Rechercher une page de manuel

Chercher une autre page de manuel:


Langue: en

Version: March 2006 (debian - 07/07/09)

Section: 1 (Commandes utilisateur)


orterun, mpirun, mpiexec - Execute serial and parallel jobs in Open MPI.

Note: mpirun, mpiexec, and orterun are all exact synonyms for each other. Using any of the names will result in exactly identical behavior.


Single Process Multiple Data (SPMD) Model:

mpirun [ options ] <program> [ <args> ]

Multiple Instruction Multiple Data (MIMD) Model:

mpirun [ global_options ]
       [ local_options1 ] <program1> [ <args1> ] :
       [ local_options2 ] <program2> [ <args2> ] :
       ... : 
       [ local_optionsN ] <programN> [ <argsN> ]

Note that in both models, invoking mpirun via an absolute path name is equivalent to specifying the --prefix option with a <dir> value equivalent to the directory where mpirun resides, minus its last subdirectory. For example:

    shell$ /usr/local/bin/mpirun ...

is equivalent to

    shell$ mpirun --prefix /usr/local


If you are simply looking for how to run an MPI application, you probably want to use a command line of the following form:

    shell$ mpirun [ -np X ] [ --hostfile <filename> ]  <program>

This will run X copies of <program> in your current run-time environment (if running under a supported resource manager, Open MPI's mpirun will usually automatically use the corresponding resource manager process starter, as opposed to, for example, rsh or ssh, which require the use of a hostfile, or will default to running all X copies on the localhost), scheduling (by default) in a round-robin fashion by CPU slot. See the rest of this page for more details.


mpirun will send the name of the directory where it was invoked on the local node to each of the remote nodes, and attempt to change to that directory. See the "Current Working Directory" section below for further details.
Pass these run-time arguments to every new process. These must always be the last arguments to mpirun. If an app context file is used, <args> will be ignored.
The program executable. This is identified as the first non-recognized argument to mpirun.
-aborted, --aborted <#>
Set the maximum number of aborted processes to display.
--app <appfile>
Provide an appfile, ignoring all other command line options.
-bynode, --bynode
Allocate (map) the processes by node in a round-robin scheme.
-byslot, --byslot
Allocate (map) the processes by slot in a round-robin scheme. This is the default.
-c <#>
Synonym for -np.
-debug, --debug
Invoke the user-level debugger indicated by the orte_base_user_debugger MCA parameter.
-debugger, --debugger
Sequence of debuggers to search for when --debug is used (i.e. a synonym for orte_base_user_debugger MCA parameter).
-gmca, --gmca <key> <value>
Pass global MCA parameters that are applicable to all contexts. <key> is the parameter name; <value> is the parameter value.
-h, --help
Display help for this command
-H <host1,host2,...,hostN>
Synonym for -host.
-host, --host <host1,host2,...,hostN>
List of hosts on which to invoke processes.
-hostfile, --hostfile <hostfile>
Provide a hostfile to use.
-machinefile, --machinefile <machinefile>
Synonym for -hostfile.
-mca, --mca <key> <value>
Send arguments to various MCA modules. See the "MCA" section, below.
-n, --n <#>
Synonym for -np.
-nolocal, --nolocal
Do not run any copies of the launched application on the same node as orterun is running. This option will override listing the localhost with --host or any other host-specifying mechanism.
-nooversubscribe, --nooversubscribe
Do not oversubscribe any nodes; error (without starting any processes) if the requested number of processes would cause oversubscription. This option implicitly sets "max_slots" equal to the "slots" value for each node.
-np <#>
Run this many copies of the program on the given nodes. This option indicates that the specified file is an executable program and not an application context. If no value is provided for the number of copies to execute (i.e., neither the "-np" nor its synonyms are provided on the command line), Open MPI will automatically execute a copy of the program on each process slot (see below for description of a "process slot"). This feature, however, can only be used in the SPMD model and will return an error (without beginning execution of the application) otherwise.
-nw, --nw
Launch the processes and do not wait for their completion. mpirun will complete as soon as successful launch occurs.
-path, --path <path>
<path> that will be used when attempting to locate requested executables.
--prefix <dir>
Prefix directory that will be used to set the PATH and LD_LIBRARY_PATH on the remote node before invoking Open MPI or the target process. See the "Remote Execution" section, below.
-q, --quiet
Suppress informative messages from orterun during application execution.
--tmpdir <dir>
Set the root for the session directory tree for mpirun only.
-tv, --tv
Launch processes under the TotalView debugger. Deprecated backwards compatibility flag. Synonym for --debug.
--universe <username@hostname:universe_name>
For this application, set the universe name as:
-v, --verbose
Be verbose
-V, --version
Print version number. If no other arguments are given, this will also cause orterun to exit.
-wd <dir>
Synonym for -wdir.
-wdir <dir>
Change to the directory <dir> before the user's program executes. See the "Current Working Directory" section for notes on relative paths. Note: If the -wdir option appears both on the command line and in an application context, the context will take precedence over the command line.
-x <env>
Export the specified environment variables to the remote nodes before executing the program. Existing environment variables can be specified (see the Examples section, below), or new variable names specified with corresponding values. The parser for the -x option is not very sophisticated; it does not even understand quoted values. Users are advised to set variables in the environment, and then use -x to export (not define) them.

