Scroll to navigation

MPI4PY(3) MPI for Python MPI4PY(3)

NAME

mpi4py - MPI for Python

Lisandro Dalcin
dalcinl@gmail.com
September 29, 2024

Abstract

MPI for Python provides Python bindings for the Message Passing Interface (MPI) standard, allowing Python applications to exploit multiple processors on workstations, clusters and supercomputers.

This package builds on the MPI specification and provides an object oriented interface resembling the MPI-2 C++ bindings. It supports point-to-point (sends, receives) and collective (broadcasts, scatters, gathers) communication of any picklable Python object, as well as efficient communication of Python objects exposing the Python buffer interface (e.g. NumPy arrays and builtin bytes/array/memoryview objects).

INTRODUCTION

Over the last years, high performance computing has become an affordable resource to many more researchers in the scientific community than ever before. The conjunction of quality open source software and commodity hardware strongly influenced the now widespread popularity of Beowulf class clusters and cluster of workstations.

Among many parallel computational models, message-passing has proven to be an effective one. This paradigm is specially suited for (but not limited to) distributed memory architectures and is used in today’s most demanding scientific and engineering application related to modeling, simulation, design, and signal processing. However, portable message-passing parallel programming used to be a nightmare in the past because of the many incompatible options developers were faced to. Fortunately, this situation definitely changed after the MPI Forum released its standard specification.

High performance computing is traditionally associated with software development using compiled languages. However, in typical applications programs, only a small part of the code is time-critical enough to require the efficiency of compiled languages. The rest of the code is generally related to memory management, error handling, input/output, and user interaction, and those are usually the most error prone and time-consuming lines of code to write and debug in the whole development process. Interpreted high-level languages can be really advantageous for this kind of tasks.

For implementing general-purpose numerical computations, MATLAB [1] is the dominant interpreted programming language. In the open source side, Octave and Scilab are well known, freely distributed software packages providing compatibility with the MATLAB language. In this work, we present MPI for Python, a new package enabling applications to exploit multiple processors using standard MPI “look and feel” in Python scripts.

[1]
MATLAB is a registered trademark of The MathWorks, Inc.

What is MPI?

MPI, [mpi-using] [mpi-ref] the Message Passing Interface, is a standardized and portable message-passing system designed to function on a wide variety of parallel computers. The standard defines the syntax and semantics of library routines and allows users to write portable programs in the main scientific programming languages (Fortran, C, or C++).

Since its release, the MPI specification [mpi-std1] [mpi-std2] has become the leading standard for message-passing libraries for parallel computers. Implementations are available from vendors of high-performance computers and from well known open source projects like MPICH [mpi-mpich] and Open MPI [mpi-openmpi].

What is Python?

Python is a modern, easy to learn, powerful programming language. It has efficient high-level data structures and a simple but effective approach to object-oriented programming with dynamic typing and dynamic binding. It supports modules and packages, which encourages program modularity and code reuse. Python’s elegant syntax, together with its interpreted nature, make it an ideal language for scripting and rapid application development in many areas on most platforms.

The Python interpreter and the extensive standard library are available in source or binary form without charge for all major platforms, and can be freely distributed. It is easily extended with new functions and data types implemented in C or C++. Python is also suitable as an extension language for customizable applications.

Python is an ideal candidate for writing the higher-level parts of large-scale scientific applications [Hinsen97] and driving simulations in parallel architectures [Beazley97] like clusters of PC’s or SMP’s. Python codes are quickly developed, easily maintained, and can achieve a high degree of integration with other libraries written in compiled languages.

As this work started and evolved, some ideas were borrowed from well known MPI and Python related open source projects from the Internet.

OOMPI
  • It has no relation with Python, but is an excellent object oriented approach to MPI.
  • It is a C++ class library specification layered on top of the C bindings that encapsulates MPI into a functional class hierarchy.
  • It provides a flexible and intuitive interface by adding some abstractions, like Ports and Messages, which enrich and simplify the syntax.

Pypar
  • Its interface is rather minimal. There is no support for communicators or process topologies.
  • It does not require the Python interpreter to be modified or recompiled, but does not permit interactive parallel runs.
  • General (picklable) Python objects of any type can be communicated. There is good support for numeric arrays, practically full MPI bandwidth can be achieved.

pyMPI
  • It rebuilds the Python interpreter providing a built-in module for message passing. It does permit interactive parallel runs, which are useful for learning and debugging.
  • It provides an interface suitable for basic parallel programming. There is not full support for defining new communicators or process topologies.
  • General (picklable) Python objects can be messaged between processors. There is not support for numeric arrays.

Scientific Python
  • It provides a collection of Python modules that are useful for scientific computing.
  • There is an interface to MPI and BSP (Bulk Synchronous Parallel programming).
  • The interface is simple but incomplete and does not resemble the MPI specification. There is support for numeric arrays.


Additionally, we would like to mention some available tools for scientific computing and software development with Python.

  • NumPy is a package that provides array manipulation and computational capabilities similar to those found in IDL, MATLAB, or Octave. Using NumPy, it is possible to write many efficient numerical data processing applications directly in Python without using any C, C++ or Fortran code.
  • SciPy is an open source library of scientific tools for Python, gathering a variety of high level science and engineering modules together as a single package. It includes modules for graphics and plotting, optimization, integration, special functions, signal and image processing, genetic algorithms, ODE solvers, and others.
  • Cython is a language that makes writing C extensions for the Python language as easy as Python itself. The Cython language is very close to the Python language, but Cython additionally supports calling C functions and declaring C types on variables and class attributes. This allows the compiler to generate very efficient C code from Cython code. This makes Cython the ideal language for wrapping for external C libraries, and for fast C modules that speed up the execution of Python code.
  • SWIG is a software development tool that connects programs written in C and C++ with a variety of high-level programming languages like Perl, Tcl/Tk, Ruby and Python. Issuing header files to SWIG is the simplest approach to interfacing C/C++ libraries from a Python module.

[mpi-std1]
MPI Forum. MPI: A Message Passing Interface Standard. International Journal of Supercomputer Applications, volume 8, number 3-4, pages 159-416, 1994.
[mpi-std2]
MPI Forum. MPI: A Message Passing Interface Standard. High Performance Computing Applications, volume 12, number 1-2, pages 1-299, 1998.
[mpi-using]
William Gropp, Ewing Lusk, and Anthony Skjellum. Using MPI: portable parallel programming with the message-passing interface. MIT Press, 1994.
[mpi-ref]
Mark Snir, Steve Otto, Steven Huss-Lederman, David Walker, and Jack Dongarra. MPI - The Complete Reference, volume 1, The MPI Core. MIT Press, 2nd. edition, 1998.
[mpi-mpich]
W. Gropp, E. Lusk, N. Doss, and A. Skjellum. A high-performance, portable implementation of the MPI message passing interface standard. Parallel Computing, 22(6):789-828, September 1996.
[mpi-openmpi]
Edgar Gabriel, Graham E. Fagg, George Bosilca, Thara Angskun, Jack J. Dongarra, Jeffrey M. Squyres, Vishal Sahay, Prabhanjan Kambadur, Brian Barrett, Andrew Lumsdaine, Ralph H. Castain, David J. Daniel, Richard L. Graham, and Timothy S. Woodall. Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation. In Proceedings, 11th European PVM/MPI Users’ Group Meeting, Budapest, Hungary, September 2004.
[Hinsen97]
Konrad Hinsen. The Molecular Modelling Toolkit: a case study of a large scientific application in Python. In Proceedings of the 6th International Python Conference, pages 29-35, San Jose, Ca., October 1997.
[Beazley97]
David M. Beazley and Peter S. Lomdahl. Feeding a large-scale physics application to Python. In Proceedings of the 6th International Python Conference, pages 21-29, San Jose, Ca., October 1997.

OVERVIEW

MPI for Python provides an object oriented approach to message passing which grounds on the standard MPI-2 C++ bindings. The interface was designed with focus in translating MPI syntax and semantics of standard MPI-2 bindings for C++ to Python. Any user of the standard C/C++ MPI bindings should be able to use this module without need of learning a new interface.

Communicating Python Objects and Array Data

The Python standard library supports different mechanisms for data persistence. Many of them rely on disk storage, but pickling and marshaling can also work with memory buffers.

The pickle modules provide user-extensible facilities to serialize general Python objects using ASCII or binary formats. The marshal module provides facilities to serialize built-in Python objects using a binary format specific to Python, but independent of machine architecture issues.

MPI for Python can communicate any built-in or user-defined Python object taking advantage of the features provided by the pickle module. These facilities will be routinely used to build binary representations of objects to communicate (at sending processes), and restoring them back (at receiving processes).

Although simple and general, the serialization approach (i.e., pickling and unpickling) previously discussed imposes important overheads in memory as well as processor usage, especially in the scenario of objects with large memory footprints being communicated. Pickling general Python objects, ranging from primitive or container built-in types to user-defined classes, necessarily requires computer resources. Processing is also needed for dispatching the appropriate serialization method (that depends on the type of the object) and doing the actual packing. Additional memory is always needed, and if its total amount is not known a priori, many reallocations can occur. Indeed, in the case of large numeric arrays, this is certainly unacceptable and precludes communication of objects occupying half or more of the available memory resources.

MPI for Python supports direct communication of any object exporting the single-segment buffer interface. This interface is a standard Python mechanism provided by some types (e.g., strings and numeric arrays), allowing access in the C side to a contiguous memory buffer (i.e., address and length) containing the relevant data. This feature, in conjunction with the capability of constructing user-defined MPI datatypes describing complicated memory layouts, enables the implementation of many algorithms involving multidimensional numeric arrays (e.g., image processing, fast Fourier transforms, finite difference schemes on structured Cartesian grids) directly in Python, with negligible overhead, and almost as fast as compiled Fortran, C, or C++ codes.

Communicators

In MPI for Python, Comm is the base class of communicators. The Intracomm and Intercomm classes are subclasses of the Comm class. The Comm.Is_inter method (and Comm.Is_intra, provided for convenience but not part of the MPI specification) is defined for communicator objects and can be used to determine the particular communicator class.

The two predefined intracommunicator instances are available: COMM_SELF and COMM_WORLD. From them, new communicators can be created as needed.

The number of processes in a communicator and the calling process rank can be respectively obtained with methods Comm.Get_size and Comm.Get_rank. The associated process group can be retrieved from a communicator by calling the Comm.Get_group method, which returns an instance of the Group class. Set operations with Group objects like like Group.Union, Group.Intersection and Group.Difference are fully supported, as well as the creation of new communicators from these groups using Comm.Create and Intracomm.Create_group.

New communicator instances can be obtained with the Comm.Clone, Comm.Dup and Comm.Split methods, as well methods Intracomm.Create_intercomm and Intercomm.Merge.

Virtual topologies (Cartcomm, Graphcomm and Distgraphcomm classes, which are specializations of the Intracomm class) are fully supported. New instances can be obtained from intracommunicator instances with factory methods Intracomm.Create_cart and Intracomm.Create_graph.

Point-to-Point Communications

Point to point communication is a fundamental capability of message passing systems. This mechanism enables the transmission of data between a pair of processes, one side sending, the other receiving.

MPI provides a set of send and receive functions allowing the communication of typed data with an associated tag. The type information enables the conversion of data representation from one architecture to another in the case of heterogeneous computing environments; additionally, it allows the representation of non-contiguous data layouts and user-defined datatypes, thus avoiding the overhead of (otherwise unavoidable) packing/unpacking operations. The tag information allows selectivity of messages at the receiving end.

Blocking Communications

MPI provides basic send and receive functions that are blocking. These functions block the caller until the data buffers involved in the communication can be safely reused by the application program.

In MPI for Python, the Comm.Send, Comm.Recv and Comm.Sendrecv methods of communicator objects provide support for blocking point-to-point communications within Intracomm and Intercomm instances. These methods can communicate memory buffers. The variants Comm.send, Comm.recv and Comm.sendrecv can communicate general Python objects.

Nonblocking Communications

On many systems, performance can be significantly increased by overlapping communication and computation. This is particularly true on systems where communication can be executed autonomously by an intelligent, dedicated communication controller.

MPI provides nonblocking send and receive functions. They allow the possible overlap of communication and computation. Non-blocking communication always come in two parts: posting functions, which begin the requested operation; and test-for-completion functions, which allow to discover whether the requested operation has completed.

In MPI for Python, the Comm.Isend and Comm.Irecv methods initiate send and receive operations, respectively. These methods return a Request instance, uniquely identifying the started operation. Its completion can be managed using the Request.Test, Request.Wait and Request.Cancel methods. The management of Request objects and associated memory buffers involved in communication requires a careful, rather low-level coordination. Users must ensure that objects exposing their memory buffers are not accessed at the Python level while they are involved in nonblocking message-passing operations.

Persistent Communications

Often a communication with the same argument list is repeatedly executed within an inner loop. In such cases, communication can be further optimized by using persistent communication, a particular case of nonblocking communication allowing the reduction of the overhead between processes and communication controllers. Furthermore , this kind of optimization can also alleviate the extra call overheads associated to interpreted, dynamic languages like Python.

In MPI for Python, the Comm.Send_init and Comm.Recv_init methods create persistent requests for a send and receive operation, respectively. These methods return an instance of the Prequest class, a subclass of the Request class. The actual communication can be effectively started using the Prequest.Start method, and its completion can be managed as previously described.

Collective Communications

Collective communications allow the transmittal of data between multiple processes of a group simultaneously. The syntax and semantics of collective functions is consistent with point-to-point communication. Collective functions communicate typed data, but messages are not paired with an associated tag; selectivity of messages is implied in the calling order. Additionally, collective functions come in blocking versions only.

The more commonly used collective communication operations are the following.

  • Barrier synchronization across all group members.
  • Global communication functions
  • Broadcast data from one member to all members of a group.
  • Gather data from all members to one member of a group.
  • Scatter data from one member to all members of a group.

Global reduction operations such as sum, maximum, minimum, etc.

In MPI for Python, the Comm.Bcast, Comm.Scatter, Comm.Gather, Comm.Allgather, Comm.Alltoall methods provide support for collective communications of memory buffers. The lower-case variants Comm.bcast, Comm.scatter, Comm.gather, Comm.allgather and Comm.alltoall can communicate general Python objects. The vector variants (which can communicate different amounts of data to each process) Comm.Scatterv, Comm.Gatherv, Comm.Allgatherv, Comm.Alltoallv and Comm.Alltoallw are also supported, they can only communicate objects exposing memory buffers.

Global reduction operations on memory buffers are accessible through the Comm.Reduce, Comm.Reduce_scatter, Comm.Allreduce, Intracomm.Scan and Intracomm.Exscan methods. The lower-case variants Comm.reduce, Comm.allreduce, Intracomm.scan and Intracomm.exscan can communicate general Python objects; however, the actual required reduction computations are performed sequentially at some process. All the predefined (i.e., SUM, PROD, MAX, etc.) reduction operations can be applied.

Support for GPU-aware MPI

Several MPI implementations, including Open MPI and MVAPICH, support passing GPU pointers to MPI calls to avoid explicit data movement between host and device. On the Python side, support for handling GPU arrays have been implemented in many libraries related GPU computation such as CuPy, Numba, PyTorch, and PyArrow. To maximize interoperability across library boundaries, two kinds of zero-copy data exchange protocols have been defined and agreed upon: DLPack and CUDA Array Interface (CAI).

MPI for Python provides an experimental support for GPU-aware MPI. This feature requires:

1.
mpi4py is built against a GPU-aware MPI library.
2.
The Python GPU arrays are compliant with either of the protocols.

See the Tutorial section for further information. We note that

  • Whether or not a MPI call can work for GPU arrays depends on the underlying MPI implementation, not on mpi4py.
  • This support is currently experimental and subject to change in the future.

Dynamic Process Management

In the context of the MPI-1 specification, a parallel application is static; that is, no processes can be added to or deleted from a running application after it has been started. Fortunately, this limitation was addressed in MPI-2. The new specification added a process management model providing a basic interface between an application and external resources and process managers.

This MPI-2 extension can be really useful, especially for sequential applications built on top of parallel modules, or parallel applications with a client/server model. The MPI-2 process model provides a mechanism to create new processes and establish communication between them and the existing MPI application. It also provides mechanisms to establish communication between two existing MPI applications, even when one did not start the other.

In MPI for Python, new independent process groups can be created by calling the Intracomm.Spawn method within an intracommunicator. This call returns a new intercommunicator (i.e., an Intercomm instance) at the parent process group. The child process group can retrieve the matching intercommunicator by calling the Comm.Get_parent class method. At each side, the new intercommunicator can be used to perform point to point and collective communications between the parent and child groups of processes.

Alternatively, disjoint groups of processes can establish communication using a client/server approach. Any server application must first call the Open_port function to open a port and the Publish_name function to publish a provided service, and next call the Intracomm.Accept method. Any client applications can first find a published service by calling the Lookup_name function, which returns the port where a server can be contacted; and next call the Intracomm.Connect method. Both Intracomm.Accept and Intracomm.Connect methods return an Intercomm instance. When connection between client/server processes is no longer needed, all of them must cooperatively call the Comm.Disconnect method. Additionally, server applications should release resources by calling the Unpublish_name and Close_port functions.

One-Sided Communications

One-sided communications (also called Remote Memory Access, RMA) supplements the traditional two-sided, send/receive based MPI communication model with a one-sided, put/get based interface. One-sided communication that can take advantage of the capabilities of highly specialized network hardware. Additionally, this extension lowers latency and software overhead in applications written using a shared-memory-like paradigm.

The MPI specification revolves around the use of objects called windows; they intuitively specify regions of a process’s memory that have been made available for remote read and write operations. The published memory blocks can be accessed through three functions for put (remote send), get (remote write), and accumulate (remote update or reduction) data items. A much larger number of functions support different synchronization styles; the semantics of these synchronization operations are fairly complex.

In MPI for Python, one-sided operations are available by using instances of the Win class. New window objects are created by calling the Win.Create method at all processes within a communicator and specifying a memory buffer . When a window instance is no longer needed, the Win.Free method should be called.

The three one-sided MPI operations for remote write, read and reduction are available through calling the methods Win.Put, Win.Get, and Win.Accumulate respectively within a Win instance. These methods need an integer rank identifying the target process and an integer offset relative the base address of the remote memory block being accessed.

The one-sided operations read, write, and reduction are implicitly nonblocking, and must be synchronized by using two primary modes. Active target synchronization requires the origin process to call the Win.Start and Win.Complete methods at the origin process, and target process cooperates by calling the Win.Post and Win.Wait methods. There is also a collective variant provided by the Win.Fence method. Passive target synchronization is more lenient, only the origin process calls the Win.Lock and Win.Unlock methods. Locks are used to protect remote accesses to the locked remote window and to protect local load/store accesses to a locked local window.

Parallel Input/Output

The POSIX standard provides a model of a widely portable file system. However, the optimization needed for parallel input/output cannot be achieved with this generic interface. In order to ensure efficiency and scalability, the underlying parallel input/output system must provide a high-level interface supporting partitioning of file data among processes and a collective interface supporting complete transfers of global data structures between process memories and files. Additionally, further efficiencies can be gained via support for asynchronous input/output, strided accesses to data, and control over physical file layout on storage devices. This scenario motivated the inclusion in the MPI-2 standard of a custom interface in order to support more elaborated parallel input/output operations.

The MPI specification for parallel input/output revolves around the use objects called files. As defined by MPI, files are not just contiguous byte streams. Instead, they are regarded as ordered collections of typed data items. MPI supports sequential or random access to any integral set of these items. Furthermore, files are opened collectively by a group of processes.

