION

MPI_COMPLEX - COMPLEX

MPI_DOUBLE_COMPLEX - DOUBLE_COMPLEX

MPI_LONG_DOUBLE_COMPLEX - LONG DOUBLE COMPLEX

MPI_REAL4 - REAL*4

MPI_REAL8 - REAL*8

MPI_REAL16 - REAL*16

MPI_COMPLEX8 - COMPLEX*8

MPI_COMPLEX16 - COMPLEX*16

MPI_COMPLEX32 - COMPLEX*32

MPI types:

MPI_AINT - MPI_Aint

MPI_OFFSET - MPI_Offset

The following types are used with MPI_MAXLOC and MPI_MINLOC.

MPI_FLOAT_INT - struct {float, int}

MPI_DOUBLE_INT - struct {double, int}

MPI_LONG_DOUBLE_INT - struct {long double, int}

MPI_SHORT_INT - struct {short, int}

MPI_2INT - struct {int, int}

MPI_LONG_INT - struct {long, int}

MPI_2INTEGER - array INTEGER(2, *)

MPI_2REAL - array REAL(2, *)

MPI_2DOUBLE_PRECISION - array DOUBLE PRECISION(2, *)

VERSION INFORMATION

MPI_VERSION - Major version number

MPI_SUBVERSION - Minor version number

COMMUNICATORS

MPI_COMM_NULL - NULL, invalid communicator

MPI_COMM_WORLD - Communicator containing all ranks

MPI_COMM_SELF - Communicator only containing the current rank

RANKS

MPI_ANY_SOURCE - Matches any source rank

MPI_PROC_NULL - No source or destination rank, no communication is performed

MPI_ROOT - Denotes that the current rank is the root of a collective operation

TAGS

MPI_ANY_TAG - Matches any tag

GROUPS

MPI_GROUP_NULL - NULL, invalid group

MPI_GROUP_EMPTY - A group with no members

MESSAGE BUFFERS

MPI_IN_PLACE - The send and receive buffers overlap in a collective operation

MPI_BOTTOM - Bottom of the address space

BUFFERS

MPI_MAX_PROCESSOR_NAME - Size of buffer for MPI_Get_processor_name()

MPI_MAX_OBJECT_NAME - Size of buffer for MPI_Comm_get_name() and MPI_Comm_set_name().

MPI_BSEND_OVERHEAD - Overhead per message using buffered send functions.

MPI_MAX_ERROR_STRING - Size of buffer for MPI_Error_string().

KEYVALS

MPI_KEYVAL_INVALID - NULL, invalid keyval

MPI_TAG_UB - Maximal tag

MPI_HOST - Host process ( MPI_PROC_NULL ) in this implementation

MPI_IO - IO process ( MPI_ANY_SOURCE ) in this implementation

MPI_WTIME_IS_GLOBAL - Is time globally coherent? (0 in this implementation)

OPERATIONS

MPI_OP_NULL - NULL, invalid operation

MPI_MAX - Return the maximum

MPI_MIN - Return the minimum

MPI_SUM - Return the sum

MPI_PROD - Return the product

MPI_LAND - Return the logical and

MPI_BAND - Return the bitwise and

MPI_LOR - Return the logical or

MPI_BOR - Return the bitwise or

MPI_LXOR - Return the logical xor

MPI_BXOR - Return the bitwise xor

MPI_MAXLOC - Return the maximum and location of the maximum

MPI_MINLOC - Return the minimum and location of the minimum

MPI_OP_MAX - Index of largest MPI-defined operation

ERROR RETURN VALUES

MPI_SUCCESS - No error

MPI_ERR_OTHER - Other error

MPI_ERR_COMM - Invalid communicator

MPI_ERR_COUNT - Invalid count

MPI_ERR_TYPE - Invalid data type

MPI_ERR_BUFFER - Invalid buffer

MPI_ERR_ROOT - Invalid root

MPI_ERR_TAG - Invalid tag

MPI_ERR_RANK - Invalid rank

MPI_ERR_ARG - Invalid argument

MPI_ERR_REQUEST - Invalid request

MPI_ERR_IN_STATUS - Error is returned in MPI_Status structure

MPI_ERR_PENDING - Pending exception

MPI_ERR_INTERN - Internal error. (Usually out of memory.)

MPI_ERR_GROUP - Invalid group

MPI_ERR_CANCEL - Request was cancelled

MPI_ERR_OP - Invalid operation

MPI_ERR_TOPOLOGY - Invalid topology

MPI_ERR_DIMS - Invalid dimensions

MPI_ERR_UNKNOWN - Unknown error

MPI_ERR_TRUNCATE - Message was truncated

MPI_ERR_KEYVAL - Invalid keyval

MPI_ERR_LASTCODE - Last used error code

ERROR HANDLERS

MPI_ERRHANDLER_NULL - Invalid error handler

MPI_ERRORS_ARE_FATAL - Default error handler that aborts the MPI process

MPI_ERRORS_RETURN - Causes an error code to be returned from MPI functions

MPI_ERRORS_THROW_EXCEPTIONS - Causes an exception to be raised when errors occur. MPI Functions not called from the C++ wrapper will not cause exceptions to be raised. (This avoids unwinding through arbitrary C or Fortran code, which many not be compiled with exception support.) Instead, MPI_ERR_PENDING will be returned. In this implementation, call MPI::throw_exception() to raise the correct exception when this occurs.

COMPARISONS

MPI_IDENT - Identical groups or communicators

MPI_CONGRUENT - Congruent groups or communicators

MPI_SIMILAR - Similar groups or communicators

MPI_UNEQUAL - Different groups or communicators

TOPOLOGIES

MPI_CART - Cartesian topology

MPI_GRAPH - Graph topology

PREDEFINED CONSTANTS

MPI_UNDEFINED - Undefined value

MPI_REQUEST_NULL - Invalid request

MPI_STATUS_IGNORE - No status structure required

MPI_STATUSES_IGNORE - No status structures required

FUNCTION TYPES

MPI_User_function - User defined reduction operation

MPI_Handler_function - User defined error handler

MPI_Copy_function - User defined attribute copy operation

MPI_Delete_function - User defined attribute delete operation

MPI_NULL_COPY_FN - A copy function that does nothing

MPI_DUP_FN - Default copy function that copies attributes by value

MPI_NULL_DELETE_FN - A delete function that does nothing

TYPES

MPI_STATUS - Status structure. Contains MPI_TAG , MPI_SOURCE , MPI_ERROR , and size

MPI_Request - Request handle

MPI_Datatype - Data type handle

MPI_Group - Group handle

MPI_Comm - Communicator handle

MPI_Op - Operation handle

MPI_Errhandler - Error handler handle

MPI_Aint - Address sized integer

MPI_Offset - Address-offset sized integer

MPI_Fint - FORTRAN integer

• Lockless
• MPI Documentation
• MPI_constants