The following options are useful for developers; they are not generally useful to most ORTE and/or MPI users:

-d, --debug-devel
Enable debugging of the OpenRTE (the run-time layer in Open MPI). This is not generally useful for most users.
Enable debugging of any OpenRTE daemons used by this application.
Enable debugging of any OpenRTE daemons used by this application, storing output in files.
Do not detach OpenRTE daemons used by this application.


One invocation of mpirun starts an MPI application running under Open MPI. If the application is single process multiple data (SPMD), the application can be specified on the mpirun command line.

If the application is multiple instruction multiple data (MIMD), comprising of multiple programs, the set of programs and argument can be specified in one of two ways: Extended Command Line Arguments, and Application Context.

An application context describes the MIMD program set including all arguments in a separate file. This file essentially contains multiple mpirun command lines, less the command name itself. The ability to specify different options for different instantiations of a program is another reason to use an application context.

Extended command line arguments allow for the description of the application layout on the command line using colons (:) to separate the specification of programs and arguments. Some options are globally set across all specified programs (e.g. --hostfile), while others are specific to a single program (e.g. -np).

Process Slots

Open MPI uses "slots" to represent a potential location for a process. Hence, a node with 2 slots means that 2 processes can be launched on that node. For performance, the community typically equates a "slot" with a physical CPU, thus ensuring that any process assigned to that slot has a dedicated processor. This is not, however, a requirement for the operation of Open MPI.

Slots can be specified in hostfiles after the hostname. For example: slots=4
Indicates that there are 4 process slots on host1.

If no slots value is specified, then Open MPI will automatically assign a default value of "slots=1" to that host.

When running under resource managers (e.g., SLURM, Torque, etc.), Open MPI will obtain both the hostnames and the number of slots directly from the resource manger. For example, if running under a SLURM job, Open MPI will automatically receive the hosts that SLURM has allocated to the job as well as how many slots on each node that SLURM says are usable - in most high-performance environments, the slots will equate to the number of processors on the node.

When deciding where to launch processes, Open MPI will first fill up all available slots before oversubscribing (see "Location Nomenclature", below, for more details on the scheduling algorithms available). Unless told otherwise, Open MPI will arbitrarily oversubscribe nodes. For example, if the only node available is the localhost, Open MPI will run as many processes as specified by the -n (or one of its variants) command line option on the localhost (although they may run quite slowly, since they'll all be competing for CPU and other resources).

Limits can be placed on oversubscription with the "max_slots" attribute in the hostfile. For example: slots=4 max_slots=6
Indicates that there are 4 process slots on host2. Further, Open MPI is limited to launching a maximum of 6 processes on host2. slots=2 max_slots=2
Indicates that there are 2 process slots on host3 and that no oversubscription is allowed (similar to the --nooversubscribe option). max_slots=2
Shorthand; same as listing "slots=2 max_slots=2".

Note that Open MPI's support for resource managers does not currently set the "max_slots" values for hosts. If you wish to prevent oversubscription in such scenarios, use the --nooversubscribe option.