The common patterns for accessing a shared file (broadcast, scatter, gather, reduction) is expressed by using user-defined datatypes. Compared to the communication patterns of point-to-point and collective communications, this approach has the advantage of added flexibility and expressiveness. Data access operations (read and write) are defined for different kinds of positioning (using explicit offsets, individual file pointers, and shared file pointers), coordination (non-collective and collective), and synchronism (blocking, nonblocking, and split collective with begin/end phases).

In MPI for Python, all MPI input/output operations are performed through instances of the File class. File handles are obtained by calling the File.Open method at all processes within a communicator and providing a file name and the intended access mode. After use, they must be closed by calling the File.Close method. Files even can be deleted by calling method File.Delete.

After creation, files are typically associated with a per-process view. The view defines the current set of data visible and accessible from an open file as an ordered set of elementary datatypes. This data layout can be set and queried with the File.Set_view and File.Get_view methods respectively.

Actual input/output operations are achieved by many methods combining read and write calls with different behavior regarding positioning, coordination, and synchronism. Summing up, MPI for Python provides the thirty (30) methods defined in MPI-2 for reading from or writing to files using explicit offsets or file pointers (individual or shared), in blocking or nonblocking and collective or noncollective versions.

Environmental Management

Initialization and Exit

Module functions Init or Init_thread and Finalize provide MPI initialization and finalization respectively. Module functions Is_initialized and Is_finalized provide the respective tests for initialization and finalization.

NOTE:

MPI_Init() or MPI_Init_thread() is actually called when you import the MPI module from the mpi4py package, but only if MPI is not already initialized. In such case, calling Init or Init_thread from Python is expected to generate an MPI error, and in turn an exception will be raised.


NOTE:

MPI_Finalize() is registered (by using Python C/API function Py_AtExit()) for being automatically called when Python processes exit, but only if mpi4py actually initialized MPI. Therefore, there is no need to call Finalize from Python to ensure MPI finalization.


Implementation Information

  • The MPI version number can be retrieved from module function Get_version. It returns a two-integer tuple (version, subversion).
  • The Get_processor_name function can be used to access the processor name.
  • The values of predefined attributes attached to the world communicator can be obtained by calling the Comm.Get_attr method within the COMM_WORLD instance.

Timers

MPI timer functionalities are available through the Wtime and Wtick functions.

Error Handling

In order to facilitate handle sharing with other Python modules interfacing MPI-based parallel libraries, the predefined MPI error handlers ERRORS_RETURN and ERRORS_ARE_FATAL can be assigned to and retrieved from communicators using methods Comm.Set_errhandler and Comm.Get_errhandler, and similarly for windows and files. New custom error handlers can be created with Comm.Create_errhandler.

When the predefined error handler ERRORS_RETURN is set, errors returned from MPI calls within Python code will raise an instance of the exception class Exception, which is a subclass of the standard Python exception RuntimeError.

NOTE:

After import, mpi4py overrides the default MPI rules governing inheritance of error handlers. The ERRORS_RETURN error handler is set in the predefined COMM_SELF and COMM_WORLD communicators, as well as any new Comm, Win, or File instance created through mpi4py. If you ever pass such handles to C/C++/Fortran library code, it is recommended to set the ERRORS_ARE_FATAL error handler on them to ensure MPI errors do not pass silently.


WARNING:

Importing with from mpi4py.MPI import * will cause a name clashing with the standard Python Exception base class.


TUTORIAL

WARNING:

Under construction. Contributions very welcome!


TIP:

Rolf Rabenseifner at HLRS developed a comprehensive MPI-3.1/4.0 course with slides and a large set of exercises including solutions. This material is available online for self-study. The slides and exercises show the C, Fortran, and Python (mpi4py) interfaces. For performance reasons, most Python exercises use NumPy arrays and communication routines involving buffer-like objects.


TIP:

Victor Eijkhout at TACC authored the book Parallel Programming for Science and Engineering. This book is available online in PDF and HTML formats. The book covers parallel programming with MPI and OpenMP in C/C++ and Fortran, and MPI in Python using mpi4py.


MPI for Python supports convenient, pickle-based communication of generic Python object as well as fast, near C-speed, direct array data communication of buffer-provider objects (e.g., NumPy arrays).

Communication of generic Python objects

You have to use methods with all-lowercase names, like Comm.send, Comm.recv, Comm.bcast, Comm.scatter, Comm.gather . An object to be sent is passed as a parameter to the communication call, and the received object is simply the return value.

The Comm.isend and Comm.irecv methods return Request instances; completion of these methods can be managed using the Request.test and Request.wait methods.

The Comm.recv and Comm.irecv methods may be passed a buffer object that can be repeatedly used to receive messages avoiding internal memory allocation. This buffer must be sufficiently large to accommodate the transmitted messages; hence, any buffer passed to Comm.recv or Comm.irecv must be at least as long as the pickled data transmitted to the receiver.

Collective calls like Comm.scatter, Comm.gather, Comm.allgather, Comm.alltoall expect a single value or a sequence of Comm.size elements at the root or all process. They return a single value, a list of Comm.size elements, or None.

NOTE:

MPI for Python uses the highest protocol version available in the Python runtime (see the HIGHEST_PROTOCOL constant in the pickle module). The default protocol can be changed at import time by setting the MPI4PY_PICKLE_PROTOCOL environment variable, or at runtime by assigning a different value to the PROTOCOL attribute of the pickle object within the MPI module.


Communication of buffer-like objects

You have to use method names starting with an upper-case letter, like Comm.Send, Comm.Recv, Comm.Bcast, Comm.Scatter, Comm.Gather.

In general, buffer arguments to these calls must be explicitly specified by using a 2/3-list/tuple like [data, MPI.DOUBLE], or [data, count, MPI.DOUBLE] (the former one uses the byte-size of data and the extent of the MPI datatype to define count).

For vector collectives communication operations like Comm.Scatterv and Comm.Gatherv, buffer arguments are specified as [data, count, displ, datatype], where count and displ are sequences of integral values.

Automatic MPI datatype discovery for NumPy/GPU arrays and PEP-3118 buffers is supported, but limited to basic C types (all C/C99-native signed/unsigned integral types and single/double precision real/complex floating types) and availability of matching datatypes in the underlying MPI implementation. In this case, the buffer-provider object can be passed directly as a buffer argument, the count and MPI datatype will be inferred.

If mpi4py is built against a GPU-aware MPI implementation, GPU arrays can be passed to upper-case methods as long as they have either the __dlpack__ and __dlpack_device__ methods or the __cuda_array_interface__ attribute that are compliant with the respective standard specifications. Moreover, only C-contiguous or Fortran-contiguous GPU arrays are supported. It is important to note that GPU buffers must be fully ready before any MPI routines operate on them to avoid race conditions. This can be ensured by using the synchronization API of your array library. mpi4py does not have access to any GPU-specific functionality and thus cannot perform this operation automatically for users.


Running Python scripts with MPI

Most MPI programs can be run with the command mpiexec. In practice, running Python programs looks like:

$ mpiexec -n 4 python script.py


to run the program with 4 processors.

Point-to-Point Communication

Python objects (pickle under the hood):

from mpi4py import MPI
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
if rank == 0:

data = {'a': 7, 'b': 3.14}
comm.send(data, dest=1, tag=11) elif rank == 1:
data = comm.recv(source=0, tag=11)


Python objects with non-blocking communication:

from mpi4py import MPI
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
if rank == 0:

data = {'a': 7, 'b': 3.14}
req = comm.isend(data, dest=1, tag=11)
req.wait() elif rank == 1:
req = comm.irecv(source=0, tag=11)
data = req.wait()


NumPy arrays (the fast way!):

from mpi4py import MPI
import numpy
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
# passing MPI datatypes explicitly
if rank == 0:

data = numpy.arange(1000, dtype='i')
comm.Send([data, MPI.INT], dest=1, tag=77) elif rank == 1:
data = numpy.empty(1000, dtype='i')
comm.Recv([data, MPI.INT], source=0, tag=77) # automatic MPI datatype discovery if rank == 0:
data = numpy.arange(100, dtype=numpy.float64)
comm.Send(data, dest=1, tag=13) elif rank == 1:
data = numpy.empty(100, dtype=numpy.float64)
comm.Recv(data, source=0, tag=13)



Collective Communication

Broadcasting a Python dictionary:

from mpi4py import MPI
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
if rank == 0:

data = {'key1' : [7, 2.72, 2+3j],
'key2' : ( 'abc', 'xyz')} else:
data = None data = comm.bcast(data, root=0)


Scattering Python objects:

from mpi4py import MPI
comm = MPI.COMM_WORLD
size = comm.Get_size()
rank = comm.Get_rank()
if rank == 0:

data = [(i+1)**2 for i in range(size)] else:
data = None data = comm.scatter(data, root=0) assert data == (rank+1)**2


Gathering Python objects:

from mpi4py import MPI
comm = MPI.COMM_WORLD
size = comm.Get_size()
rank = comm.Get_rank()
data = (rank+1)**2
data = comm.gather(data, root=0)
if rank == 0:

for i in range(size):
assert data[i] == (i+1)**2 else:
assert data is None


Broadcasting a NumPy array:

from mpi4py import MPI
import numpy as np
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
if rank == 0:

data = np.arange(100, dtype='i') else:
data = np.empty(100, dtype='i') comm.Bcast(data, root=0) for i in range(100):
assert data[i] == i


Scattering NumPy arrays:

from mpi4py import MPI
import numpy as np
comm = MPI.COMM_WORLD
size = comm.Get_size()
rank = comm.Get_rank()
sendbuf = None
if rank == 0:

sendbuf = np.empty([size, 100], dtype='i')
sendbuf.T[:,:] = range(size) recvbuf = np.empty(100, dtype='i') comm.Scatter(sendbuf, recvbuf, root=0) assert np.allclose(recvbuf, rank)


Gathering NumPy arrays:

from mpi4py import MPI
import numpy as np
comm = MPI.COMM_WORLD
size = comm.Get_size()
rank = comm.Get_rank()
sendbuf = np.zeros(100, dtype='i') + rank
recvbuf = None
if rank == 0:

recvbuf = np.empty([size, 100], dtype='i') comm.Gather(sendbuf, recvbuf, root=0) if rank == 0:
for i in range(size):
assert np.allclose(recvbuf[i,:], i)


Parallel matrix-vector product:

from mpi4py import MPI
import numpy
def matvec(comm, A, x):

m = A.shape[0] # local rows
p = comm.Get_size()
xg = numpy.zeros(m*p, dtype='d')
comm.Allgather([x, MPI.DOUBLE],
[xg, MPI.DOUBLE])
y = numpy.dot(A, xg)
return y



Input/Output (MPI-IO)

Collective I/O with NumPy arrays:

from mpi4py import MPI
import numpy as np
amode = MPI.MODE_WRONLY|MPI.MODE_CREATE
comm = MPI.COMM_WORLD
fh = MPI.File.Open(comm, "./datafile.contig", amode)
buffer = np.empty(10, dtype=np.int)
buffer[:] = comm.Get_rank()
offset = comm.Get_rank()*buffer.nbytes
fh.Write_at_all(offset, buffer)
fh.Close()


Non-contiguous Collective I/O with NumPy arrays and datatypes:

from mpi4py import MPI
import numpy as np
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
size = comm.Get_size()
amode = MPI.MODE_WRONLY|MPI.MODE_CREATE
fh = MPI.File.Open(comm, "./datafile.noncontig", amode)
item_count = 10
buffer = np.empty(item_count, dtype='i')
buffer[:] = rank
filetype = MPI.INT.Create_vector(item_count, 1, size)
filetype.Commit()
displacement = MPI.INT.Get_size()*rank
fh.Set_view(displacement, filetype=filetype)
fh.Write_all(buffer)
filetype.Free()
fh.Close()



Dynamic Process Management

Compute Pi - Master (or parent, or client) side:

#!/usr/bin/env python
from mpi4py import MPI
import numpy
import sys
comm = MPI.COMM_SELF.Spawn(sys.executable,

args=['cpi.py'],
maxprocs=5) N = numpy.array(100, 'i') comm.Bcast([N, MPI.INT], root=MPI.ROOT) PI = numpy.array(0.0, 'd') comm.Reduce(None, [PI, MPI.DOUBLE],
op=MPI.SUM, root=MPI.ROOT) print(PI) comm.Disconnect()


Compute Pi - Worker (or child, or server) side:

#!/usr/bin/env python
from mpi4py import MPI
import numpy
comm = MPI.Comm.Get_parent()
size = comm.Get_size()
rank = comm.Get_rank()
N = numpy.array(0, dtype='i')
comm.Bcast([N, MPI.INT], root=0)
h = 1.0 / N; s = 0.0
for i in range(rank, N, size):

x = h * (i + 0.5)
s += 4.0 / (1.0 + x**2) PI = numpy.array(s * h, dtype='d') comm.Reduce([PI, MPI.DOUBLE], None,
op=MPI.SUM, root=0) comm.Disconnect()



GPU-aware MPI + Python GPU arrays

Reduce-to-all CuPy arrays:

from mpi4py import MPI
import cupy as cp
comm = MPI.COMM_WORLD
size = comm.Get_size()
rank = comm.Get_rank()
sendbuf = cp.arange(10, dtype='i')
recvbuf = cp.empty_like(sendbuf)
cp.cuda.get_current_stream().synchronize()
comm.Allreduce(sendbuf, recvbuf)
assert cp.allclose(recvbuf, sendbuf*size)



One-Sided Communication (RMA)

Read from (write to) the entire RMA window:

import numpy as np
from mpi4py import MPI
from mpi4py.util import dtlib
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
datatype = MPI.FLOAT
np_dtype = dtlib.to_numpy_dtype(datatype)
itemsize = datatype.Get_size()
N = 10
win_size = N * itemsize if rank == 0 else 0
win = MPI.Win.Allocate(win_size, comm=comm)
buf = np.empty(N, dtype=np_dtype)
if rank == 0:

buf.fill(42)
win.Lock(rank=0)
win.Put(buf, target_rank=0)
win.Unlock(rank=0)
comm.Barrier() else:
comm.Barrier()
win.Lock(rank=0)
win.Get(buf, target_rank=0)
win.Unlock(rank=0)
assert np.all(buf == 42)


Accessing a part of the RMA window using the target argument, which is defined as (offset, count, datatype):

import numpy as np
from mpi4py import MPI
from mpi4py.util import dtlib
comm = MPI.COMM_WORLD
rank = comm.Get_rank()
datatype = MPI.FLOAT
np_dtype = dtlib.to_numpy_dtype(datatype)
itemsize = datatype.Get_size()
N = comm.Get_size() + 1
win_size = N * itemsize if rank == 0 else 0
win = MPI.Win.Allocate(

size=win_size,
disp_unit=itemsize,
comm=comm, ) if rank == 0:
mem = np.frombuffer(win, dtype=np_dtype)
mem[:] = np.arange(len(mem), dtype=np_dtype) comm.Barrier() buf = np.zeros(3, dtype=np_dtype) target = (rank, 2, datatype) win.Lock(rank=0) win.Get(buf, target_rank=0, target=target) win.Unlock(rank=0) assert np.all(buf == [rank, rank+1, 0])



Wrapping with SWIG

C source:

/* file: helloworld.c */
void sayhello(MPI_Comm comm)
{

int size, rank;
MPI_Comm_size(comm, &size);
MPI_Comm_rank(comm, &rank);
printf("Hello, World! "
"I am process %d of %d.\n",
rank, size); }


SWIG interface file:

// file: helloworld.i
%module helloworld
%{
#include <mpi.h>
#include "helloworld.c"
}%
%include mpi4py/mpi4py.i
%mpi4py_typemap(Comm, MPI_Comm);
void sayhello(MPI_Comm comm);


Try it in the Python prompt:

>>> from mpi4py import MPI
>>> import helloworld
>>> helloworld.sayhello(MPI.COMM_WORLD)
Hello, World! I am process 0 of 1.



Wrapping with F2Py

Fortran 90 source:

! file: helloworld.f90
subroutine sayhello(comm)

use mpi
implicit none
integer :: comm, rank, size, ierr
call MPI_Comm_size(comm, size, ierr)
call MPI_Comm_rank(comm, rank, ierr)
print *, 'Hello, World! I am process ',rank,' of ',size,'.' end subroutine sayhello


Compiling example using f2py

$ f2py -c --f90exec=mpif90 helloworld.f90 -m helloworld


Try it in the Python prompt:

>>> from mpi4py import MPI
>>> import helloworld
>>> fcomm = MPI.COMM_WORLD.py2f()
>>> helloworld.sayhello(fcomm)
Hello, World! I am process 0 of 1.



MPI4PY

The MPI for Python package.

The Message Passing Interface (MPI) is a standardized and portable message-passing system designed to function on a wide variety of parallel computers. The MPI standard defines the syntax and semantics of library routines and allows users to write portable programs in the main scientific programming languages (Fortran, C, or C++). Since its release, the MPI specification has become the leading standard for message-passing libraries for parallel computers.

MPI for Python provides MPI bindings for the Python programming language, allowing any Python program to exploit multiple processors. This package build on the MPI specification and provides an object oriented interface which closely follows MPI-2 C++ bindings.

Runtime configuration options

This object has attributes exposing runtime configuration options that become effective at import time of the MPI module.

Attributes Summary

initialize Automatic MPI initialization at import
threads Request initialization with thread support
thread_level Level of thread support to request
finalize Automatic MPI finalization at exit
fast_reduce Use tree-based reductions for objects
recv_mprobe Use matched probes to receive objects
irecv_bufsz Default buffer size in bytes for irecv()
errors Error handling policy

Attributes Documentation

Automatic MPI initialization at import.
bool
True

SEE ALSO:

MPI4PY_RC_INITIALIZE



Request initialization with thread support.
bool
True

SEE ALSO:

MPI4PY_RC_THREADS



Level of thread support to request.
str
"multiple"
"multiple", "serialized", "funneled", "single"

SEE ALSO:

MPI4PY_RC_THREAD_LEVEL



Automatic MPI finalization at exit.
None or bool
None

SEE ALSO:

MPI4PY_RC_FINALIZE



Use tree-based reductions for objects.
bool
True

SEE ALSO:

MPI4PY_RC_FAST_REDUCE



Use matched probes to receive objects.
bool
True

SEE ALSO:

MPI4PY_RC_RECV_MPROBE



Default buffer size in bytes for irecv().
int
32768

SEE ALSO:

MPI4PY_RC_IRECV_BUFSZ


Added in version 4.0.0.


Error handling policy.
str
"exception"
"exception", "default", "abort", "fatal"

SEE ALSO:

MPI4PY_RC_ERRORS



Example

MPI for Python features automatic initialization and finalization of the MPI execution environment. By using the mpi4py.rc object, MPI initialization and finalization can be handled programmatically:

import mpi4py
mpi4py.rc.initialize = False  # do not initialize MPI automatically
mpi4py.rc.finalize = False    # do not finalize MPI automatically
from mpi4py import MPI # import the 'MPI' module
MPI.Init()      # manual initialization of the MPI environment
...             # your finest code here ...
MPI.Finalize()  # manual finalization of the MPI environment


Environment variables

The following environment variables override the corresponding attributes of the mpi4py.rc and MPI.pickle objects at import time of the MPI module.

NOTE:

For variables of boolean type, accepted values are 0 and 1 (interpreted as False and True, respectively), and strings specifying a YAML boolean value (case-insensitive).


bool
True

Whether to automatically initialize MPI at import time of the mpi4py.MPI module.

