17.2.77. MPI_Comm_spawn
MPI_Comm_spawn — Spawns a number of identical binaries.
17.2.77.1. SYNTAX
17.2.77.1.1. C Syntax
#include <mpi.h>
int MPI_Comm_spawn(const char *command, char *argv[], int maxprocs,
MPI_Info info, int root, MPI_Comm comm,
MPI_Comm *intercomm, int array_of_errcodes[])
17.2.77.1.2. Fortran Syntax
USE MPI
! or the older form: INCLUDE 'mpif.h'
MPI_COMM_SPAWN(COMMAND, ARGV, MAXPROCS, INFO, ROOT, COMM,
INTERCOMM, ARRAY_OF_ERRCODES, IERROR)
CHARACTER*(*) COMMAND, ARGV(*)
INTEGER INFO, MAXPROCS, ROOT, COMM, INTERCOMM,
ARRAY_OF_ERRCODES(*), IERROR
17.2.77.1.3. Fortran 2008 Syntax
USE mpi_f08
MPI_Comm_spawn(command, argv, maxprocs, info, root, comm, intercomm,
array_of_errcodes, ierror)
CHARACTER(LEN=*), INTENT(IN) :: command, argv(*)
INTEGER, INTENT(IN) :: maxprocs, root
TYPE(MPI_Info), INTENT(IN) :: info
TYPE(MPI_Comm), INTENT(IN) :: comm
TYPE(MPI_Comm), INTENT(OUT) :: intercomm
INTEGER :: array_of_errcodes(*)
INTEGER, OPTIONAL, INTENT(OUT) :: ierror
17.2.77.2. INPUT PARAMETERS
command
: Name of program to be spawned (string, significant only at root).argv
: Arguments to command (array of strings, significant only at root).maxprocs
: Maximum number of processes to start (integer, significant only at root).info
: A set of key-value pairs telling the runtime system where and how to start the processes (handle, significant only at root).root
: Rank of process in which previous arguments are examined (integer).comm
: Intracommunicator containing group of spawning processes (handle).
17.2.77.3. OUTPUT PARAMETER
intercomm
: Intercommunicator between original group and the newly spawned group (handle).array_of_errcodes
: One code per process (array of integers).ierror
: Fortran only: Error status (integer).
17.2.77.4. DESCRIPTION
MPI_Comm_spawn tries to start maxprocs identical copies of the MPI program specified by command, establishing communication with them and returning an intercommunicator. The spawned processes are referred to as children. The children have their own MPI_COMM_WORLD, which is separate from that of the parents. MPI_Comm_spawn is collective over comm, and also may not return until MPI_Init has been called in the children. Similarly, MPI_Init in the children may not return until all parents have called MPI_Comm_spawn. In this sense, MPI_Comm_spawn in the parents and MPI_Init in the children form a collective operation over the union of parent and child processes. The intercommunicator returned by MPI_Comm_spawn contains the parent processes in the local group and the child processes in the remote group. The ordering of processes in the local and remote groups is the same as the as the ordering of the group of the comm in the parents and of MPI_COMM_WORLD of the children, respectively. This intercommunicator can be obtained in the children through the function MPI_Comm_get_parent.
The MPI standard allows an implementation to use the MPI_UNIVERSE_SIZE attribute of MPI_COMM_WORLD to specify the number of processes that will be active in a program. Although this implementation of the MPI standard defines MPI_UNIVERSE_SIZE, it does not allow the user to set its value. If you try to set the value of MPI_UNIVERSE_SIZE, you will get an error message.
The command Argument
The command argument is a string containing the name of a program to be spawned. The string is null-terminated in C. In Fortran, leading and trailing spaces are stripped. MPI looks for the file first in the working directory of the spawning process.
The argv Argument
argv is an array of strings containing arguments that are passed to the program. The first element of argv is the first argument passed to command, not, as is conventional in some contexts, the command itself. The argument list is terminated by NULL in C and an empty string in Fortran (note that it is the MPI application’s responsibility to ensure that the last entry of the argv array is an empty string; the compiler will not automatically insert it). In Fortran, leading and trailing spaces are always stripped, so that a string consisting of all spaces is considered an empty string. The constant MPI_ARGV_NULL may be used in C and Fortran to indicate an empty argument list. In C, this constant is the same as NULL.
In C, the MPI_Comm_spawn argument argv differs from the argv argument of main in two respects. First, it is shifted by one element. Specifically, argv[0] of main contains the name of the program (given by command). argv[1] of main corresponds to argv[0] in MPI_Comm_spawn, argv[2] of main to argv[1] of MPI_Comm_spawn, and so on. Second, argv of MPI_Comm_spawn must be null-terminated, so that its length can be determined. Passing an argv of MPI_ARGV_NULL to MPI_Comm_spawn results in main receiving argc of 1 and an argv whose element 0 is the name of the program.