In scenarios where the user wishes to launch an application across all available slots by not providing a "-n" option on the mpirun command line, Open MPI will launch a process on each process slot for each host within the provided environment. For example, if a hostfile has been provided, then Open MPI will spawn processes on each identified host up to the "slots=x" limit if oversubscription is not allowed. If oversubscription is allowed (the default), then Open MPI will spawn processes on each host up to the "max_slots=y" limit if that value is provided. In all cases, the "-bynode" and "-byslot" mapping directives will be enforced to ensure proper placement of process ranks.

Location Nomenclature

As described above, mpirun can specify arbitrary locations in the current Open MPI universe. Locations can be specified either by CPU or by node.

Note: This nomenclature does not force Open MPI to bind processes to CPUs -- specifying a location "by CPU" is really a convenience mechanism for SMPs that ultimately maps down to a specific node.

Specifying locations by node will launch one copy of an executable per specified node. Using the --bynode option tells Open MPI to use all available nodes. Using the --byslot option tells Open MPI to use all slots on an available node before allocating resources on the next available node. For example:

mpirun --bynode -np 4 a.out
Runs one copy of the the executable a.out on all available nodes in the Open MPI universe. MPI_COMM_WORLD rank 0 will be on node0, rank 1 will be on node1, etc. Regardless of how many slots are available on each of the nodes.
mpirun --byslot -np 4 a.out
Runs one copy of the the executable a.out on each slot on a given node before running the executable on other available nodes.

Specifying Hosts

Hosts can be specified in a number of ways. The most common of which is in a 'hostfile' or 'machinefile'. If our hostfile contain the following information:

   shell$ cat my-hostfile
   node00 slots=2
   node01 slots=2
   node02 slots=2

mpirun --hostfile my-hostfile -np 3 a.out
This will run one copy of the executable a.out on hosts node00,node01, and node02.

Another method for specifying hosts is directly on the command line. Here can can include and exclude hosts from the set of hosts to run on. For example:

mpirun -np 3 --host a a.out
Runs three copies of the executable a.out on host a.
mpirun -np 3 --host a,b,c a.out
Runs one copy of the executable a.out on hosts a, b, and c.
mpirun -np 3 --hostfile my-hostfile --host node00 a.out
Runs three copies of the executable a.out on host node00.
mpirun -np 3 --hostfile my-hostfile --host node10 a.out
This will prompt an error since node10 is not in my-hostfile; mpirun will abort.
shell$ mpirun -np 1 --host a hostname : -np 2 --host b,c uptime
Runs one copy of the executable hostname on host a. And runs one copy of the executable uptime on hosts b and c.

No Local Launch

Using the --nolocal option to orterun tells the system to not launch any of the application processes on the same node that orterun is running. While orterun typically blocks and consumes few system resources, this option can be helpful for launching very large jobs where orterun may actually need to use noticable amounts of memory and/or processing time. --nolocal allows orteun to run without sharing the local node with the launched applications, and likewise allows the launched applications to run unhindered by orterun's system usage.

Note that --nolocal will override any other specification to launch the application on the local node. It will disqualify the localhost from being capable of running any processes in the application.

shell$ mpirun -np 1 --host localhost --nolocal hostname
This example will result in an error because orterun will not find anywhere to launch the application.

No Oversubscription

Using the --nooversubscribe option causes Open MPI to implicitly set the "max_slots" value to be the same as the "slots" value for each node. This can be especially helpful when running jobs under a resource manager because Open MPI currently only sets the "slots" value for each node that it obtains from the resource manager.

Application Context or Executable Program?

To distinguish the two different forms, mpirun looks on the command line for --app option. If it is specified, then the file named on the command line is assumed to be an application context. If it is not specified, then the file is assumed to be an executable program.

Locating Files

If no relative or absolute path is specified for a file, Open MPI will look for files by searching the directories in the user's PATH environment variable as defined on the source node(s).

If a relative directory is specified, it must be relative to the initial working directory determined by the specific starter used. For example when using the rsh or ssh starters, the initial directory is $HOME by default. Other starters may set the initial directory to the current working directory from the invocation of mpirun.

Current Working Directory

The -wdir mpirun option (and its synonym, -wd) allows the user to change to an arbitrary directory before the program is invoked. It can also be used in application context files to specify working directories on specific nodes and/or for specific applications.