SEE ALSO:

mpi4py.rc.initialize


Added in version 4.0.0.


None | bool
None
None, True, False

Whether to automatically finalize MPI at exit time of the Python process.

SEE ALSO:

mpi4py.rc.finalize


Added in version 4.0.0.


bool
True

Whether to initialize MPI with thread support.

SEE ALSO:

mpi4py.rc.threads


Added in version 3.1.0.


"multiple"
"single", "funneled", "serialized", "multiple"

The level of required thread support.

SEE ALSO:

mpi4py.rc.thread_level


Added in version 3.1.0.


bool
True

Whether to use tree-based reductions for objects.

SEE ALSO:

mpi4py.rc.fast_reduce


Added in version 3.1.0.


bool
True

Whether to use matched probes to receive objects.

SEE ALSO:

mpi4py.rc.recv_mprobe



bool
True

Default buffer size in bytes for irecv().

SEE ALSO:

mpi4py.rc.irecv_bufsz


Added in version 4.0.0.


"exception"
"exception", "default", "abort", "fatal"

Controls default MPI error handling policy.

SEE ALSO:

mpi4py.rc.errors


Added in version 3.1.0.


int
pickle.HIGHEST_PROTOCOL

Controls the default pickle protocol to use when communicating Python objects.

SEE ALSO:

PROTOCOL attribute of the MPI.pickle object within the MPI module.


Added in version 3.1.0.


int
262144

Controls the default buffer size threshold for switching from in-band to out-of-band buffer handling when using pickle protocol version 5 or higher.

SEE ALSO:

THRESHOLD attribute of the MPI.pickle object within the MPI module.


Added in version 3.1.2.


Miscellaneous functions

Support for the MPI profiling interface.
  • name (str) – Name of the profiler library to load.
  • path (sequence of str, optional) – Additional paths to search for the profiler.

None


Return the directory in the package that contains header files.

Extension modules that need to compile against mpi4py should use this function to locate the appropriate include directory. Using Python distutils (or perhaps NumPy distutils):

import mpi4py
Extension('extension_name', ...

include_dirs=[..., mpi4py.get_include()])




Return a dictionary with information about MPI.

Changed in version 4.0.0: By default, this function returns an empty dictionary. However, downstream packagers and distributors may alter such behavior. To that end, MPI information must be provided under an mpi section within a UTF-8 encoded INI-style configuration file mpi.cfg located at the top-level package directory. The configuration file is read and parsed using the configparser module.

dict[str, str]


MPI4PY.MPI

Classes

Ancillary

Datatype Datatype object.
Status Status object.
Request Request handler.
Prequest Persistent request handler.
Grequest Generalized request handler.
Op Reduction operation.
Group Group of processes.
Info Info object.
Session Session context.

Communication

Comm Communication context.
Intracomm Intracommunicator.
Topocomm Topology intracommunicator.
Cartcomm Cartesian topology intracommunicator.
Graphcomm General graph topology intracommunicator.
Distgraphcomm Distributed graph topology intracommunicator.
Intercomm Intercommunicator.
Message Matched message.

One-sided operations

Win Remote memory access context.

Input/Output

File File I/O context.

Error handling

Errhandler Error handler.
Exception Exception class.

Auxiliary

Pickle Pickle/unpickle Python objects.
buffer Buffer.

Functions

Version inquiry

Get_version() Obtain the version number of the MPI standard.
Get_library_version() Obtain the version string of the MPI library.

Initialization and finalization

Init() Initialize the MPI execution environment.
Init_thread([required]) Initialize the MPI execution environment.
Finalize() Terminate the MPI execution environment.
Is_initialized() Indicate whether Init has been called.
Is_finalized() Indicate whether Finalize has completed.
Query_thread() Return the level of thread support provided by the MPI library.
Is_thread_main() Indicate whether this thread called Init or Init_thread.

Memory allocation

Alloc_mem(size[, info]) Allocate memory for message passing and remote memory access.
Free_mem(mem) Free memory allocated with Alloc_mem.

Address manipulation

Get_address(location) Get the address of a location in memory.
Aint_add(base, disp) Return the sum of base address and displacement.
Aint_diff(addr1, addr2) Return the difference between absolute addresses.

Timer

Wtick() Return the resolution of Wtime.
Wtime() Return an elapsed time on the calling processor.

Error handling

Get_error_class(errorcode) Convert an error code into an error class.
Get_error_string(errorcode) Return the error string for a given error class or error code.
Add_error_class() Add an error class to the known error classes.
Add_error_code(errorclass) Add an error code to an error class.
Add_error_string(errorcode, string) Associate an error string with an error class or error code.
Remove_error_class(errorclass) Remove an error class from the known error classes.
Remove_error_code(errorcode) Remove an error code from the known error codes.
Remove_error_string(errorcode) Remove error string association from error class or error code.

Dynamic process management

Open_port([info]) Return an address used to connect group of processes.
Close_port(port_name) Close a port.
Publish_name(service_name, port_name[, info]) Publish a service name.
Unpublish_name(service_name, port_name[, info]) Unpublish a service name.
Lookup_name(service_name[, info]) Lookup a port name given a service name.

Miscellanea

Attach_buffer(buf) Attach a user-provided buffer for sending in buffered mode.
Detach_buffer() Remove an existing attached buffer.
Flush_buffer() Block until all buffered messages have been transmitted.
Iflush_buffer() Nonblocking flush for buffered messages.
Compute_dims(nnodes, dims) Return a balanced distribution of processes per coordinate direction.
Get_processor_name() Obtain the name of the calling processor.
Register_datarep(datarep, read_fn, write_fn, ...) Register user-defined data representations.
Pcontrol(level) Control profiling.

Utilities

get_vendor() Information about the underlying MPI implementation.

Attributes

UNDEFINED Constant UNDEFINED of type int
ANY_SOURCE Constant ANY_SOURCE of type int
ANY_TAG Constant ANY_TAG of type int
PROC_NULL Constant PROC_NULL of type int
ROOT Constant ROOT of type int
BOTTOM Constant BOTTOM of type BottomType
IN_PLACE Constant IN_PLACE of type InPlaceType
BUFFER_AUTOMATIC Constant BUFFER_AUTOMATIC of type BufferAutomaticType
KEYVAL_INVALID Constant KEYVAL_INVALID of type int
TAG_UB Constant TAG_UB of type int
IO Constant IO of type int
WTIME_IS_GLOBAL Constant WTIME_IS_GLOBAL of type int
UNIVERSE_SIZE Constant UNIVERSE_SIZE of type int
APPNUM Constant APPNUM of type int
LASTUSEDCODE Constant LASTUSEDCODE of type int
WIN_BASE Constant WIN_BASE of type int
WIN_SIZE Constant WIN_SIZE of type int
WIN_DISP_UNIT Constant WIN_DISP_UNIT of type int
WIN_CREATE_FLAVOR Constant WIN_CREATE_FLAVOR of type int
WIN_FLAVOR Constant WIN_FLAVOR of type int
WIN_MODEL Constant WIN_MODEL of type int
SUCCESS Constant SUCCESS of type int
ERR_LASTCODE Constant ERR_LASTCODE of type int
ERR_COMM Constant ERR_COMM of type int
ERR_GROUP Constant ERR_GROUP of type int
ERR_TYPE Constant ERR_TYPE of type int
ERR_REQUEST Constant ERR_REQUEST of type int
ERR_OP Constant ERR_OP of type int
ERR_ERRHANDLER Constant ERR_ERRHANDLER of type int
ERR_BUFFER Constant ERR_BUFFER of type int
ERR_COUNT Constant ERR_COUNT of type int
ERR_TAG Constant ERR_TAG of type int
ERR_RANK Constant ERR_RANK of type int
ERR_ROOT Constant ERR_ROOT of type int
ERR_TRUNCATE Constant ERR_TRUNCATE of type int
ERR_IN_STATUS Constant ERR_IN_STATUS of type int
ERR_PENDING Constant ERR_PENDING of type int
ERR_TOPOLOGY Constant ERR_TOPOLOGY of type int
ERR_DIMS Constant ERR_DIMS of type int
ERR_ARG Constant ERR_ARG of type int
ERR_OTHER Constant ERR_OTHER of type int
ERR_UNKNOWN Constant ERR_UNKNOWN of type int
ERR_INTERN Constant ERR_INTERN of type int
ERR_INFO Constant ERR_INFO of type int
ERR_FILE Constant ERR_FILE of type int
ERR_WIN Constant ERR_WIN of type int
ERR_KEYVAL Constant ERR_KEYVAL of type int
ERR_INFO_KEY Constant ERR_INFO_KEY of type int
ERR_INFO_VALUE Constant ERR_INFO_VALUE of type int
ERR_INFO_NOKEY Constant ERR_INFO_NOKEY of type int
ERR_ACCESS Constant ERR_ACCESS of type int
ERR_AMODE Constant ERR_AMODE of type int
ERR_BAD_FILE Constant ERR_BAD_FILE of type int
ERR_FILE_EXISTS Constant ERR_FILE_EXISTS of type int
ERR_FILE_IN_USE Constant ERR_FILE_IN_USE of type int
ERR_NO_SPACE Constant ERR_NO_SPACE of type int
ERR_NO_SUCH_FILE Constant ERR_NO_SUCH_FILE of type int
ERR_IO Constant ERR_IO of type int
ERR_READ_ONLY Constant ERR_READ_ONLY of type int
ERR_CONVERSION Constant ERR_CONVERSION of type int
ERR_DUP_DATAREP Constant ERR_DUP_DATAREP of type int
ERR_UNSUPPORTED_DATAREP Constant ERR_UNSUPPORTED_DATAREP of type int
ERR_UNSUPPORTED_OPERATION Constant ERR_UNSUPPORTED_OPERATION of type int
ERR_NAME Constant ERR_NAME of type int
ERR_NO_MEM Constant ERR_NO_MEM of type int
ERR_NOT_SAME Constant ERR_NOT_SAME of type int
ERR_PORT Constant ERR_PORT of type int
ERR_QUOTA Constant ERR_QUOTA of type int
ERR_SERVICE Constant ERR_SERVICE of type int
ERR_SPAWN Constant ERR_SPAWN of type int
ERR_BASE Constant ERR_BASE of type int
ERR_SIZE Constant ERR_SIZE of type int
ERR_DISP Constant ERR_DISP of type int
ERR_ASSERT Constant ERR_ASSERT of type int
ERR_LOCKTYPE Constant ERR_LOCKTYPE of type int
ERR_RMA_CONFLICT Constant ERR_RMA_CONFLICT of type int
ERR_RMA_SYNC Constant ERR_RMA_SYNC of type int
ERR_RMA_RANGE Constant ERR_RMA_RANGE of type int
ERR_RMA_ATTACH Constant ERR_RMA_ATTACH of type int
ERR_RMA_SHARED Constant ERR_RMA_SHARED of type int
ERR_RMA_FLAVOR Constant ERR_RMA_FLAVOR of type int
ORDER_C Constant ORDER_C of type int
ORDER_F Constant ORDER_F of type int
ORDER_FORTRAN Constant ORDER_FORTRAN of type int
TYPECLASS_INTEGER Constant TYPECLASS_INTEGER of type int
TYPECLASS_REAL Constant TYPECLASS_REAL of type int
TYPECLASS_COMPLEX Constant TYPECLASS_COMPLEX of type int
DISTRIBUTE_NONE Constant DISTRIBUTE_NONE of type int
DISTRIBUTE_BLOCK Constant DISTRIBUTE_BLOCK of type int
DISTRIBUTE_CYCLIC Constant DISTRIBUTE_CYCLIC of type int
DISTRIBUTE_DFLT_DARG Constant DISTRIBUTE_DFLT_DARG of type int
COMBINER_NAMED Constant COMBINER_NAMED of type int
COMBINER_DUP Constant COMBINER_DUP of type int
COMBINER_CONTIGUOUS Constant COMBINER_CONTIGUOUS of type int
COMBINER_VECTOR Constant COMBINER_VECTOR of type int
COMBINER_HVECTOR Constant COMBINER_HVECTOR of type int
COMBINER_INDEXED Constant COMBINER_INDEXED of type int
COMBINER_HINDEXED Constant COMBINER_HINDEXED of type int
COMBINER_INDEXED_BLOCK Constant COMBINER_INDEXED_BLOCK of type int
COMBINER_HINDEXED_BLOCK Constant COMBINER_HINDEXED_BLOCK of type int
COMBINER_STRUCT Constant COMBINER_STRUCT of type int
COMBINER_SUBARRAY Constant COMBINER_SUBARRAY of type int
COMBINER_DARRAY Constant COMBINER_DARRAY of type int
COMBINER_RESIZED Constant COMBINER_RESIZED of type int
COMBINER_VALUE_INDEX Constant COMBINER_VALUE_INDEX of type int
COMBINER_F90_REAL Constant COMBINER_F90_REAL of type int
COMBINER_F90_COMPLEX Constant COMBINER_F90_COMPLEX of type int
COMBINER_F90_INTEGER Constant COMBINER_F90_INTEGER of type int
IDENT Constant IDENT of type int
CONGRUENT Constant CONGRUENT of type int
SIMILAR Constant SIMILAR of type int
UNEQUAL Constant UNEQUAL of type int
CART Constant CART of type int
GRAPH Constant GRAPH of type int
DIST_GRAPH Constant DIST_GRAPH of type int
UNWEIGHTED Constant UNWEIGHTED of type int
WEIGHTS_EMPTY Constant WEIGHTS_EMPTY of type int
COMM_TYPE_SHARED Constant COMM_TYPE_SHARED of type int
BSEND_OVERHEAD Constant BSEND_OVERHEAD of type int
WIN_FLAVOR_CREATE Constant WIN_FLAVOR_CREATE of type int
WIN_FLAVOR_ALLOCATE Constant WIN_FLAVOR_ALLOCATE of type int
WIN_FLAVOR_DYNAMIC Constant WIN_FLAVOR_DYNAMIC of type int
WIN_FLAVOR_SHARED Constant WIN_FLAVOR_SHARED of type int
WIN_SEPARATE Constant WIN_SEPARATE of type int
WIN_UNIFIED Constant WIN_UNIFIED of type int
MODE_NOCHECK Constant MODE_NOCHECK of type int
MODE_NOSTORE Constant MODE_NOSTORE of type int
MODE_NOPUT Constant MODE_NOPUT of type int
MODE_NOPRECEDE Constant MODE_NOPRECEDE of type int
MODE_NOSUCCEED Constant MODE_NOSUCCEED of type int
LOCK_EXCLUSIVE Constant LOCK_EXCLUSIVE of type int
LOCK_SHARED Constant LOCK_SHARED of type int
MODE_RDONLY Constant MODE_RDONLY of type int
MODE_WRONLY Constant MODE_WRONLY of type int
MODE_RDWR Constant MODE_RDWR of type int
MODE_CREATE Constant MODE_CREATE of type int
MODE_EXCL Constant MODE_EXCL of type int
MODE_DELETE_ON_CLOSE Constant MODE_DELETE_ON_CLOSE of type int
MODE_UNIQUE_OPEN Constant MODE_UNIQUE_OPEN of type int
MODE_SEQUENTIAL Constant MODE_SEQUENTIAL of type int
MODE_APPEND Constant MODE_APPEND of type int
SEEK_SET Constant SEEK_SET of type int
SEEK_CUR Constant SEEK_CUR of type int
SEEK_END Constant SEEK_END of type int
DISPLACEMENT_CURRENT Constant DISPLACEMENT_CURRENT of type int
DISP_CUR Constant DISP_CUR of type int
THREAD_SINGLE Constant THREAD_SINGLE of type int
THREAD_FUNNELED Constant THREAD_FUNNELED of type int
THREAD_SERIALIZED Constant THREAD_SERIALIZED of type int
THREAD_MULTIPLE Constant THREAD_MULTIPLE of type int
VERSION Constant VERSION of type int
SUBVERSION Constant SUBVERSION of type int
MAX_PROCESSOR_NAME Constant MAX_PROCESSOR_NAME of type int
MAX_ERROR_STRING Constant MAX_ERROR_STRING of type int
MAX_PORT_NAME Constant MAX_PORT_NAME of type int
MAX_INFO_KEY Constant MAX_INFO_KEY of type int
MAX_INFO_VAL Constant MAX_INFO_VAL of type int
MAX_OBJECT_NAME Constant MAX_OBJECT_NAME of type int
MAX_DATAREP_STRING Constant MAX_DATAREP_STRING of type int
MAX_LIBRARY_VERSION_STRING Constant MAX_LIBRARY_VERSION_STRING of type int
DATATYPE_NULL Object DATATYPE_NULL of type Datatype
PACKED Object PACKED of type Datatype
BYTE Object BYTE of type Datatype
AINT Object AINT of type Datatype
OFFSET Object OFFSET of type Datatype
COUNT Object COUNT of type Datatype
CHAR Object CHAR of type Datatype
WCHAR Object WCHAR of type Datatype
SIGNED_CHAR Object SIGNED_CHAR of type Datatype
SHORT Object SHORT of type Datatype
INT Object INT of type Datatype
LONG Object LONG of type Datatype
LONG_LONG Object LONG_LONG of type Datatype
UNSIGNED_CHAR Object UNSIGNED_CHAR of type Datatype
UNSIGNED_SHORT Object UNSIGNED_SHORT of type Datatype
UNSIGNED Object UNSIGNED of type Datatype
UNSIGNED_LONG Object UNSIGNED_LONG of type Datatype
UNSIGNED_LONG_LONG Object UNSIGNED_LONG_LONG of type Datatype
FLOAT Object FLOAT of type Datatype
DOUBLE Object DOUBLE of type Datatype
LONG_DOUBLE Object LONG_DOUBLE of type Datatype
C_BOOL Object C_BOOL of type Datatype
INT8_T Object INT8_T of type Datatype
INT16_T Object INT16_T of type Datatype
INT32_T Object INT32_T of type Datatype
INT64_T Object INT64_T of type Datatype
UINT8_T Object UINT8_T of type Datatype
UINT16_T Object UINT16_T of type Datatype
UINT32_T Object UINT32_T of type Datatype
UINT64_T Object UINT64_T of type Datatype
C_COMPLEX Object C_COMPLEX of type Datatype
C_FLOAT_COMPLEX Object C_FLOAT_COMPLEX of type Datatype
C_DOUBLE_COMPLEX Object C_DOUBLE_COMPLEX of type Datatype
C_LONG_DOUBLE_COMPLEX Object C_LONG_DOUBLE_COMPLEX of type Datatype
CXX_BOOL Object CXX_BOOL of type Datatype
CXX_FLOAT_COMPLEX Object CXX_FLOAT_COMPLEX of type Datatype
CXX_DOUBLE_COMPLEX Object CXX_DOUBLE_COMPLEX of type Datatype
CXX_LONG_DOUBLE_COMPLEX Object CXX_LONG_DOUBLE_COMPLEX of type Datatype
SHORT_INT Object SHORT_INT of type Datatype
INT_INT Object INT_INT of type Datatype
TWOINT Object TWOINT of type Datatype
LONG_INT Object LONG_INT of type Datatype
FLOAT_INT Object FLOAT_INT of type Datatype
DOUBLE_INT Object DOUBLE_INT of type Datatype
LONG_DOUBLE_INT Object LONG_DOUBLE_INT of type Datatype
CHARACTER Object CHARACTER of type Datatype
LOGICAL Object LOGICAL of type Datatype
INTEGER Object INTEGER of type Datatype
REAL Object REAL of type Datatype
DOUBLE_PRECISION Object DOUBLE_PRECISION of type Datatype
COMPLEX Object COMPLEX of type Datatype
DOUBLE_COMPLEX Object DOUBLE_COMPLEX of type Datatype
LOGICAL1 Object LOGICAL1 of type Datatype
LOGICAL2 Object LOGICAL2 of type Datatype
LOGICAL4 Object LOGICAL4 of type Datatype
LOGICAL8 Object LOGICAL8 of type Datatype
INTEGER1 Object INTEGER1 of type Datatype
INTEGER2 Object INTEGER2 of type Datatype
INTEGER4 Object INTEGER4 of type Datatype
INTEGER8 Object INTEGER8 of type Datatype
INTEGER16 Object INTEGER16 of type Datatype
REAL2 Object REAL2 of type Datatype
REAL4 Object REAL4 of type Datatype
REAL8 Object REAL8 of type Datatype
REAL16 Object REAL16 of type Datatype
COMPLEX4 Object COMPLEX4 of type Datatype
COMPLEX8 Object COMPLEX8 of type Datatype
COMPLEX16 Object COMPLEX16 of type Datatype
COMPLEX32 Object COMPLEX32 of type Datatype
UNSIGNED_INT Object UNSIGNED_INT of type Datatype
SIGNED_SHORT Object SIGNED_SHORT of type Datatype
SIGNED_INT Object SIGNED_INT of type Datatype
SIGNED_LONG Object SIGNED_LONG of type Datatype
SIGNED_LONG_LONG Object SIGNED_LONG_LONG of type Datatype
BOOL Object BOOL of type Datatype
SINT8_T Object SINT8_T of type Datatype
SINT16_T Object SINT16_T of type Datatype
SINT32_T Object SINT32_T of type Datatype
SINT64_T Object SINT64_T of type Datatype
F_BOOL Object F_BOOL of type Datatype
F_INT Object F_INT of type Datatype
F_FLOAT Object F_FLOAT of type Datatype
F_DOUBLE Object F_DOUBLE of type Datatype
F_COMPLEX Object F_COMPLEX of type Datatype
F_FLOAT_COMPLEX Object F_FLOAT_COMPLEX of type Datatype
F_DOUBLE_COMPLEX Object F_DOUBLE_COMPLEX of type Datatype
REQUEST_NULL Object REQUEST_NULL of type Request
MESSAGE_NULL Object MESSAGE_NULL of type Message
MESSAGE_NO_PROC Object MESSAGE_NO_PROC of type Message
OP_NULL Object OP_NULL of type Op
MAX Object MAX of type Op
MIN Object MIN of type Op
SUM Object SUM of type Op
PROD Object PROD of type Op
LAND Object LAND of type Op
BAND Object BAND of type Op
LOR Object LOR of type Op
BOR Object BOR of type Op
LXOR Object LXOR of type Op
BXOR Object BXOR of type Op
MAXLOC Object MAXLOC of type Op
MINLOC Object MINLOC of type Op
REPLACE Object REPLACE of type Op
NO_OP Object NO_OP of type Op
GROUP_NULL Object GROUP_NULL of type Group
GROUP_EMPTY Object GROUP_EMPTY of type Group
INFO_NULL Object INFO_NULL of type Info
INFO_ENV Object INFO_ENV of type Info
ERRHANDLER_NULL Object ERRHANDLER_NULL of type Errhandler
ERRORS_RETURN Object ERRORS_RETURN of type Errhandler
ERRORS_ARE_FATAL Object ERRORS_ARE_FATAL of type Errhandler
COMM_NULL Object COMM_NULL of type Comm
COMM_SELF Object COMM_SELF of type Intracomm
COMM_WORLD Object COMM_WORLD of type Intracomm
WIN_NULL Object WIN_NULL of type Win
FILE_NULL Object FILE_NULL of type File
pickle Object pickle of type Pickle