The maxprocs Argument
Open MPI tries to spawn maxprocs processes. If it is unable to spawn maxprocs processes, it raises an error of class MPI_ERR_SPAWN. If MPI is able to spawn the specified number of processes, MPI_Comm_spawn returns successfully and the number of spawned processes, m, is given by the size of the remote group of intercomm.
A spawn call with the default behavior is called hard. A spawn call for which fewer than maxprocs processes may be returned is called soft.
The info Argument
The info argument is an opaque handle of type MPI_Info in C and
INTEGER in Fortran. It is a container for a number of user-specified
(key,value) pairs. key and value are strings (null-terminated
char *
in C, character*(*)
in Fortran). Routines to create and
manipulate the info argument are described in Section 4.10 of the
MPI-2 standard.
For the SPAWN calls, info provides additional, implementation-dependent instructions to MPI and the runtime system on how to start processes. An application may pass MPI_INFO_NULL in C or Fortran. Portable programs not requiring detailed control over process locations should use MPI_INFO_NULL.
The following keys for info are recognized in Open MPI. (The reserved values mentioned in Section 5.3.4 of the MPI-2 standard are not implemented.)
Key Type Description
--- ---- -----------
host char * Host on which the process should be
spawned. See the orte_host man
page for an explanation of how this
will be used.
hostfile char * Hostfile containing the hosts on which
the processes are to be spawned. See
the orte_hostfile man page for
an explanation of how this will be
used.
add-host char * Add the specified host to the list of
hosts known to this job and use it for
the associated process. This will be
used similarly to the -host option.
add-hostfile char * Hostfile containing hosts to be added
to the list of hosts known to this job
and use it for the associated
process. This will be used similarly
to the -hostfile option.
wdir char * Directory where the executable is
located. If files are to be
pre-positioned, then this location is
the desired working directory at time
of execution - if not specified, then
it will automatically be set to
ompi_preload_files_dest_dir.
ompi_prefix char * Same as the --prefix command line
argument to mpirun.
ompi_preload_binary bool If set to true, pre-position the
specified executable onto the remote
host. A destination directory must
also be provided.
ompi_preload_files char * A comma-separated list of files that
are to be pre-positioned in addition
to the executable. Note that this
option does not depend upon
ompi_preload_binary - files can
be moved to the target even if an
executable is not moved.
ompi_stdin_target char * Comma-delimited list of ranks to
receive stdin when forwarded.
ompi_non_mpi bool If set to true, launching a non-MPI
application; the returned communicator
will be MPI_COMM_NULL. Failure to set
this flag when launching a non-MPI
application will cause both the child
and parent jobs to "hang".
ompi_param char * Pass an OMPI MCA parameter to the
child job. If that parameter already
exists in the environment, the value
will be overwritten by the provided
value.
mapper char * Mapper to be used for this job
map_by char * Mapping directive indicating how
processes are to be mapped (slot,
node, socket, etc.).
rank_by char * Ranking directive indicating how
processes are to be ranked (slot,
node, socket, etc.).
bind_to char * Binding directive indicating how
processes are to be bound (core, slot,
node, socket, etc.).
path char * List of directories to search for
the executable
npernode char * Number of processes to spawn on
each node of the allocation
pernode bool Equivalent to npernode of 1
ppr char * Spawn specified number of processes
on each of the identified object type
env char * Newline-delimited list of envars to
be passed to the spawned procs
bool info keys are actually strings but are evaluated as follows: if the string value is a number, it is converted to an integer and cast to a boolean (meaning that zero integers are false and non-zero values are true). If the string value is (case-insensitive) “yes” or “true”, the boolean is true. If the string value is (case-insensitive) “no” or “false”, the boolean is false. All other string values are unrecognized, and therefore false.
The root Argument
All arguments before the root argument are examined only on the process whose rank in comm is equal to root. The value of these arguments on other processes is ignored.
The array_of_errcodes Argument
The array_of_errcodes is an array of length maxprocs in which MPI reports the status of the processes that MPI was requested to start. If all maxprocs processes were spawned, array_of_errcodes is filled in with the value MPI_SUCCESS. If anyof the processes are not spawned, array_of_errcodes is filled in with the value MPI_ERR_SPAWN. In C or Fortran, an application may pass MPI_ERRCODES_IGNORE if it is not interested in the error codes.
17.2.77.5. NOTES
Completion of MPI_Comm_spawn in the parent does not necessarily mean that MPI_Init has been called in the children (although the returned intercommunicator can be used immediately).
17.2.77.6. 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-3.1 standard for more information.