If the -wdir option appears both in a context file and on the command line, the context file directory will override the command line value.

If the -wdir option is specified, Open MPI will attempt to change to the specified directory on all of the remote nodes. If this fails, mpirun will abort.

If the -wdir option is not specified, Open MPI will send the directory name where mpirun was invoked to each of the remote nodes. The remote nodes will try to change to that directory. If they are unable (e.g., if the directory does not exit on that node), then Open MPI will use the default directory determined by the starter.

All directory changing occurs before the user's program is invoked; it does not wait until MPI_INIT is called.

Standard I/O

Open MPI directs UNIX standard input to /dev/null on all processes except the MPI_COMM_WORLD rank 0 process. The MPI_COMM_WORLD rank 0 process inherits standard input from mpirun. Note: The node that invoked mpirun need not be the same as the node where the MPI_COMM_WORLD rank 0 process resides. Open MPI handles the redirection of mpirun's standard input to the rank 0 process.

Open MPI directs UNIX standard output and error from remote nodes to the node that invoked mpirun and prints it on the standard output/error of mpirun. Local processes inherit the standard output/error of mpirun and transfer to it directly.

Thus it is possible to redirect standard I/O for Open MPI applications by using the typical shell redirection procedure on mpirun.

      shell$ mpirun -np 2 my_app < my_input > my_output

Note that in this example only the MPI_COMM_WORLD rank 0 process will receive the stream from my_input on stdin. The stdin on all the other nodes will be tied to /dev/null. However, the stdout from all nodes will be collected into the my_output file.

Signal Propagation

When orterun receives a SIGTERM and SIGINT, it will attempt to kill the entire job by sending all processes in the job a SIGTERM, waiting a small number of seconds, then sending all processes in the job a SIGKILL. SIGUSR1 and SIGUSR2 signals received by orterun are propagated to all processes in the job. Other signals are not currently propagated by orterun.

Process Termination / Signal Handling

During the run of an MPI application, if any rank dies abnormally (either exiting before invoking MPI_FINALIZE, or dying as the result of a signal), mpirun will print out an error message and kill the rest of the MPI application.

User signal handlers should probably avoid trying to cleanup MPI state (Open MPI is, currently, neither thread-safe nor async-signal-safe). For example, if a segmentation fault occurs in MPI_SEND (perhaps because a bad buffer was passed in) and a user signal handler is invoked, if this user handler attempts to invoke MPI_FINALIZE, Bad Things could happen since Open MPI was already "in" MPI when the error occurred. Since mpirun will notice that the process died due to a signal, it is probably not necessary (and safest) for the user to only clean up non-MPI state.

Process Environment

Processes in the MPI application inherit their environment from the Open RTE daemon upon the node on which they are running. The environment is typically inherited from the user's shell. On remote nodes, the exact environment is determined by the boot MCA module used. The rsh launch module, for example, uses either rsh/ssh to launch the Open RTE daemon on remote nodes, and typically executes one or more of the user's shell-setup files before launching the Open RTE daemon. When running dynamically linked applications which require the LD_LIBRARY_PATH environment variable to be set, care must be taken to ensure that it is correctly set when booting Open MPI.

See the "Remote Execution" section for more details.

Remote Execution

Open MPI requires that the PATH environment variable be set to find executables on remote nodes (this is typically only necessary in rsh- or ssh-based environments -- batch/scheduled environments typically copy the current environment to the execution of remote jobs, so if the current environment has PATH and/or LD_LIBRARY_PATH set properly, the remote nodes will also have it set properly). If Open MPI was compiled with shared library support, it may also be necessary to have the LD_LIBRARY_PATH environment variable set on remote nodes as well (especially to find the shared libraries required to run user MPI applications).

However, it is not always desirable or possible to edit shell startup files to set PATH and/or LD_LIBRARY_PATH. The --prefix option is provided for some simple configurations where this is not possible.

The --prefix option takes a single argument: the base directory on the remote node where Open MPI is installed. Open MPI will use this directory to set the remote PATH and LD_LIBRARY_PATH before executing any Open MPI or user applications. This allows running Open MPI jobs without having pre-configued the PATH and LD_LIBRARY_PATH on the remote nodes.