MPI4PY.TYPING

Added in version 4.0.0.

This module provides type aliases used to add type hints to the various functions and methods within the MPI module.

SEE ALSO:

Documentation of the typing standard module.



Types Summary

SupportsBuffer Python buffer protocol.
SupportsDLPack DLPack data interchange protocol.
SupportsCAI CUDA Array Interface (CAI) protocol.
Buffer Buffer-like object.
Bottom Start of the address range.
InPlace In-place buffer argument.
Aint Address-sized integral type.
Count Integral type for counts.
Displ Integral type for displacements.
Offset Integral type for offsets.
TypeSpec Datatype specification.
BufSpec Buffer specification.
BufSpecB Buffer specification (block).
BufSpecV Buffer specification (vector).
BufSpecW Buffer specification (generalized).
TargetSpec Target specification.

Types Documentation

Python buffer protocol.

SEE ALSO:

Buffer Protocol



DLPack data interchange protocol.

SEE ALSO:

dlpack:python-spec



CUDA Array Interface (CAI) protocol.

SEE ALSO:

numba:cuda-array-interface



Buffer-like object.

alias of SupportsBuffer | SupportsDLPack | SupportsCAI


Start of the address range.

alias of BottomType | None


In-place buffer argument.

alias of InPlaceType | None


Address-sized integral type.

alias of numbers.Integral


Integral type for counts.

alias of numbers.Integral


Integral type for displacements.

alias of numbers.Integral


Integral type for offsets.

alias of numbers.Integral


Datatype specification.

alias of Datatype | str


Buffer specification.
  • Buffer
  • Tuple[Buffer, Count]
  • Tuple[Buffer, TypeSpec]
  • Tuple[Buffer, Count, TypeSpec]
  • Tuple[Bottom, Count, Datatype]

alias of SupportsBuffer | SupportsDLPack | SupportsCAI | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Integral] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Datatype | str] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Integral, Datatype | str] | Tuple[BottomType | None, Integral, Datatype] | List[Any]


Buffer specification (block).
  • Buffer
  • Tuple[Buffer, Count]
  • Tuple[Buffer, TypeSpec]
  • Tuple[Buffer, Count, TypeSpec]

alias of SupportsBuffer | SupportsDLPack | SupportsCAI | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Integral] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Datatype | str] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Integral, Datatype | str] | List[Any]


Buffer specification (vector).
  • Buffer
  • Tuple[Buffer, Sequence[Count]]
  • Tuple[Buffer, Tuple[Sequence[Count], Sequence[Displ]]]
  • Tuple[Buffer, TypeSpec]
  • Tuple[Buffer, Sequence[Count], TypeSpec]
  • Tuple[Buffer, Tuple[Sequence[Count], Sequence[Displ]], TypeSpec]
  • Tuple[Buffer, Sequence[Count], Sequence[Displ], TypeSpec]
  • Tuple[Bottom, Tuple[Sequence[Count], Sequence[Displ]], Datatype]
  • Tuple[Bottom, Sequence[Count], Sequence[Displ], Datatype]

alias of SupportsBuffer | SupportsDLPack | SupportsCAI | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Sequence[Integral]] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Tuple[Sequence[Integral], Sequence[Integral]]] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Datatype | str] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Sequence[Integral], Datatype | str] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Tuple[Sequence[Integral], Sequence[Integral]], Datatype | str] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Sequence[Integral], Sequence[Integral], Datatype | str] | Tuple[BottomType | None, Tuple[Sequence[Integral], Sequence[Integral]], Datatype] | Tuple[BottomType | None, Sequence[Integral], Sequence[Integral], Datatype] | List[Any]


Buffer specification (generalized).
  • Tuple[Buffer, Sequence[Datatype]]
  • Tuple[Buffer, Tuple[Sequence[Count], Sequence[Displ]], Sequence[Datatype]]
  • Tuple[Buffer, Sequence[Count], Sequence[Displ], Sequence[Datatype]]
  • Tuple[Bottom, Tuple[Sequence[Count], Sequence[Displ]], Sequence[Datatype]]
  • Tuple[Bottom, Sequence[Count], Sequence[Displ], Sequence[Datatype]]

alias of Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Sequence[Datatype]] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Tuple[Sequence[Integral], Sequence[Integral]], Sequence[Datatype]] | Tuple[SupportsBuffer | SupportsDLPack | SupportsCAI, Sequence[Integral], Sequence[Integral], Sequence[Datatype]] | Tuple[BottomType | None, Tuple[Sequence[Integral], Sequence[Integral]], Sequence[Datatype]] | Tuple[BottomType | None, Sequence[Integral], Sequence[Integral], Sequence[Datatype]] | List[Any]


Target specification.
  • Displ
  • Tuple[()]
  • Tuple[Displ]
  • Tuple[Displ, Count]
  • Tuple[Displ, Count, Datatype]

alias of Integral | Tuple | Tuple[Integral] | Tuple[Integral, Integral] | Tuple[Integral, Integral, Datatype | str] | List[Any]


Invariant TypeVar.

Invariant TypeVar.

Invariant TypeVar.

Invariant TypeVar.

MPI4PY.FUTURES

Added in version 3.0.0.

This package provides a high-level interface for asynchronously executing callables on a pool of worker processes using MPI for inter-process communication.

The mpi4py.futures package is based on concurrent.futures from the Python standard library. More precisely, mpi4py.futures provides the MPIPoolExecutor class as a concrete implementation of the abstract class Executor. The submit() interface schedules a callable to be executed asynchronously and returns a Future object representing the execution of the callable. Future instances can be queried for the call result or exception. Sets of Future instances can be passed to the wait() and as_completed() functions.

SEE ALSO:

Documentation of the concurrent.futures standard module.



MPIPoolExecutor

The MPIPoolExecutor class uses a pool of MPI processes to execute calls asynchronously. By performing computations in separate processes, it allows to side-step the global interpreter lock but also means that only picklable objects can be executed and returned. The __main__ module must be importable by worker processes, thus MPIPoolExecutor instances may not work in the interactive interpreter.

MPIPoolExecutor takes advantage of the dynamic process management features introduced in the MPI-2 standard. In particular, the MPI.Intracomm.Spawn method of MPI.COMM_SELF is used in the master (or parent) process to spawn new worker (or child) processes running a Python interpreter. The master process uses a separate thread (one for each MPIPoolExecutor instance) to communicate back and forth with the workers. The worker processes serve the execution of tasks in the main (and only) thread until they are signaled for completion.

NOTE:

The worker processes must import the main script in order to unpickle any callable defined in the __main__ module and submitted from the master process. Furthermore, the callables may need access to other global variables. At the worker processes, mpi4py.futures executes the main script code (using the runpy module) under the __worker__ namespace to define the __main__ module. The __main__ and __worker__ modules are added to sys.modules (both at the master and worker processes) to ensure proper pickling and unpickling.


WARNING:

During the initial import phase at the workers, the main script cannot create and use new MPIPoolExecutor instances. Otherwise, each worker would attempt to spawn a new pool of workers, leading to infinite recursion. mpi4py.futures detects such recursive attempts to spawn new workers and aborts the MPI execution environment. As the main script code is run under the __worker__ namespace, the easiest way to avoid spawn recursion is using the idiom if __name__ == '__main__': ... in the main script.


An Executor subclass that executes calls asynchronously using a pool of at most max_workers processes. If max_workers is None or not given, its value is determined from the MPI4PY_FUTURES_MAX_WORKERS environment variable if set, or the MPI universe size if set, otherwise a single worker process is spawned. If max_workers is lower than or equal to 0, then a ValueError will be raised.

initializer is an optional callable that is called at the start of each worker process before executing any tasks; initargs is a tuple of arguments passed to the initializer. If initializer raises an exception, all pending tasks and any attempt to submit new tasks to the pool will raise a BrokenExecutor exception.

Other parameters:

  • python_exe: Path to the Python interpreter executable used to spawn worker processes, otherwise sys.executable is used.
  • python_args: list or iterable with additional command line flags to pass to the Python executable. Command line flags determined from inspection of sys.flags, sys.warnoptions and sys._xoptions in are passed unconditionally.
  • mpi_info: dict or iterable yielding (key, value) pairs. These (key, value) pairs are passed (through an MPI.Info object) to the MPI.Intracomm.Spawn call used to spawn worker processes. This mechanism allows telling the MPI runtime system where and how to start the processes. Check the documentation of the backend MPI implementation about the set of keys it interprets and the corresponding format for values.
  • globals: dict or iterable yielding (name, value) pairs to initialize the main module namespace in worker processes.
  • main: If set to False, do not import the __main__ module in worker processes. Setting main to False prevents worker processes from accessing definitions in the parent __main__ namespace.
  • path: list or iterable with paths to append to sys.path in worker processes to extend the module search path.
  • wdir: Path to set the current working directory in worker processes using os.chdir(). The initial working directory is set by the MPI implementation. Quality MPI implementations should honor a wdir info key passed through mpi_info, although such feature is not mandatory.
  • env: dict or iterable yielding (name, value) pairs with environment variables to update os.environ in worker processes. The initial environment is set by the MPI implementation. MPI implementations may allow setting the initial environment through mpi_info, however such feature is not required nor recommended by the MPI standard.
  • use_pkl5: If set to True, use pickle5 with out-of-band buffers for interprocess communication. If use_pkl5 is set to None or not given, its value is determined from the MPI4PY_FUTURES_USE_PKL5 environment variable. Using pickle5 with out-of-band buffers may benefit applications dealing with large buffer-like objects like NumPy arrays. See mpi4py.util.pkl5 for additional information.
  • backoff: float value specifying the maximum number of seconds a worker thread or process suspends execution with time.sleep() while idle-waiting. If not set, its value is determined from the MPI4PY_FUTURES_BACKOFF environment variable if set, otherwise the default value of 0.001 seconds is used. Lower values will reduce latency and increase execution throughput for very short-lived tasks, albeit at the expense of spinning CPU cores and increased energy consumption.

Schedule the callable, func, to be executed as func(*args, **kwargs) and returns a Future object representing the execution of the callable.

executor = MPIPoolExecutor(max_workers=1)
future = executor.submit(pow, 321, 1234)
print(future.result())



Equivalent to map(func, *iterables) except func is executed asynchronously and several calls to func may be made concurrently, out-of-order, in separate processes. The returned iterator raises a TimeoutError if __next__() is called and the result isn’t available after timeout seconds from the original call to map(). timeout can be an int or a float. If timeout is not specified or None, there is no limit to the wait time. If a call raises an exception, then that exception will be raised when its value is retrieved from the iterator. This method chops iterables into a number of chunks which it submits to the pool as separate tasks. The (approximate) size of these chunks can be specified by setting chunksize to a positive integer. For very long iterables, using a large value for chunksize can significantly improve performance compared to the default size of one. By default, the returned iterator yields results in-order, waiting for successive tasks to complete . This behavior can be changed by passing the keyword argument unordered as True, then the result iterator will yield a result as soon as any of the tasks complete.

executor = MPIPoolExecutor(max_workers=3)
for result in executor.map(pow, [2]*32, range(32)):

print(result)



Equivalent to itertools.starmap(func, iterable). Used instead of map() when argument parameters are already grouped in tuples from a single iterable (the data has been “pre-zipped”). map(func, *iterable) is equivalent to starmap(func, zip(*iterable)).

executor = MPIPoolExecutor(max_workers=3)
iterable = ((2, n) for n in range(32))
for result in executor.starmap(pow, iterable):

print(result)



Signal the executor that it should free any resources that it is using when the currently pending futures are done executing. Calls to submit() and map() made after shutdown() will raise RuntimeError.

If wait is True then this method will not return until all the pending futures are done executing and the resources associated with the executor have been freed. If wait is False then this method will return immediately and the resources associated with the executor will be freed when all pending futures are done executing. Regardless of the value of wait, the entire Python program will not exit until all pending futures are done executing.

If cancel_futures is True, this method will cancel all pending futures that the executor has not started running. Any futures that are completed or running won’t be cancelled, regardless of the value of cancel_futures.

You can avoid having to call this method explicitly if you use the with statement, which will shutdown the executor instance (waiting as if shutdown() were called with wait set to True).

import time
with MPIPoolExecutor(max_workers=1) as executor:

future = executor.submit(time.sleep, 2) assert future.done()



Signal the executor that it should allocate eagerly any required resources (in particular, MPI worker processes). If wait is True, then bootup() will not return until the executor resources are ready to process submissions. Resources are automatically allocated in the first call to submit(), thus calling bootup() explicitly is seldom needed.

Number or worker processes in the pool.


If the max_workers parameter to MPIPoolExecutor is None or not given, the MPI4PY_FUTURES_MAX_WORKERS environment variable provides a fallback value for the maximum number of MPI worker processes to spawn.

Added in version 3.1.0.


If the use_pkl5 keyword argument to MPIPoolExecutor is None or not given, the MPI4PY_FUTURES_USE_PKL5 environment variable provides a fallback value for whether the executor should use pickle5 with out-of-band buffers for interprocess communication. Accepted values are 0 and 1 (interpreted as False and True, respectively), and strings specifying a YAML boolean value (case-insensitive). Using pickle5 with out-of-band buffers may benefit applications dealing with large buffer-like objects like NumPy arrays. See mpi4py.util.pkl5 for additional information.

Added in version 4.0.0.


If the backoff keyword argument to MPIPoolExecutor is not given, the MPI4PY_FUTURES_BACKOFF environment variable can be set to a float value specifying the maximum number of seconds a worker thread or process suspends execution with time.sleep() while idle-waiting. If not set, the default backoff value is 0.001 seconds. Lower values will reduce latency and increase execution throughput for very short-lived tasks, albeit at the expense of spinning CPU cores and increased energy consumption.

Added in version 4.0.0.


NOTE:

As the master process uses a separate thread to perform MPI communication with the workers, the backend MPI implementation should provide support for MPI.THREAD_MULTIPLE. However, some popular MPI implementations do not support yet concurrent MPI calls from multiple threads. Additionally, users may decide to initialize MPI with a lower level of thread support. If the level of thread support in the backend MPI is less than MPI.THREAD_MULTIPLE, mpi4py.futures will use a global lock to serialize MPI calls. If the level of thread support is less than MPI.THREAD_SERIALIZED, mpi4py.futures will emit a RuntimeWarning.


WARNING:

If the level of thread support in the backend MPI is less than MPI.THREAD_SERIALIZED (i.e, it is either MPI.THREAD_SINGLE or MPI.THREAD_FUNNELED), in theory mpi4py.futures cannot be used. Rather than raising an exception, mpi4py.futures emits a warning and takes a “cross-fingers” attitude to continue execution in the hope that serializing MPI calls with a global lock will actually work.


MPICommExecutor

Legacy MPI-1 implementations (as well as some vendor MPI-2 implementations) do not support the dynamic process management features introduced in the MPI-2 standard. Additionally, job schedulers and batch systems in supercomputing facilities may pose additional complications to applications using the MPI_Comm_spawn() routine.

With these issues in mind, mpi4py.futures supports an additional, more traditional, SPMD-like usage pattern requiring MPI-1 calls only. Python applications are started the usual way, e.g., using the mpiexec command. Python code should make a collective call to the MPICommExecutor context manager to partition the set of MPI processes within a MPI communicator in one master processes and many workers processes. The master process gets access to an MPIPoolExecutor instance to submit tasks. Meanwhile, the worker process follow a different execution path and team-up to execute the tasks submitted from the master.

Besides alleviating the lack of dynamic process management features in legacy MPI-1 or partial MPI-2 implementations, the MPICommExecutor context manager may be useful in classic MPI-based Python applications willing to take advantage of the simple, task-based, master/worker approach available in the mpi4py.futures package.

Context manager for MPIPoolExecutor. This context manager splits a MPI (intra)communicator comm (defaults to MPI.COMM_WORLD if not provided or None) in two disjoint sets: a single master process (with rank root in comm) and the remaining worker processes. These sets are then connected through an intercommunicator. The target of the with statement is assigned either an MPIPoolExecutor instance (at the master) or None (at the workers).

from mpi4py import MPI
from mpi4py.futures import MPICommExecutor
with MPICommExecutor(MPI.COMM_WORLD, root=0) as executor:

if executor is not None:
future = executor.submit(abs, -42)
assert future.result() == 42
answer = set(executor.map(abs, [-42, 42]))
assert answer == {42}



WARNING:

If MPICommExecutor is passed a communicator of size one (e.g., MPI.COMM_SELF), then the executor instance assigned to the target of the with statement will execute all submitted tasks in a single worker thread, thus ensuring that task execution still progress asynchronously. However, the GIL will prevent the main and worker threads from running concurrently in multicore processors. Moreover, the thread context switching may harm noticeably the performance of CPU-bound tasks. In case of I/O-bound tasks, the GIL is not usually an issue, however, as a single worker thread is used, it progress one task at a time. We advice against using MPICommExecutor with communicators of size one and suggest refactoring your code to use instead a ThreadPoolExecutor.


Command line

Recalling the issues related to the lack of support for dynamic process management features in MPI implementations, mpi4py.futures supports an alternative usage pattern where Python code (either from scripts, modules, or zip files) is run under command line control of the mpi4py.futures package by passing -m mpi4py.futures to the python executable. The mpi4py.futures invocation should be passed a pyfile path to a script (or a zipfile/directory containing a __main__.py file). Additionally, mpi4py.futures accepts -m mod to execute a module named mod, -c cmd to execute a command string cmd, or even - to read commands from standard input (sys.stdin). Summarizing, mpi4py.futures can be invoked in the following ways:

  • $ mpiexec -n numprocs python -m mpi4py.futures pyfile [arg] ...
  • $ mpiexec -n numprocs python -m mpi4py.futures -m mod [arg] ...
  • $ mpiexec -n numprocs python -m mpi4py.futures -c cmd [arg] ...
  • $ mpiexec -n numprocs python -m mpi4py.futures - [arg] ...

Before starting the main script execution, mpi4py.futures splits MPI.COMM_WORLD in one master (the process with rank 0 in MPI.COMM_WORLD) and numprocs - 1 workers and connects them through an MPI intercommunicator. Afterwards, the master process proceeds with the execution of the user script code, which eventually creates MPIPoolExecutor instances to submit tasks. Meanwhile, the worker processes follow a different execution path to serve the master. Upon successful termination of the main script at the master, the entire MPI execution environment exists gracefully. In case of any unhandled exception in the main script, the master process calls MPI.COMM_WORLD.Abort(1) to prevent deadlocks and force termination of entire MPI execution environment.

WARNING:

Running scripts under command line control of mpi4py.futures is quite similar to executing a single-process application that spawn additional workers as required. However, there is a very important difference users should be aware of. All MPIPoolExecutor instances created at the master will share the pool of workers. Tasks submitted at the master from many different executors will be scheduled for execution in random order as soon as a worker is idle. Any executor can easily starve all the workers (e.g., by calling MPIPoolExecutor.map() with long iterables). If that ever happens, submissions from other executors will not be serviced until free workers are available.


SEE ALSO:

Documentation on Python command line interface.



Parallel tasks

The mpi4py.futures package favors an embarrassingly parallel execution model involving a series of sequential tasks independent of each other and executed asynchronously. Albeit unnatural, MPIPoolExecutor can still be used for handling workloads involving parallel tasks, where worker processes communicate and coordinate each other via MPI.

Access an intracommunicator grouping MPI worker processes.

Executing parallel tasks with mpi4py.futures requires following some rules, cf. highlighted lines in example cpi.py :

  • Use MPIPoolExecutor.num_workers to determine the number of worker processes in the executor and submit exactly one callable per worker process using the MPIPoolExecutor.submit() method.
  • The submitted callable must use get_comm_workers() to access an intracommunicator grouping MPI worker processes. Afterwards, it is highly recommended calling the Barrier() method on the communicator. The barrier synchronization ensures that every worker process is executing the submitted callable exactly once. Afterwards, the parallel task can safely perform any kind of point-to-point or collective operation using the returned communicator.
  • The Future instances returned by MPIPoolExecutor.submit() should be collected in a sequence. Use wait() with the sequence of Future instances to ensure logical completion of the parallel task.

Utilities

The mpi4py.futures package provides additional utilities for handling Future instances.

Gather a collection of futures in a new future.
fs – Collection of futures.
New future producing as result a list with results from fs.


Compose the completion of a future with result and exception handlers.
  • future – Input future instance.
  • resulthook – Function to be called once the input future completes with success. Once the input future finish running with success, its result value is the input argument for resulthook. The result of resulthook is set as the result of the output future. If resulthook is None, the output future is completed directly with the result of the input future.
  • excepthook – Function to be called once the input future completes with failure. Once the input future finish running with failure, its exception value is the input argument for excepthook. If excepthook returns an Exception instance, it is set as the exception of the output future. Otherwise, the result of excepthook is set as the result of the output future. If excepthook is None, the output future is set as failed with the exception from the input future.

Output future instance to be completed once the input future is completed and either resulthook or excepthook finish executing.


Examples

Computing the Julia set

The following julia.py script computes the Julia set and dumps an image to disk in binary PGM format. The code starts by importing MPIPoolExecutor from the mpi4py.futures package. Next, some global constants and functions implement the computation of the Julia set. The computations are protected with the standard if __name__ == '__main__': ... idiom. The image is computed by whole scanlines submitting all these tasks at once using the map method. The result iterator yields scanlines in-order as the tasks complete. Finally, each scanline is dumped to disk.

julia.py

from mpi4py.futures import MPIPoolExecutor
x0, x1, w = -2.0, +2.0, 640*2
y0, y1, h = -1.5, +1.5, 480*2
dx = (x1 - x0) / w
dy = (y1 - y0) / h
c = complex(0, 0.65)
def julia(x, y):

z = complex(x, y)
n = 255
while abs(z) < 3 and n > 1:
z = z**2 + c
n -= 1
return n def julia_line(k):
line = bytearray(w)
y = y1 - k * dy
for j in range(w):
x = x0 + j * dx
line[j] = julia(x, y)
return line if __name__ == '__main__':
with MPIPoolExecutor() as executor:
image = executor.map(julia_line, range(h))
with open('julia.pgm', 'wb') as f:
f.write(b'P5 %d %d %d\n' % (w, h, 255))
for line in image:
f.write(line)


The recommended way to execute the script is by using the mpiexec command specifying one MPI process (master) and (optional but recommended) the desired MPI universe size, which determines the number of additional dynamically spawned processes (workers). The MPI universe size is provided either by a batch system or set by the user via command-line arguments to mpiexec or environment variables. Below we provide examples for MPICH and Open MPI implementations [1]. In all of these examples, the mpiexec command launches a single master process running the Python interpreter and executing the main script. When required, mpi4py.futures spawns the pool of 16 worker processes. The master submits tasks to the workers and waits for the results. The workers receive incoming tasks, execute them, and send back the results to the master.

When using MPICH implementation or its derivatives based on the Hydra process manager, users can set the MPI universe size via the -usize argument to mpiexec:

$ mpiexec -n 1 -usize 17 python julia.py


or, alternatively, by setting the MPIEXEC_UNIVERSE_SIZE environment variable:

$ env MPIEXEC_UNIVERSE_SIZE=17 mpiexec -n 1 python julia.py


In the Open MPI implementation, the MPI universe size can be set via the -host argument to mpiexec:

$ mpiexec -n 1 -host localhost:17 python julia.py


Another way to specify the number of workers is to use the mpi4py.futures-specific environment variable MPI4PY_FUTURES_MAX_WORKERS:

$ env MPI4PY_FUTURES_MAX_WORKERS=16 mpiexec -n 1 python julia.py


Note that in this case, the MPI universe size is ignored.

Alternatively, users may decide to execute the script in a more traditional way, that is, all the MPI processes are started at once. The user script is run under command-line control of mpi4py.futures passing the -m flag to the python executable:

$ mpiexec -n 17 python -m mpi4py.futures julia.py


As explained previously, the 17 processes are partitioned in one master and 16 workers. The master process executes the main script while the workers execute the tasks submitted by the master.

[1]
When using an MPI implementation other than MPICH or Open MPI, please check the documentation of the implementation and/or batch system for the ways to specify the desired MPI universe size.

Computing Pi (parallel task)

The number \pi can be approximated via numerical integration with the simple midpoint rule, that is:


\pi = \int_{0}^{1} \frac{4}{1+x^2} \,dx \approx
\frac{1}{n} \sum_{i=1}^{n} \frac{4}{1 + \left[\frac{1}{n} \left(i-\frac{1}{2}\right) \right]^2} .

The following cpi.py script computes such approximations using mpi4py.futures with a parallel task involving a collective reduction operation. Highlighted lines correspond to the rules discussed in Parallel tasks.

cpi.py

import math
import sys
from mpi4py.futures import MPIPoolExecutor, wait
from mpi4py.futures import get_comm_workers
def compute_pi(n):

# Access intracommunicator and synchronize
comm = get_comm_workers()
comm.Barrier()
rank = comm.Get_rank()
size = comm.Get_size()
# Local computation
h = 1.0 / n
s = 0.0
for i in range(rank + 1, n + 1, size):
x = h * (i - 0.5)
s += 4.0 / (1.0 + x**2)
pi_partial = s * h
# Parallel reduce-to-all
pi = comm.allreduce(pi_partial)
# All workers return the same value
return pi if __name__ == '__main__':
n = int(sys.argv[1]) if len(sys.argv) > 1 else 256
with MPIPoolExecutor() as executor:
# Submit exactly one callable per worker
P = executor.num_workers
fs = [executor.submit(compute_pi, n) for _ in range(P)]
# Wait for all workers to finish
wait(fs)
# Get result from the first future object.
# In this particular example, due to using reduce-to-all,
# all the other future objects hold the same result value.
pi = fs[0].result()
print(
f"pi: {pi:.16f}, error: {abs(pi - math.pi):.3e}",
f"({n:d} intervals, {P:d} workers)",
)


To run in modern MPI-2 mode:

$ env MPI4PY_FUTURES_MAX_WORKERS=4 mpiexec -n 1 python cpi.py 128
pi: 3.1415977398528137, error: 5.086e-06 (128 intervals, 4 workers)
$ env MPI4PY_FUTURES_MAX_WORKERS=8 mpiexec -n 1 python cpi.py 512
pi: 3.1415929714812316, error: 3.179e-07 (512 intervals, 8 workers)


To run in legacy MPI-1 mode:

$ mpiexec -n 5 python -m mpi4py.futures cpi.py 128
pi: 3.1415977398528137, error: 5.086e-06 (128 intervals, 4 workers)
$ mpiexec -n 9 python -m mpi4py.futures cpi.py 512
pi: 3.1415929714812316, error: 3.179e-07 (512 intervals, 8 workers)


Citation

If mpi4py.futures been significant to a project that leads to an academic publication, please acknowledge our work by citing the following article [mpi4py-futures]:

[mpi4py-futures]
M. Rogowski, S. Aseeri, D. Keyes, and L. Dalcin, mpi4py.futures: MPI-Based Asynchronous Task Execution for Python, IEEE Transactions on Parallel and Distributed Systems, 34(2):611-622, 2023. https://doi.org/10.1109/TPDS.2022.3225481

MPI4PY.UTIL

Added in version 3.1.0.

The mpi4py.util package collects miscellaneous utilities within the intersection of Python and MPI.

mpi4py.util.dtlib

Added in version 3.1.0.

The mpi4py.util.dtlib module provides converter routines between NumPy and MPI datatypes.

Convert NumPy datatype to MPI datatype.
dtype (DTypeLike) – NumPy dtype-like object.
Datatype


Convert MPI datatype to NumPy datatype.
datatype (Datatype) – MPI datatype.
dtype[Any]


mpi4py.util.pkl5

Added in version 3.1.0.

pickle protocol 5 (see PEP 574) introduced support for out-of-band buffers, allowing for more efficient handling of certain object types with large memory footprints.

MPI for Python uses the traditional in-band handling of buffers. This approach is appropriate for communicating non-buffer Python objects, or buffer-like objects with small memory footprints. For point-to-point communication, in-band buffer handling allows for the communication of a pickled stream with a single MPI message, at the expense of additional CPU and memory overhead in the pickling and unpickling steps.

The mpi4py.util.pkl5 module provides communicator wrapper classes reimplementing pickle-based point-to-point and collective communication methods using pickle protocol 5. Handling out-of-band buffers necessarily involves multiple MPI messages, thus increasing latency and hurting performance in case of small size data. However, in case of large size data, the zero-copy savings of out-of-band buffer handling more than offset the extra latency costs. Additionally, these wrapper methods overcome the infamous 2 GiB message count limit (MPI-1 to MPI-3).

NOTE:

Support for pickle protocol 5 is available in the pickle module within the Python standard library since Python 3.8. Previous Python 3 releases can use the pickle5 backport, which is available on PyPI and can be installed with:

python -m pip install pickle5




Request.

Custom request class for nonblocking communications.

NOTE:

Request is not a subclass of mpi4py.MPI.Request


Free a communication request.


Free a communication request.


Cancel a communication request.


Non-destructive test for the completion of a request.
status (Status | None)
bool


Test for the completion of a request.
status (Status | None)
tuple[bool, Any | None]


Wait for a request to complete.
status (Status | None)
Any


Non-destructive test for the completion of all requests.


Test for the completion of all requests.


Wait for all requests to complete.



Message.

Custom message class for matching probes.

NOTE:

Message is not a subclass of mpi4py.MPI.Message


Do nothing.


Blocking receive of matched message.
status (Status | None)
Any


Nonblocking receive of matched message.
Request





Communicator.

Base communicator wrapper class.

Blocking send in standard mode.
  • obj (Any)
  • dest (int)
  • tag (int)

None


Blocking send in buffered mode.
  • obj (Any)
  • dest (int)
  • tag (int)

None


Blocking send in synchronous mode.
  • obj (Any)
  • dest (int)
  • tag (int)

None


Nonblocking send in standard mode.
  • obj (Any)
  • dest (int)
  • tag (int)

Request


Nonblocking send in buffered mode.
  • obj (Any)
  • dest (int)
  • tag (int)

Request


Nonblocking send in synchronous mode.
  • obj (Any)
  • dest (int)
  • tag (int)

Request


Blocking receive.
  • buf (Buffer | None)
  • source (int)
  • tag (int)
  • status (Status | None)

Any


Nonblocking receive.

WARNING:

This method cannot be supported reliably and raises RuntimeError.


  • buf (Buffer | None)
  • source (int)
  • tag (int)

Request


Send and receive.
  • sendobj (Any)
  • dest (int)
  • sendtag (int)
  • recvbuf (Buffer | None)
  • source (int)
  • recvtag (int)
  • status (Status | None)

Any


Blocking test for a matched message.
  • source (int)
  • tag (int)
  • status (Status | None)

Message


Nonblocking test for a matched message.
  • source (int)
  • tag (int)
  • status (Status | None)

Message | None


Broadcast.

Added in version 3.1.0.

  • obj (Any)
  • root (int)

Any


Gather.

Added in version 4.0.0.

  • sendobj (Any)
  • root (int)

list[Any] | None


Scatter.

Added in version 4.0.0.

  • sendobj (Sequence[Any] | None)
  • root (int)

Any


Gather to All.

Added in version 4.0.0.

sendobj (Any)
list[Any]


All to All Scatter/Gather.

Added in version 4.0.0.

sendobj (Sequence[Any])
list[Any]



Intracommunicator.

Intracommunicator wrapper class.


Intercommunicator.

Intercommunicator wrapper class.


Examples

test-pkl5-1.py

import numpy as np
from mpi4py import MPI
from mpi4py.util import pkl5
comm = pkl5.Intracomm(MPI.COMM_WORLD)  # comm wrapper
size = comm.Get_size()
rank = comm.Get_rank()
dst = (rank + 1) % size
src = (rank - 1) % size
sobj = np.full(1024**3, rank, dtype='i4')  # > 4 GiB
sreq = comm.isend(sobj, dst, tag=42)
robj = comm.recv (None, src, tag=42)
sreq.Free()
assert np.min(robj) == src
assert np.max(robj) == src


test-pkl5-2.py

import numpy as np
from mpi4py import MPI
from mpi4py.util import pkl5
comm = pkl5.Intracomm(MPI.COMM_WORLD)  # comm wrapper
size = comm.Get_size()
rank = comm.Get_rank()
dst = (rank + 1) % size
src = (rank - 1) % size
sobj = np.full(1024**3, rank, dtype='i4')  # > 4 GiB
sreq = comm.isend(sobj, dst, tag=42)
status = MPI.Status()
rmsg = comm.mprobe(status=status)
assert status.Get_source() == src
assert status.Get_tag() == 42
rreq = rmsg.irecv()
robj = rreq.wait()
sreq.Free()
assert np.max(robj) == src
assert np.min(robj) == src


mpi4py.util.pool

Added in version 4.0.0.

SEE ALSO:

This module intends to be a drop-in replacement for the multiprocessing.pool interface from the Python standard library. The Pool class exposed here is implemented as a thin wrapper around MPIPoolExecutor.


NOTE:

The mpi4py.futures package offers a higher level interface for asynchronously pushing tasks to MPI worker process, allowing for a clear separation between submitting tasks and waiting for the results.


Pool using MPI processes as workers.
__init__(processes=None, initializer=None, initargs=(), **kwargs)
Initialize a new Pool instance.
  • processes (int | None) – Number of worker processes.
  • initializer (Callable[[...], None] | None) – An callable used to initialize workers processes.
  • initargs (Iterable[Any]) – A tuple of arguments to pass to the initializer.
  • kwargs (Any)

None

NOTE:

Additional keyword arguments are passed down to the MPIPoolExecutor constructor.


WARNING:

The maxtasksperchild and context arguments of multiprocessing.pool.Pool are not supported. Specifying maxtasksperchild or context with a value other than None will issue a warning of category UserWarning.



Call func with arguments args and keyword arguments kwds.

Equivalent to func(*args, **kwds).

  • func (Callable[[...], T])
  • args (Iterable[Any])
  • kwds (Mapping[str, Any])

T


Asynchronous version of apply() returning ApplyResult.
  • func (Callable[..., T])
  • args (Iterable[Any])
  • kwds (Mapping[str, Any])
  • callback (Callable[[T], object] | None)
  • error_callback (Callable[[BaseException], object] | None)

AsyncResult[T]


Apply func to each element in iterable.

Equivalent to list(map(func, iterable)).

Block until all results are ready and return them in a list.

The iterable is choped into a number of chunks which are submitted as separate tasks. The (approximate) size of these chunks can be specified by setting chunksize to a positive integer.

Consider using imap() or imap_unordered() with explicit chunksize for better efficiency.

  • func (Callable[[S], T])
  • iterable (Iterable[S])
  • chunksize (int | None)

list[T]


Asynchronous version of map() returning MapResult.
  • func (Callable[[S], T])
  • iterable (Iterable[S])
  • chunksize (int | None)
  • callback (Callable[[T], None] | None)
  • error_callback (Callable[[BaseException], None] | None)

MapResult[T]


Like map() but return an iterator.

Equivalent to map(func, iterable).

  • func (Callable[[S], T])
  • iterable (Iterable[S])
  • chunksize (int)

Iterator[T]


Like imap() but ordering of results is arbitrary.
  • func (Callable[[S], T])
  • iterable (Iterable[S])
  • chunksize (int)

Iterator[T]


Apply func to each argument tuple in iterable.