Open MPI adds the basename of the current node's "bindir" (the directory where Open MPI's executables are installed) to the prefix and uses that to set the PATH on the remote node. Similarly, Open MPI adds the basename of the current node's "libdir" (the directory where Open MPI's libraries are installed) to the prefix and uses that to set the LD_LIBRARY_PATH on the remote node. For example:

Local bindir:
Local libdir:

If the following command line is used:

    shell$ mpirun --prefix /remote/node/directory

Open MPI will add "/remote/node/directory/bin" to the PATH and "/remote/node/directory/lib64" to the D_LIBRARY_PATH on the remote node before attempting to execute anything.

Note that --prefix can be set on a per-context basis, allowing for different values for different nodes.

The --prefix option is not sufficient if the installation paths on the remote node are different than the local node (e.g., if "/lib" is used on the local node, but "/lib64" is used on the remote node), or if the installation paths are something other than a subdirectory under a common prefix.

Note that executing mpirun via an absolute pathname is equivalent to specifying --prefix without the last subdirectory in the absolute pathname to mpirun. For example:

    shell$ /usr/local/bin/mpirun ...

is equivalent to

    shell$ mpirun --prefix /usr/local

Exported Environment Variables

All environment variables that are named in the form OMPI_* will automatically be exported to new processes on the local and remote nodes. The -x option to mpirun can be used to export specific environment variables to the new processes. While the syntax of the -x option allows the definition of new variables, note that the parser for this option is currently not very sophisticated - it does not even understand quoted values. Users are advised to set variables in the environment and use -x to export them; not to define them.

MCA (Modular Component Architecture)

The -mca switch allows the passing of parameters to various MCA modules. MCA modules have direct impact on MPI programs because they allow tunable parameters to be set at run time (such as which BTL communication device driver to use, what parameters to pass to that BTL, etc.).

The -mca switch takes two arguments: <key> and <value>. The <key> argument generally specifies which MCA module will receive the value. For example, the <key> "btl" is used to select which BTL to be used for transporting MPI messages. The <value> argument is the value that is passed. For example:

mpirun -mca btl tcp,self -np 1 foo
Tells Open MPI to use the "tcp" and "self" BTLs, and to run a single copy of "foo" an allocated node.
mpirun -mca btl self -np 1 foo
Tells Open MPI to use the "self" BTL, and to run a single copy of "foo" an allocated node.

The -mca switch can be used multiple times to specify different <key> and/or <value> arguments. If the same <key> is specified more than once, the <value>s are concatenated with a comma (",") separating them.

Note: The -mca switch is simply a shortcut for setting environment variables. The same effect may be accomplished by setting corresponding environment variables before running mpirun. The form of the environment variables that Open MPI sets are:


Note that the -mca switch overrides any previously set environment variables. Also note that unknown <key> arguments are still set as environment variable -- they are not checked (by mpirun) for correctness. Illegal or incorrect <value> arguments may or may not be reported -- it depends on the specific MCA module.


Be sure to also see the examples in the "Location Nomenclature" section, above.
mpirun -np 1 prog1
Load and execute prog1 on one node. Search the user's $PATH for the executable file on each node.
mpirun -np 8 --byslot prog1
Run 8 copies of prog1 wherever Open MPI wants to run them.
mpirun -np 4 -mca btl ib,tcp,self prog1
Run 4 copies of prog1 using the "ib", "tcp", and "self" BTL's for the transport of MPI messages.


mpirun returns 0 if all ranks started by mpirun exit after calling MPI_FINALIZE. A non-zero value is returned if an internal error occurred in mpirun, or one or more ranks exited before calling MPI_FINALIZE. If an internal error occurred in mpirun, the corresponding error code is returned. In the event that one or more ranks exit before calling MPI_FINALIZE, the return value of the rank of the process that mpirun first notices died before calling MPI_FINALIZE will be returned. Note that, in general, this will be the first rank that died but is not guaranteed to be so.

However, note that if the -nw switch is used, the return value from mpirun does not indicate the exit status of the ranks.

Les bouton occupent 0% du temps processeur !
-- Jayce - \o/ --