Equivalent to list(itertools.starmap(func, iterable)).

Block until all results are ready and return them in a list.

The iterable is choped into a number of chunks which are submitted as separate tasks. The (approximate) size of these chunks can be specified by setting chunksize to a positive integer.

Consider using istarmap() or istarmap_unordered() with explicit chunksize for better efficiency.

  • func (Callable[[...], T])
  • iterable (Iterable[Iterable[Any]])
  • chunksize (int | None)

list[T]


Asynchronous version of starmap() returning MapResult.
  • func (Callable[..., T])
  • iterable (Iterable[Iterable[Any]])
  • chunksize (int | None)
  • callback (Callable[[T], None] | None)
  • error_callback (Callable[[BaseException], None] | None)

MapResult[T]


Like starmap() but return an iterator.

Equivalent to itertools.starmap(func, iterable).

  • func (Callable[[...], T])
  • iterable (Iterable[Iterable[Any]])
  • chunksize (int)

Iterator[T]


Like istarmap() but ordering of results is arbitrary.
  • func (Callable[[...], T])
  • iterable (Iterable[Iterable[Any]])
  • chunksize (int)

Iterator[T]


Prevent any more tasks from being submitted to the pool.


Stop the worker processes without completing pending tasks.


Wait for the worker processes to exit.



Bases: Pool

Pool using threads as workers.


Asynchronous result.
Return the result when it arrives.

If timeout is not None and the result does not arrive within timeout seconds then raise TimeoutError.

If the remote call raised an exception then that exception will be reraised.

timeout (float | None)
T


Wait until the result is available or timeout seconds pass.
timeout (float | None)
None


Return whether the call has completed.


Return whether the call completed without raising an exception.

If the result is not ready then raise ValueError.




Bases: AsyncResult

Result type of apply_async().


Bases: AsyncResult

Result type of map_async() and starmap_async().


mpi4py.util.sync

Added in version 4.0.0.

The mpi4py.util.sync module provides parallel synchronization utilities.

Sequential execution

Sequential execution.

Context manager for sequential execution within a group of MPI processes.

The implementation is based in MPI-1 point-to-point communication. A process with rank i waits in a blocking receive until the previous process rank i-1 finish executing and signals the next rank i with a send.

__init__(comm, tag=0)
Initialize sequential execution.
  • comm (Intracomm) – Intracommunicator context.
  • tag (int) – Tag for point-to-point communication.

None


__enter__()
Enter sequential execution.


__exit__(*exc)
Exit sequential execution.
exc (object)
None


Begin sequential execution.


End sequential execution.



Global counter

Global counter.

Produce consecutive values within a group of MPI processes. The counter interface is close to that of itertools.count.

The implementation is based in MPI-3 one-sided operations. A root process (typically rank 0) holds the counter, and its value is queried and incremented with an atomic RMA fetch-and-add operation.

__init__(start=0, step=1, *, typecode='i', comm=COMM_SELF, info=INFO_NULL, root=0)
Initialize global counter.
  • start (int) – Start value.
  • step (int) – Increment value.
  • typecode (str) – Type code as defined in the array module.
  • comm (Intracomm) – Intracommunicator context.
  • info (Info) – Info object for RMA context creation.
  • root (int) – Process rank holding the counter memory.

None


__iter__()
Implement iter(self).


__next__()
Implement next(self).


Return current value and increment.
incr (int | None) – Increment value.
The counter value before incrementing.
int


Free counter resources.



Mutual exclusion

Mutual exclusion.

Establish a critical section or mutual exclusion among MPI processes.

The mutex interface is close to that of threading.Lock and threading.RLock, allowing the use of either recursive or non-recursive mutual exclusion. However, a mutex should be used within a group of MPI processes, not threads.

In non-recursive mode, the semantics of Mutex are somewhat different than these of threading.Lock:

  • Once acquired, a mutex is held and owned by a process until released.
  • Trying to acquire a mutex already held raises RuntimeError.
  • Trying to release a mutex not yet held raises RuntimeError.

This mutex implementation uses the scalable and fair spinlock algorithm from [mcs-paper] and took inspiration from the MPI-3 RMA implementation of [uam-book].

__init__(*, recursive=False, comm=COMM_SELF, info=INFO_NULL)
Initialize mutex object.
  • comm (Intracomm) – Intracommunicator context.
  • recursive (bool) – Whether to allow recursive acquisition.
  • info (Info) – Info object for RMA context creation.

None


__enter__()
Acquire mutex.


__exit__(*exc)
Release mutex.
exc (object)
None


Acquire mutex, blocking or non-blocking.
blocking (bool) – If True, block until the mutex is held.
True if the mutex is held, False otherwise.
bool


Release mutex.


Return whether the mutex is held.


Return the recursion count.


Free mutex resources.



[mcs-paper]
John M. Mellor-Crummey and Michael L. Scott. Algorithms for scalable synchronization on shared-memory multiprocessors. ACM Transactions on Computer Systems, 9(1):21-65, February 1991. https://doi.org/10.1145/103727.103729
[uam-book]
William Gropp, Torsten Hoefler, Rajeev Thakur, Ewing Lusk. Using Advanced MPI - Modern Features of the Message-Passing Interface. Chapter 4, Section 4.7, Pages 130-131. The MIT Press, November 2014. https://mitpress.mit.edu/9780262527637/using-advanced-mpi/

Condition variable

Condition variable.

A condition variable allows one or more MPI processes to wait until they are notified by another processes.

The condition variable interface is close to that of threading.Condition, allowing the use of either recursive or non-recursive mutual exclusion. However, the condition variable should be used within a group of MPI processes, not threads.

This condition variable implementation uses a MPI-3 RMA-based scalable and fair circular queue algorithm to track the set of waiting processes.

__init__(mutex=None, *, recursive=True, comm=COMM_SELF, info=INFO_NULL)
Initialize condition variable.
  • mutex (Mutex | None) – Mutual exclusion object.
  • recursive (bool) – Whether to allow recursive acquisition.
  • comm (Intracomm) – Intracommunicator context.
  • info (Info) – Info object for RMA context creation.

None


__enter__()
Acquire the underlying mutex.


__exit__(*exc)
Release the underlying mutex.
exc (object)
None


Acquire the underlying mutex.
blocking (bool)
bool


Release the underlying mutex.


Return whether the underlying mutex is held.


Wait until notified by another process.
Always True.
Literal[True]


Wait until a predicate evaluates to True.
predicate (Callable[[], T]) – callable returning a boolean.
The result of predicate once it evaluates to True.
T


Wake up one or more processes waiting on this condition.
n (int) – Maximum number of processes to wake up.
The actual number of processes woken up.
int


Wake up all processes waiting on this condition.
The actual number of processes woken up.
int


Free condition resources.



Semaphore object

Semaphore object.

A semaphore object manages an internal counter which is decremented by each acquire() call and incremented by each release() call. The internal counter never reaches a value below zero; when acquire() finds that it is zero, it blocks and waits until some other process calls release().

The semaphore interface is close to that of threading.Semaphore and threading.BoundedSemaphore, allowing the use of either bounded (default) or unbounded semaphores. With a bounded semaphore, the internal counter never exceeds its initial value; otherwise release() raises ValueError.

This semaphore implementation uses a global Counter and a Condition variable to handle waiting and and notification.

__init__(value=1, *, bounded=True, comm=COMM_SELF, info=INFO_NULL)
Initialize semaphore object.
  • value (int) – Initial value for internal counter.
  • bounded (bool) – Bound internal counter to initial value.
  • comm (Intracomm) – Intracommunicator context.
  • info (Info) – Info object for RMA context creation.

None


__enter__()
Acquire semaphore.


__exit__(*exc)
Release semaphore.
exc (object)
None


Acquire semaphore, decrementing the internal counter by one.
blocking (bool) – If True, block until the semaphore is acquired.
True if the semaphore is acquired, False otherwise.
bool


Release semaphore, incrementing the internal counter by one or more.
n (int) – Increment for the internal counter.
None


Free semaphore resources.



Examples

test-sync-1.py

from mpi4py import MPI
from mpi4py.util.sync import Counter, Sequential
comm = MPI.COMM_WORLD
counter = Counter(comm)
with Sequential(comm):

value = next(counter) counter.free() assert comm.rank == value


test-sync-2.py

from mpi4py import MPI
from mpi4py.util.sync import Counter, Mutex
comm = MPI.COMM_WORLD
mutex = Mutex(comm)
counter = Counter(comm)
with mutex:

value = next(counter) counter.free() mutex.free() assert (
list(range(comm.size)) ==
sorted(comm.allgather(value)) )


MPI4PY.RUN

Added in version 3.0.0.

At import time, mpi4py initializes the MPI execution environment calling MPI_Init_thread() and installs an exit hook to automatically call MPI_Finalize() just before the Python process terminates. Additionally, mpi4py overrides the default ERRORS_ARE_FATAL error handler in favor of ERRORS_RETURN, which allows translating MPI errors in Python exceptions. These departures from standard MPI behavior may be controversial, but are quite convenient within the highly dynamic Python programming environment. Third-party code using mpi4py can just from mpi4py import MPI and perform MPI calls without the tedious initialization/finalization handling. MPI errors, once translated automatically to Python exceptions, can be dealt with the common tryexceptfinally clauses; unhandled MPI exceptions will print a traceback which helps in locating problems in source code.

Unfortunately, the interplay of automatic MPI finalization and unhandled exceptions may lead to deadlocks. In unattended runs, these deadlocks will drain the battery of your laptop, or burn precious allocation hours in your supercomputing facility.

Exceptions and deadlocks

Consider the following snippet of Python code. Assume this code is stored in a standard Python script file and run with mpiexec in two or more processes.

deadlock.py

from mpi4py import MPI
assert MPI.COMM_WORLD.Get_size() > 1
rank = MPI.COMM_WORLD.Get_rank()
if rank == 0:

1/0
MPI.COMM_WORLD.send(None, dest=1, tag=42) elif rank == 1:
MPI.COMM_WORLD.recv(source=0, tag=42)


Process 0 raises ZeroDivisionError exception before performing a send call to process 1. As the exception is not handled, the Python interpreter running in process 0 will proceed to exit with non-zero status. However, as mpi4py installed a finalizer hook to call MPI_Finalize() before exit, process 0 will block waiting for other processes to also enter the MPI_Finalize() call. Meanwhile, process 1 will block waiting for a message to arrive from process 0, thus never reaching to MPI_Finalize(). The whole MPI execution environment is irremediably in a deadlock state.

To alleviate this issue, mpi4py offers a simple, alternative command line execution mechanism based on using the -m flag and implemented with the runpy module. To use this features, Python code should be run passing -m mpi4py in the command line invoking the Python interpreter. In case of unhandled exceptions, the finalizer hook will call MPI_Abort() on the MPI_COMM_WORLD communicator, thus effectively aborting the MPI execution environment.

WARNING:

When a process is forced to abort, resources (e.g. open files) are not cleaned-up and any registered finalizers (either with the atexit module, the Python C/API function Py_AtExit(), or even the C standard library function atexit()) will not be executed. Thus, aborting execution is an extremely impolite way of ensuring process termination. However, MPI provides no other mechanism to recover from a deadlock state.


Command line

The use of -m mpi4py to execute Python code on the command line resembles that of the Python interpreter.

  • mpiexec -n numprocs python -m mpi4py pyfile [arg] ...
  • mpiexec -n numprocs python -m mpi4py -m mod [arg] ...
  • mpiexec -n numprocs python -m mpi4py -c cmd [arg] ...
  • mpiexec -n numprocs python -m mpi4py - [arg] ...

<pyfile>
Execute the Python code contained in pyfile, which must be a filesystem path referring to either a Python file, a directory containing a __main__.py file, or a zipfile containing a __main__.py file.

Search sys.path for the named module mod and execute its contents.

Execute the Python code in the cmd string command.

-
Read commands from standard input (sys.stdin).

SEE ALSO:

Documentation on Python command line interface.



MPI4PY.BENCH

Added in version 3.0.0.

REFERENCE

mpi4py.MPI Message Passing Interface.

mpi4py.MPI

Message Passing Interface.

Classes

BottomType Type of BOTTOM.
BufferAutomaticType Type of BUFFER_AUTOMATIC.
Cartcomm Cartesian topology intracommunicator.
Comm Communication context.
Datatype Datatype object.
Distgraphcomm Distributed graph topology intracommunicator.
Errhandler Error handler.
File File I/O context.
Graphcomm General graph topology intracommunicator.
Grequest Generalized request handler.
Group Group of processes.
InPlaceType Type of IN_PLACE.
Info Info object.
Intercomm Intercommunicator.
Intracomm Intracommunicator.
Message Matched message.
Op Reduction operation.
Pickle Pickle/unpickle Python objects.
Prequest Persistent request handler.
Request Request handler.
Session Session context.
Status Status object.
Topocomm Topology intracommunicator.
Win Remote memory access context.
buffer Buffer.
memory alias of buffer

mpi4py.MPI.BottomType


mpi4py.MPI.BufferAutomaticType

Bases: int

Type of BUFFER_AUTOMATIC.



mpi4py.MPI.Cartcomm

Bases: Topocomm

Cartesian topology intracommunicator.


Methods Summary

Get_cart_rank(coords) Translate logical coordinates to ranks.
Get_coords(rank) Translate ranks to logical coordinates.
Get_dim() Return number of dimensions.
Get_topo() Return information on the cartesian topology.
Shift(direction, disp) Return a process ranks for data shifting with Sendrecv.
Sub(remain_dims) Return a lower-dimensional Cartesian topology.

Attributes Summary

coords Coordinates.
dim Number of dimensions.
dims Dimensions.
ndim Number of dimensions.
periods Periodicity.
topo Topology information.

Methods Documentation

Translate logical coordinates to ranks.
coords (Sequence[int])
int


Translate ranks to logical coordinates.
rank (int)
list[int]


Return number of dimensions.


Return information on the cartesian topology.
tuple[list[int], list[int], list[int]]


Return a process ranks for data shifting with Sendrecv.
  • direction (int)
  • disp (int)

tuple[int, int]


Return a lower-dimensional Cartesian topology.
remain_dims (Sequence[bool])
Cartcomm


Attributes Documentation

Coordinates.

Number of dimensions.

Dimensions.

Number of dimensions.

Periodicity.

Topology information.


mpi4py.MPI.Comm

Bases: object

Communication context.


Methods Summary

Abort([errorcode]) Terminate the MPI execution environment.
Ack_failed([num_to_ack]) Acknowledge failures on a communicator.
Agree(flag) Blocking agreement.
Allgather(sendbuf, recvbuf) Gather to All.
Allgather_init(sendbuf, recvbuf[, info]) Persistent Gather to All.
Allgatherv(sendbuf, recvbuf) Gather to All Vector.
Allgatherv_init(sendbuf, recvbuf[, info]) Persistent Gather to All Vector.
Allreduce(sendbuf, recvbuf[, op]) Reduce to All.
Allreduce_init(sendbuf, recvbuf[, op, info]) Persistent Reduce to All.
Alltoall(sendbuf, recvbuf) All to All Scatter/Gather.
Alltoall_init(sendbuf, recvbuf[, info]) Persistent All to All Scatter/Gather.
Alltoallv(sendbuf, recvbuf) All to All Scatter/Gather Vector.
Alltoallv_init(sendbuf, recvbuf[, info]) Persistent All to All Scatter/Gather Vector.
Alltoallw(sendbuf, recvbuf) All to All Scatter/Gather General.
Alltoallw_init(sendbuf, recvbuf[, info]) Persistent All to All Scatter/Gather General.
Attach_buffer(buf) Attach a user-provided buffer for sending in buffered mode.
Barrier() Barrier synchronization.
Barrier_init([info]) Persistent Barrier.
Bcast(buf[, root]) Broadcast data from one process to all other processes.
Bcast_init(buf[, root, info]) Persistent Broadcast.
Bsend(buf, dest[, tag]) Blocking send in buffered mode.
Bsend_init(buf, dest[, tag]) Persistent request for a send in buffered mode.
Call_errhandler(errorcode) Call the error handler installed on a communicator.
Clone() Clone an existing communicator.
Compare(comm) Compare two communicators.
Create(group) Create communicator from group.
Create_errhandler(errhandler_fn) Create a new error handler for communicators.
Create_keyval([copy_fn, delete_fn, nopython]) Create a new attribute key for communicators.
Delete_attr(keyval) Delete attribute value associated with a key.
Detach_buffer() Remove an existing attached buffer.
Disconnect() Disconnect from a communicator.
Dup([info]) Duplicate a communicator.
Dup_with_info(info) Duplicate a communicator with hints.
Flush_buffer() Block until all buffered messages have been transmitted.
Free() Free a communicator.
Free_keyval(keyval) Free an attribute key for communicators.
Gather(sendbuf, recvbuf[, root]) Gather data to one process from all other processes.
Gather_init(sendbuf, recvbuf[, root, info]) Persistent Gather.
Gatherv(sendbuf, recvbuf[, root]) Gather Vector.
Gatherv_init(sendbuf, recvbuf[, root, info]) Persistent Gather Vector.
Get_attr(keyval) Retrieve attribute value by key.
Get_errhandler() Get the error handler for a communicator.
Get_failed() Extract the group of failed processes.
Get_group() Access the group associated with a communicator.
Get_info() Return the current hints for a communicator.
Get_name() Get the print name for this communicator.
Get_parent() Return the parent intercommunicator for this process.
Get_rank() Return the rank of this process in a communicator.
Get_size() Return the number of processes in a communicator.
Get_topology() Return the type of topology (if any) associated with a communicator.
Iagree(flag) Nonblocking agreement.
Iallgather(sendbuf, recvbuf) Nonblocking Gather to All.
Iallgatherv(sendbuf, recvbuf) Nonblocking Gather to All Vector.
Iallreduce(sendbuf, recvbuf[, op]) Nonblocking Reduce to All.
Ialltoall(sendbuf, recvbuf) Nonblocking All to All Scatter/Gather.
Ialltoallv(sendbuf, recvbuf) Nonblocking All to All Scatter/Gather Vector.
Ialltoallw(sendbuf, recvbuf) Nonblocking All to All Scatter/Gather General.
Ibarrier() Nonblocking Barrier.
Ibcast(buf[, root]) Nonblocking Broadcast.
Ibsend(buf, dest[, tag]) Nonblocking send in buffered mode.
Idup([info]) Nonblocking duplicate a communicator.
Idup_with_info(info) Nonblocking duplicate a communicator with hints.
Iflush_buffer() Nonblocking flush for buffered messages.
Igather(sendbuf, recvbuf[, root]) Nonblocking Gather.
Igatherv(sendbuf, recvbuf[, root]) Nonblocking Gather Vector.
Improbe([source, tag, status]) Nonblocking test for a matched message.
Iprobe([source, tag, status]) Nonblocking test for a message.
Irecv(buf[, source, tag]) Nonblocking receive.
Ireduce(sendbuf, recvbuf[, op, root]) Nonblocking Reduce to Root.
Ireduce_scatter(sendbuf, recvbuf[, ...]) Nonblocking Reduce-Scatter (vector version).
Ireduce_scatter_block(sendbuf, recvbuf[, op]) Nonblocking Reduce-Scatter Block (regular, non-vector version).
Irsend(buf, dest[, tag]) Nonblocking send in ready mode.
Is_inter() Return whether the communicator is an intercommunicator.
Is_intra() Return whether the communicator is an intracommunicator.
Is_revoked() Indicate whether the communicator has been revoked.
Iscatter(sendbuf, recvbuf[, root]) Nonblocking Scatter.
Iscatterv(sendbuf, recvbuf[, root]) Nonblocking Scatter Vector.
Isend(buf, dest[, tag]) Nonblocking send.
Isendrecv(sendbuf, dest[, sendtag, recvbuf, ...]) Nonblocking send and receive.
Isendrecv_replace(buf, dest[, sendtag, ...]) Send and receive a message.
Ishrink() Nonblocking shrink a communicator to remove all failed processes.
Issend(buf, dest[, tag]) Nonblocking send in synchronous mode.
Join(fd) Interconnect two processes connected by a socket.
Mprobe([source, tag, status]) Blocking test for a matched message.
Precv_init(buf, partitions[, source, tag, info]) Create request for a partitioned recv operation.
Probe([source, tag, status]) Blocking test for a message.
Psend_init(buf, partitions, dest[, tag, info]) Create request for a partitioned send operation.
Recv(buf[, source, tag, status]) Blocking receive.
Recv_init(buf[, source, tag]) Create a persistent request for a receive.
Reduce(sendbuf, recvbuf[, op, root]) Reduce to Root.
Reduce_init(sendbuf, recvbuf[, op, root, info]) Persistent Reduce to Root.
Reduce_scatter(sendbuf, recvbuf[, ...]) Reduce-Scatter (vector version).
Reduce_scatter_block(sendbuf, recvbuf[, op]) Reduce-Scatter Block (regular, non-vector version).
Reduce_scatter_block_init(sendbuf, recvbuf) Persistent Reduce-Scatter Block (regular, non-vector version).
Reduce_scatter_init(sendbuf, recvbuf[, ...]) Persistent Reduce-Scatter (vector version).
Revoke() Revoke a communicator.
Rsend(buf, dest[, tag]) Blocking send in ready mode.
Rsend_init(buf, dest[, tag]) Persistent request for a send in ready mode.
Scatter(sendbuf, recvbuf[, root]) Scatter data from one process to all other processes.
Scatter_init(sendbuf, recvbuf[, root, info]) Persistent Scatter.
Scatterv(sendbuf, recvbuf[, root]) Scatter Vector.
Scatterv_init(sendbuf, recvbuf[, root, info]) Persistent Scatter Vector.
Send(buf, dest[, tag]) Blocking send.
Send_init(buf, dest[, tag]) Create a persistent request for a standard send.
Sendrecv(sendbuf, dest[, sendtag, recvbuf, ...]) Send and receive a message.
Sendrecv_replace(buf, dest[, sendtag, ...]) Send and receive a message.
Set_attr(keyval, attrval) Store attribute value associated with a key.
Set_errhandler(errhandler) Set the error handler for a communicator.
Set_info(info) Set new values for the hints associated with a communicator.
Set_name(name) Set the print name for this communicator.
Shrink() Shrink a communicator to remove all failed processes.
Split([color, key]) Split communicator by color and key.
Split_type(split_type[, key, info]) Split communicator by split type.
Ssend(buf, dest[, tag]) Blocking send in synchronous mode.
Ssend_init(buf, dest[, tag]) Persistent request for a send in synchronous mode.
allgather(sendobj) Gather to All.
allreduce(sendobj[, op]) Reduce to All.
alltoall(sendobj) All to All Scatter/Gather.
barrier() Barrier synchronization.
bcast(obj[, root]) Broadcast.
bsend(obj, dest[, tag]) Send in buffered mode.
f2py(arg)
free() Call Free if not null or predefined.
fromhandle(handle) Create object from MPI handle.
gather(sendobj[, root]) Gather.
ibsend(obj, dest[, tag]) Nonblocking send in buffered mode.
improbe([source, tag, status]) Nonblocking test for a matched message.
iprobe([source, tag, status]) Nonblocking test for a message.
irecv([buf, source, tag]) Nonblocking receive.
isend(obj, dest[, tag]) Nonblocking send.
issend(obj, dest[, tag]) Nonblocking send in synchronous mode.
mprobe([source, tag, status]) Blocking test for a matched message.
probe([source, tag, status]) Blocking test for a message.
py2f()
recv([buf, source, tag, status]) Receive.
reduce(sendobj[, op, root]) Reduce to Root.
scatter(sendobj[, root]) Scatter.
send(obj, dest[, tag]) Send in standard mode.
sendrecv(sendobj, dest[, sendtag, recvbuf, ...]) Send and Receive.
ssend(obj, dest[, tag]) Send in synchronous mode.

Attributes Summary

group Group.
handle MPI handle.
info Info hints.
is_inter Is intercommunicator.
is_intra Is intracommunicator.
is_topo Is a topology.
name Print name.
rank Rank of this process.
size Number of processes.
topology Topology type.

Methods Documentation

Terminate the MPI execution environment.

WARNING:

The invocation of this method prevents the execution of various Python exit and cleanup mechanisms. Use this method as a last resort to prevent parallel deadlocks in case of unrecoverable errors.


errorcode (int)
NoReturn


Acknowledge failures on a communicator.
num_to_ack (int | None)
int


Blocking agreement.
flag (int)
int


Gather to All.

Gather data from all processes and broadcast the combined data to all other processes.

  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecB)

None


Persistent Gather to All.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecB)
  • info (Info)

Prequest


Gather to All Vector.

Gather data from all processes and send it to all other processes providing different amounts of data and displacements.

  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecV)

None


Persistent Gather to All Vector.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecV)
  • info (Info)

Prequest


Reduce to All.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)

None


Persistent Reduce to All.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)
  • info (Info)

Prequest


All to All Scatter/Gather.

Send data to all processes and recv data from all processes.

  • sendbuf (BufSpecB | InPlace)
  • recvbuf (BufSpecB)

None


Persistent All to All Scatter/Gather.
  • sendbuf (BufSpecB | InPlace)
  • recvbuf (BufSpecB)
  • info (Info)

Prequest


All to All Scatter/Gather Vector.

Send data to all processes and recv data from all processes providing different amounts of data and displacements.

  • sendbuf (BufSpecV | InPlace)
  • recvbuf (BufSpecV)

None


Persistent All to All Scatter/Gather Vector.
  • sendbuf (BufSpecV | InPlace)
  • recvbuf (BufSpecV)
  • info (Info)

Prequest


All to All Scatter/Gather General.

Send/recv data to/from all processes allowing the specification of different counts, displacements, and datatypes for each dest/source.

  • sendbuf (BufSpecW | InPlace)
  • recvbuf (BufSpecW)

None


Persistent All to All Scatter/Gather General.
  • sendbuf (BufSpecW | InPlace)
  • recvbuf (BufSpecW)
  • info (Info)

Prequest


Attach a user-provided buffer for sending in buffered mode.
buf (Buffer | None)
None


Barrier synchronization.


Persistent Barrier.
info (Info)
Prequest


Broadcast data from one process to all other processes.
  • buf (BufSpec)
  • root (int)

None


Persistent Broadcast.
  • buf (BufSpec)
  • root (int)
  • info (Info)

Prequest


Blocking send in buffered mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

None


Persistent request for a send in buffered mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Call the error handler installed on a communicator.
errorcode (int)
None


Clone an existing communicator.


Compare two communicators.
comm (Comm)
int


Create communicator from group.
group (Group)
Comm


Create a new error handler for communicators.
errhandler_fn (Callable[[Comm, int], None])
Errhandler


Create a new attribute key for communicators.
  • copy_fn (Callable[[Comm, int, Any], Any] | None)
  • delete_fn (Callable[[Comm, int, Any], None] | None)
  • nopython (bool)

int


Delete attribute value associated with a key.
keyval (int)
None


Remove an existing attached buffer.
Buffer | None


Disconnect from a communicator.


Duplicate a communicator.
info (Info | None)
Self


Duplicate a communicator with hints.
info (Info)
Self


Block until all buffered messages have been transmitted.


Free a communicator.


Free an attribute key for communicators.
keyval (int)
int


Gather data to one process from all other processes.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecB | None)
  • root (int)

None


Persistent Gather.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecB | None)
  • root (int)
  • info (Info)

Prequest


Gather Vector.

Gather data to one process from all other processes providing different amounts of data and displacements.

  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecV | None)
  • root (int)

None


Persistent Gather Vector.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecV | None)
  • root (int)
  • info (Info)

Prequest


Retrieve attribute value by key.
keyval (int)
int | Any | None


Get the error handler for a communicator.
Errhandler


Extract the group of failed processes.
Group


Access the group associated with a communicator.
Group


Return the current hints for a communicator.


Get the print name for this communicator.


Return the parent intercommunicator for this process.
Intercomm


Return the rank of this process in a communicator.


Return the number of processes in a communicator.


Return the type of topology (if any) associated with a communicator.


Nonblocking agreement.
flag (Buffer)
Request


Nonblocking Gather to All.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecB)

Request


Nonblocking Gather to All Vector.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecV)

Request


Nonblocking Reduce to All.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)

Request


Nonblocking All to All Scatter/Gather.
  • sendbuf (BufSpecB | InPlace)
  • recvbuf (BufSpecB)

Request


Nonblocking All to All Scatter/Gather Vector.
  • sendbuf (BufSpecV | InPlace)
  • recvbuf (BufSpecV)

Request


Nonblocking All to All Scatter/Gather General.
  • sendbuf (BufSpecW | InPlace)
  • recvbuf (BufSpecW)

Request


Nonblocking Barrier.
Request


Nonblocking Broadcast.
  • buf (BufSpec)
  • root (int)

Request


Nonblocking send in buffered mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Nonblocking duplicate a communicator.
info (Info | None)
tuple[Self, Request]


Nonblocking duplicate a communicator with hints.
info (Info)
tuple[Self, Request]


Nonblocking flush for buffered messages.
Request


Nonblocking Gather.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecB | None)
  • root (int)

Request


Nonblocking Gather Vector.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpecV | None)
  • root (int)

Request


Nonblocking test for a matched message.
  • source (int)
  • tag (int)
  • status (Status | None)

Message | None


Nonblocking test for a message.
  • source (int)
  • tag (int)
  • status (Status | None)

bool


Nonblocking receive.
  • buf (BufSpec)
  • source (int)
  • tag (int)

Request


Nonblocking Reduce to Root.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec | None)
  • op (Op)
  • root (int)

Request


Nonblocking Reduce-Scatter (vector version).
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • recvcounts (Sequence[int] | None)
  • op (Op)

Request


Nonblocking Reduce-Scatter Block (regular, non-vector version).
  • sendbuf (BufSpecB | InPlace)
  • recvbuf (BufSpec | BufSpecB)
  • op (Op)

Request


Nonblocking send in ready mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Return whether the communicator is an intercommunicator.


Return whether the communicator is an intracommunicator.


Indicate whether the communicator has been revoked.


Nonblocking Scatter.
  • sendbuf (BufSpecB | None)
  • recvbuf (BufSpec | InPlace)
  • root (int)

Request


Nonblocking Scatter Vector.
  • sendbuf (BufSpecV | None)
  • recvbuf (BufSpec | InPlace)
  • root (int)

Request


Nonblocking send.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Nonblocking send and receive.
  • sendbuf (BufSpec)
  • dest (int)
  • sendtag (int)
  • recvbuf (BufSpec | None)
  • source (int)
  • recvtag (int)

Request


Send and receive a message.

NOTE:

This function is guaranteed not to deadlock in situations where pairs of blocking sends and receives may deadlock.


CAUTION:

A common mistake when using this function is to mismatch the tags with the source and destination ranks, which can result in deadlock.


  • buf (BufSpec)
  • dest (int)
  • sendtag (int)
  • source (int)
  • recvtag (int)

Request


Nonblocking shrink a communicator to remove all failed processes.
tuple[Comm, Request]


Nonblocking send in synchronous mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Interconnect two processes connected by a socket.
fd (int)
Intercomm


Blocking test for a matched message.
  • source (int)
  • tag (int)
  • status (Status | None)

Message


Create request for a partitioned recv operation.
  • buf (BufSpec)
  • partitions (int)
  • source (int)
  • tag (int)
  • info (Info)

Prequest


Blocking test for a message.

NOTE:

This function blocks until the message arrives.


  • source (int)
  • tag (int)
  • status (Status | None)

Literal[True]


Create request for a partitioned send operation.
  • buf (BufSpec)
  • partitions (int)
  • dest (int)
  • tag (int)
  • info (Info)

Prequest


Blocking receive.

NOTE:

This function blocks until the message is received.


  • buf (BufSpec)
  • source (int)
  • tag (int)
  • status (Status | None)

None


Create a persistent request for a receive.
  • buf (BufSpec)
  • source (int)
  • tag (int)

Prequest


Reduce to Root.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec | None)
  • op (Op)
  • root (int)

None


Persistent Reduce to Root.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec | None)
  • op (Op)
  • root (int)
  • info (Info)

Prequest


Reduce-Scatter (vector version).
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • recvcounts (Sequence[int] | None)
  • op (Op)

None


Reduce-Scatter Block (regular, non-vector version).
  • sendbuf (BufSpecB | InPlace)
  • recvbuf (BufSpec | BufSpecB)
  • op (Op)

None


Persistent Reduce-Scatter Block (regular, non-vector version).
  • sendbuf (BufSpecB | InPlace)
  • recvbuf (BufSpec | BufSpecB)
  • op (Op)
  • info (Info)

Prequest


Persistent Reduce-Scatter (vector version).
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • recvcounts (Sequence[int] | None)
  • op (Op)
  • info (Info)

Prequest


Revoke a communicator.


Blocking send in ready mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

None


Persistent request for a send in ready mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Scatter data from one process to all other processes.
  • sendbuf (BufSpecB | None)
  • recvbuf (BufSpec | InPlace)
  • root (int)

None


Persistent Scatter.
  • sendbuf (BufSpecB | None)
  • recvbuf (BufSpec | InPlace)
  • root (int)
  • info (Info)

Prequest


Scatter Vector.

Scatter data from one process to all other processes providing different amounts of data and displacements.

  • sendbuf (BufSpecV | None)
  • recvbuf (BufSpec | InPlace)
  • root (int)

None


Persistent Scatter Vector.
  • sendbuf (BufSpecV | None)
  • recvbuf (BufSpec | InPlace)
  • root (int)
  • info (Info)

Prequest


Blocking send.

NOTE:

This function may block until the message is received. Whether Send blocks or not depends on several factors and is implementation dependent.


  • buf (BufSpec)
  • dest (int)
  • tag (int)

None


Create a persistent request for a standard send.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Prequest


Send and receive a message.

NOTE:

This function is guaranteed not to deadlock in situations where pairs of blocking sends and receives may deadlock.


CAUTION:

A common mistake when using this function is to mismatch the tags with the source and destination ranks, which can result in deadlock.


  • sendbuf (BufSpec)
  • dest (int)
  • sendtag (int)
  • recvbuf (BufSpec | None)
  • source (int)
  • recvtag (int)
  • status (Status | None)

None


Send and receive a message.

NOTE:

This function is guaranteed not to deadlock in situations where pairs of blocking sends and receives may deadlock.


CAUTION:

A common mistake when using this function is to mismatch the tags with the source and destination ranks, which can result in deadlock.


  • buf (BufSpec)
  • dest (int)
  • sendtag (int)
  • source (int)
  • recvtag (int)
  • status (Status | None)

None


Store attribute value associated with a key.
  • keyval (int)
  • attrval (Any)

None


Set the error handler for a communicator.
errhandler (Errhandler)
None


Set new values for the hints associated with a communicator.
info (Info)
None


Set the print name for this communicator.
name (str)
None


Shrink a communicator to remove all failed processes.


Split communicator by color and key.
  • color (int)
  • key (int)

Comm


Split communicator by split type.
  • split_type (int)
  • key (int)
  • info (Info)

Comm


Blocking send in synchronous mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

None


Persistent request for a send in synchronous mode.
  • buf (BufSpec)
  • dest (int)
  • tag (int)

Request


Gather to All.
sendobj (Any)
list[Any]


Reduce to All.
  • sendobj (Any)
  • op (Op | Callable[[Any, Any], Any])

Any


All to All Scatter/Gather.
sendobj (Sequence[Any])
list[Any]


Barrier synchronization.

NOTE:

This method is equivalent to Barrier.




Broadcast.
  • obj (Any)
  • root (int)

Any


Send in buffered mode.
  • obj (Any)
  • dest (int)
  • tag (int)

None



Call Free if not null or predefined.


Create object from MPI handle.
handle (int)
Comm


Gather.
  • sendobj (Any)
  • root (int)

list[Any] | None


Nonblocking send in buffered mode.
  • obj (Any)
  • dest (int)
  • tag (int)

Request


Nonblocking test for a matched message.
  • source (int)
  • tag (int)
  • status (Status | None)

Message | None


Nonblocking test for a message.
  • source (int)
  • tag (int)
  • status (Status | None)

bool


Nonblocking receive.
  • buf (Buffer | None)
  • source (int)
  • tag (int)

Request


Nonblocking send.
  • obj (Any)
  • dest (int)
  • tag (int)

Request


Nonblocking send in synchronous mode.
  • obj (Any)
  • dest (int)
  • tag (int)

Request


Blocking test for a matched message.
  • source (int)
  • tag (int)
  • status (Status | None)

Message


Blocking test for a message.
  • source (int)
  • tag (int)
  • status (Status | None)

Literal[True]



Receive.
  • buf (Buffer | None)
  • source (int)
  • tag (int)
  • status (Status | None)

Any


Reduce to Root.
  • sendobj (Any)
  • op (Op | Callable[[Any, Any], Any])
  • root (int)

Any | None


Scatter.
  • sendobj (Sequence[Any] | None)
  • root (int)

Any


Send in standard mode.
  • obj (Any)
  • dest (int)
  • tag (int)

None


Send and Receive.
  • sendobj (Any)
  • dest (int)
  • sendtag (int)
  • recvbuf (Buffer | None)
  • source (int)
  • recvtag (int)
  • status (Status | None)

Any


Send in synchronous mode.
  • obj (Any)
  • dest (int)
  • tag (int)

None


Attributes Documentation

Group.

MPI handle.

Info hints.

Is intercommunicator.

Is intracommunicator.

Is a topology.

Print name.

Rank of this process.

Number of processes.

Topology type.


mpi4py.MPI.Datatype

Bases: object

Datatype object.


Methods Summary

Commit() Commit the datatype.
Create_contiguous(count) Create a contiguous datatype.
Create_darray(size, rank, gsizes, distribs, ...) Create a datatype for a distributed array on Cartesian process grids.
Create_f90_complex(p, r) Return a bounded complex datatype.
Create_f90_integer(r) Return a bounded integer datatype.
Create_f90_real(p, r) Return a bounded real datatype.
Create_hindexed(blocklengths, displacements) Create an indexed datatype.
Create_hindexed_block(blocklength, displacements) Create an indexed datatype with constant-sized blocks.
Create_hvector(count, blocklength, stride) Create a vector (strided) datatype with stride in bytes.
Create_indexed(blocklengths, displacements) Create an indexed datatype.
Create_indexed_block(blocklength, displacements) Create an indexed datatype with constant-sized blocks.
Create_keyval([copy_fn, delete_fn, nopython]) Create a new attribute key for datatypes.
Create_resized(lb, extent) Create a datatype with a new lower bound and extent.
Create_struct(blocklengths, displacements, ...) Create a general composite (struct) datatype.
Create_subarray(sizes, subsizes, starts[, order]) Create a datatype for a subarray of a multidimensional array.
Create_vector(count, blocklength, stride) Create a vector (strided) datatype.
Delete_attr(keyval) Delete attribute value associated with a key.
Dup() Duplicate a datatype.
Free() Free the datatype.
Free_keyval(keyval) Free an attribute key for datatypes.
Get_attr(keyval) Retrieve attribute value by key.
Get_contents() Return the input arguments used to create a datatype.
Get_envelope() Return the number of input arguments used to create a datatype.
Get_extent() Return lower bound and extent of datatype.
Get_name() Get the print name for this datatype.
Get_size() Return the number of bytes occupied by entries in the datatype.
Get_true_extent() Return the true lower bound and extent of a datatype.
Get_value_index(value, index) Return a predefined pair datatype.
Match_size(typeclass, size) Find a datatype matching a specified size in bytes.
Pack(inbuf, outbuf, position, comm) Pack into contiguous memory according to datatype.
Pack_external(datarep, inbuf, outbuf, position) Pack into contiguous memory according to datatype.
Pack_external_size(datarep, count) Determine the amount of space needed to pack a message.
Pack_size(count, comm) Determine the amount of space needed to pack a message.
Set_attr(keyval, attrval) Store attribute value associated with a key.
Set_name(name) Set the print name for this datatype.
Unpack(inbuf, position, outbuf, comm) Unpack from contiguous memory according to datatype.
Unpack_external(datarep, inbuf, position, outbuf) Unpack from contiguous memory according to datatype.
decode() Convenience method for decoding a datatype.
f2py(arg)
free() Call Free if not null or predefined.
fromcode(code) Get predefined MPI datatype from character code or type string.
fromhandle(handle) Create object from MPI handle.
py2f()
tocode() Get character code or type string from predefined MPI datatype.

Attributes Summary

combiner Combiner.
contents Contents.
envelope Envelope.
extent Extent.
handle MPI handle.
is_named Is a named datatype.
is_predefined Is a predefined datatype.
lb Lower bound.
name Print name.
size Size (in bytes).
true_extent True extent.
true_lb True lower bound.
true_ub True upper bound.
typechar Character code.
typestr Type string.
ub Upper bound.

Methods Documentation

Commit the datatype.


Create a contiguous datatype.
count (int)
Self


Create a datatype for a distributed array on Cartesian process grids.
  • size (int)
  • rank (int)
  • gsizes (Sequence[int])
  • distribs (Sequence[int])
  • dargs (Sequence[int])
  • psizes (Sequence[int])
  • order (int)

Self


Return a bounded complex datatype.
  • p (int)
  • r (int)

Self


Return a bounded integer datatype.
r (int)
Self


Return a bounded real datatype.
  • p (int)
  • r (int)

Self


Create an indexed datatype.

NOTE:

Displacements are measured in bytes.


  • blocklengths (Sequence[int])
  • displacements (Sequence[int])

Self


Create an indexed datatype with constant-sized blocks.

NOTE:

Displacements are measured in bytes.


  • blocklength (int)
  • displacements (Sequence[int])

Self


Create a vector (strided) datatype with stride in bytes.
  • count (int)
  • blocklength (int)
  • stride (int)

Self


Create an indexed datatype.
  • blocklengths (Sequence[int])
  • displacements (Sequence[int])

Self


Create an indexed datatype with constant-sized blocks.
  • blocklength (int)
  • displacements (Sequence[int])

Self


Create a new attribute key for datatypes.
  • copy_fn (Callable[[Datatype, int, Any], Any] | None)
  • delete_fn (Callable[[Datatype, int, Any], None] | None)
  • nopython (bool)

int


Create a datatype with a new lower bound and extent.
  • lb (int)
  • extent (int)

Self


Create a general composite (struct) datatype.

NOTE:

Displacements are measured in bytes.


  • blocklengths (Sequence[int])
  • displacements (Sequence[int])
  • datatypes (Sequence[Datatype])

Self


Create a datatype for a subarray of a multidimensional array.
  • sizes (Sequence[int])
  • subsizes (Sequence[int])
  • starts (Sequence[int])
  • order (int)

Self


Create a vector (strided) datatype.
  • count (int)
  • blocklength (int)
  • stride (int)

Self


Delete attribute value associated with a key.
keyval (int)
None


Duplicate a datatype.


Free the datatype.


Free an attribute key for datatypes.
keyval (int)
int


Retrieve attribute value by key.
keyval (int)
int | Any | None


Return the input arguments used to create a datatype.
tuple[list[int], list[int], list[int], list[Datatype]]


Return the number of input arguments used to create a datatype.
tuple[int, int, int, int, int]


Return lower bound and extent of datatype.
tuple[int, int]


Get the print name for this datatype.


Return the number of bytes occupied by entries in the datatype.


Return the true lower bound and extent of a datatype.
tuple[int, int]


Return a predefined pair datatype.
  • value (Datatype)
  • index (Datatype)

Self


Find a datatype matching a specified size in bytes.
  • typeclass (int)
  • size (int)

Self


Pack into contiguous memory according to datatype.
  • inbuf (BufSpec)
  • outbuf (BufSpec)
  • position (int)
  • comm (Comm)

int


Pack into contiguous memory according to datatype.

Uses the portable data representation external32.

  • datarep (str)
  • inbuf (BufSpec)
  • outbuf (BufSpec)
  • position (int)

int


Determine the amount of space needed to pack a message.

Uses the portable data representation external32.

NOTE:

Returns an upper bound measured in bytes.


  • datarep (str)
  • count (int)

int


Determine the amount of space needed to pack a message.

NOTE:

Returns an upper bound measured in bytes.


  • count (int)
  • comm (Comm)

int


Store attribute value associated with a key.
  • keyval (int)
  • attrval (Any)

None


Set the print name for this datatype.
name (str)
None


Unpack from contiguous memory according to datatype.
  • inbuf (BufSpec)
  • position (int)
  • outbuf (BufSpec)
  • comm (Comm)

int


Unpack from contiguous memory according to datatype.

Uses the portable data representation external32.

  • datarep (str)
  • inbuf (BufSpec)
  • position (int)
  • outbuf (BufSpec)

int


Convenience method for decoding a datatype.
tuple[Datatype, str, dict[str, Any]]



Call Free if not null or predefined.


Get predefined MPI datatype from character code or type string.
code (str)
Datatype


Create object from MPI handle.
handle (int)
Datatype



Get character code or type string from predefined MPI datatype.


Attributes Documentation

Combiner.

Contents.

Envelope.

Extent.

MPI handle.

Is a named datatype.

Is a predefined datatype.

Lower bound.

Print name.

Size (in bytes).

True extent.

True lower bound.

True upper bound.

Character code.

Type string.

Upper bound.


mpi4py.MPI.Distgraphcomm

Bases: Topocomm

Distributed graph topology intracommunicator.

comm (Distgraphcomm | None)
Self


Methods Summary

Get_dist_neighbors() Return adjacency information for a distributed graph topology.
Get_dist_neighbors_count() Return adjacency information for a distributed graph topology.

Methods Documentation

Return adjacency information for a distributed graph topology.
tuple[list[int], list[int], tuple[list[int], list[int]] | None]


Return adjacency information for a distributed graph topology.



mpi4py.MPI.Errhandler

Bases: object

Error handler.

errhandler (Errhandler | None)
Self


Methods Summary

Free() Free an error handler.
f2py(arg)
free() Call Free if not null.
fromhandle(handle) Create object from MPI handle.
py2f()

Attributes Summary

handle MPI handle.

Methods Documentation

Free an error handler.



Call Free if not null.


Create object from MPI handle.
handle (int)
Errhandler



Attributes Documentation

MPI handle.


mpi4py.MPI.File

Bases: object

File I/O context.


Methods Summary

Call_errhandler(errorcode) Call the error handler installed on a file.
Close() Close a file.
Create_errhandler(errhandler_fn) Create a new error handler for files.
Delete(filename[, info]) Delete a file.
Get_amode() Return the file access mode.
Get_atomicity() Return the atomicity mode.
Get_byte_offset(offset) Return the absolute byte position in the file.
Get_errhandler() Get the error handler for a file.
Get_group() Access the group of processes that opened the file.
Get_info() Return the current hints for a file.
Get_position() Return the current position of the individual file pointer.
Get_position_shared() Return the current position of the shared file pointer.
Get_size() Return the file size.
Get_type_extent(datatype) Return the extent of datatype in the file.
Get_view() Return the file view.
Iread(buf) Nonblocking read using individual file pointer.
Iread_all(buf) Nonblocking collective read using individual file pointer.
Iread_at(offset, buf) Nonblocking read using explicit offset.
Iread_at_all(offset, buf) Nonblocking collective read using explicit offset.
Iread_shared(buf) Nonblocking read using shared file pointer.
Iwrite(buf) Nonblocking write using individual file pointer.
Iwrite_all(buf) Nonblocking collective write using individual file pointer.
Iwrite_at(offset, buf) Nonblocking write using explicit offset.
Iwrite_at_all(offset, buf) Nonblocking collective write using explicit offset.
Iwrite_shared(buf) Nonblocking write using shared file pointer.
Open(comm, filename[, amode, info]) Open a file.
Preallocate(size) Preallocate storage space for a file.
Read(buf[, status]) Read using individual file pointer.
Read_all(buf[, status]) Collective read using individual file pointer.
Read_all_begin(buf) Start a split collective read using individual file pointer.
Read_all_end(buf[, status]) Complete a split collective read using individual file pointer.
Read_at(offset, buf[, status]) Read using explicit offset.
Read_at_all(offset, buf[, status]) Collective read using explicit offset.
Read_at_all_begin(offset, buf) Start a split collective read using explicit offset.
Read_at_all_end(buf[, status]) Complete a split collective read using explicit offset.
Read_ordered(buf[, status]) Collective read using shared file pointer.
Read_ordered_begin(buf) Start a split collective read using shared file pointer.
Read_ordered_end(buf[, status]) Complete a split collective read using shared file pointer.
Read_shared(buf[, status]) Read using shared file pointer.
Seek(offset[, whence]) Update the individual file pointer.
Seek_shared(offset[, whence]) Update the shared file pointer.
Set_atomicity(flag) Set the atomicity mode.
Set_errhandler(errhandler) Set the error handler for a file.
Set_info(info) Set new values for the hints associated with a file.
Set_size(size) Set the file size.
Set_view([disp, etype, filetype, datarep, info]) Set the file view.
Sync() Causes all previous writes to be transferred to the storage device.
Write(buf[, status]) Write using individual file pointer.
Write_all(buf[, status]) Collective write using individual file pointer.
Write_all_begin(buf) Start a split collective write using individual file pointer.
Write_all_end(buf[, status]) Complete a split collective write using individual file pointer.
Write_at(offset, buf[, status]) Write using explicit offset.
Write_at_all(offset, buf[, status]) Collective write using explicit offset.
Write_at_all_begin(offset, buf) Start a split collective write using explicit offset.
Write_at_all_end(buf[, status]) Complete a split collective write using explicit offset.
Write_ordered(buf[, status]) Collective write using shared file pointer.
Write_ordered_begin(buf) Start a split collective write using shared file pointer.
Write_ordered_end(buf[, status]) Complete a split collective write using shared file pointer.
Write_shared(buf[, status]) Write using shared file pointer.
f2py(arg)
free() Call Close if not null.
fromhandle(handle) Create object from MPI handle.
py2f()

Attributes Summary

amode Access mode.
atomicity Atomicity mode.
group Group.
group_rank Group rank.
group_size Group size.
handle MPI handle.
info Info hints.
size Size (in bytes).

Methods Documentation

Call the error handler installed on a file.
errorcode (int)
None


Close a file.


Create a new error handler for files.
errhandler_fn (Callable[[File, int], None])
Errhandler


Delete a file.
  • filename (PathLike | str | bytes)
  • info (Info)

None


Return the file access mode.


Return the atomicity mode.


Return the absolute byte position in the file.

NOTE:

Input offset is measured in etype units relative to the current file view.


offset (int)
int


Get the error handler for a file.
Errhandler


Access the group of processes that opened the file.
Group


Return the current hints for a file.


Return the current position of the individual file pointer.

NOTE:

Position is measured in etype units relative to the current file view.




Return the current position of the shared file pointer.

NOTE:

Position is measured in etype units relative to the current view.




Return the file size.


Return the extent of datatype in the file.
datatype (Datatype)
int


Return the file view.
tuple[int, Datatype, Datatype, str]


Nonblocking read using individual file pointer.
buf (BufSpec)
Request


Nonblocking collective read using individual file pointer.
buf (BufSpec)
Request


Nonblocking read using explicit offset.
  • offset (int)
  • buf (BufSpec)

Request


Nonblocking collective read using explicit offset.
  • offset (int)
  • buf (BufSpec)

Request


Nonblocking read using shared file pointer.
buf (BufSpec)
Request


Nonblocking write using individual file pointer.
buf (BufSpec)
Request


Nonblocking collective write using individual file pointer.
buf (BufSpec)
Request


Nonblocking write using explicit offset.
  • offset (int)
  • buf (BufSpec)

Request


Nonblocking collective write using explicit offset.
  • offset (int)
  • buf (BufSpec)

Request


Nonblocking write using shared file pointer.
buf (BufSpec)
Request


Open a file.
  • comm (Intracomm)
  • filename (PathLike | str | bytes)
  • amode (int)
  • info (Info)

Self


Preallocate storage space for a file.
size (int)
None


Read using individual file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Collective read using individual file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Start a split collective read using individual file pointer.
buf (BufSpec)
None


Complete a split collective read using individual file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Read using explicit offset.
  • offset (int)
  • buf (BufSpec)
  • status (Status | None)

None


Collective read using explicit offset.
  • offset (int)
  • buf (BufSpec)
  • status (Status | None)

None


Start a split collective read using explicit offset.
  • offset (int)
  • buf (BufSpec)

None


Complete a split collective read using explicit offset.
  • buf (BufSpec)
  • status (Status | None)

None


Collective read using shared file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Start a split collective read using shared file pointer.
buf (BufSpec)
None


Complete a split collective read using shared file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Read using shared file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Update the individual file pointer.
  • offset (int)
  • whence (int)

None


Update the shared file pointer.
  • offset (int)
  • whence (int)

None


Set the atomicity mode.
flag (bool)
None


Set the error handler for a file.
errhandler (Errhandler)
None


Set new values for the hints associated with a file.
info (Info)
None


Set the file size.
size (int)
None


Set the file view.
  • disp (int)
  • etype (Datatype)
  • filetype (Datatype | None)
  • datarep (str)
  • info (Info)

None


Causes all previous writes to be transferred to the storage device.


Write using individual file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Collective write using individual file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Start a split collective write using individual file pointer.
buf (BufSpec)
None


Complete a split collective write using individual file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Write using explicit offset.
  • offset (int)
  • buf (BufSpec)
  • status (Status | None)

None


Collective write using explicit offset.
  • offset (int)
  • buf (BufSpec)
  • status (Status | None)

None


Start a split collective write using explicit offset.
  • offset (int)
  • buf (BufSpec)

None


Complete a split collective write using explicit offset.
  • buf (BufSpec)
  • status (Status | None)

None


Collective write using shared file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Start a split collective write using shared file pointer.
buf (BufSpec)
None


Complete a split collective write using shared file pointer.
  • buf (BufSpec)
  • status (Status | None)

None


Write using shared file pointer.
  • buf (BufSpec)
  • status (Status | None)

None



Call Close if not null.


Create object from MPI handle.
handle (int)
File



Attributes Documentation

Access mode.

Atomicity mode.

Group.

Group rank.

Group size.

MPI handle.

Info hints.

Size (in bytes).


mpi4py.MPI.Graphcomm

Bases: Topocomm

General graph topology intracommunicator.


Methods Summary

Get_dims() Return the number of nodes and edges.
Get_neighbors(rank) Return list of neighbors of a process.
Get_neighbors_count(rank) Return number of neighbors of a process.
Get_topo() Return index and edges.

Attributes Summary

dims Number of nodes and edges.
edges Edges.
index Index.
nedges Number of edges.
neighbors Neighbors.
nneighbors Number of neighbors.
nnodes Number of nodes.
topo Topology information.

Methods Documentation

Return the number of nodes and edges.
tuple[int, int]


Return list of neighbors of a process.
rank (int)
list[int]


Return number of neighbors of a process.
rank (int)
int


Return index and edges.
tuple[list[int], list[int]]


Attributes Documentation

Number of nodes and edges.

Edges.

Index.

Number of edges.

Neighbors.

Number of neighbors.

Number of nodes.

Topology information.


mpi4py.MPI.Grequest

Bases: Request

Generalized request handler.


Methods Summary

Complete() Notify that a user-defined request is complete.
Start([query_fn, free_fn, cancel_fn, args, ...]) Create and return a user-defined request.
complete([obj]) Notify that a user-defined request is complete.

Methods Documentation

Notify that a user-defined request is complete.


Create and return a user-defined request.
  • query_fn (Callable[[...], None] | None)
  • free_fn (Callable[[...], None] | None)
  • cancel_fn (Callable[[...], None] | None)
  • args (tuple[Any] | None)
  • kwargs (dict[str, Any] | None)

Grequest


Notify that a user-defined request is complete.
obj (Any)
None



mpi4py.MPI.Group

Bases: object

Group of processes.


Methods Summary

Compare(group) Compare two groups.
Create_from_session_pset(session, pset_name) Create a new group from session and process set.
Difference(group1, group2) Create a new group from the difference of two existing groups.
Dup() Duplicate a group.
Excl(ranks) Create a new group by excluding listed members.
Free() Free a group.
Get_rank() Return the rank of this process in a group.
Get_size() Return the number of processes in a group.
Incl(ranks) Create a new group by including listed members.
Intersection(group1, group2) Create a new group from the intersection of two existing groups.
Range_excl(ranks) Create a new group by excluding ranges of members.
Range_incl(ranks) Create a new group by including ranges of members.
Translate_ranks([ranks, group]) Translate ranks in a group to those in another group.
Union(group1, group2) Create a new group from the union of two existing groups.
f2py(arg)
free() Call Free if not null or predefined.
fromhandle(handle) Create object from MPI handle.
py2f()

Attributes Summary

handle MPI handle.
rank Rank of this process.
size Number of processes.

Methods Documentation

Compare two groups.
group (Group)
int


Create a new group from session and process set.
  • session (Session)
  • pset_name (str)

Self


Create a new group from the difference of two existing groups.
  • group1 (Group)
  • group2 (Group)

Self


Duplicate a group.


Create a new group by excluding listed members.
ranks (Sequence[int])
Self


Free a group.


Return the rank of this process in a group.


Return the number of processes in a group.


Create a new group by including listed members.
ranks (Sequence[int])
Self


Create a new group from the intersection of two existing groups.
  • group1 (Group)
  • group2 (Group)

Self


Create a new group by excluding ranges of members.
ranks (Sequence[tuple[int, int, int]])
Self


Create a new group by including ranges of members.
ranks (Sequence[tuple[int, int, int]])
Self


Translate ranks in a group to those in another group.
  • ranks (Sequence[int] | None)
  • group (Group | None)

list[int]


Create a new group from the union of two existing groups.
  • group1 (Group)
  • group2 (Group)

Self



Call Free if not null or predefined.


Create object from MPI handle.
handle (int)
Group



Attributes Documentation

MPI handle.

Rank of this process.

Number of processes.


mpi4py.MPI.InPlaceType


mpi4py.MPI.Info

Bases: object

Info object.


Methods Summary

Create([items]) Create a new info object.
Create_env([args]) Create a new environment info object.
Delete(key) Remove a (key, value) pair from info.
Dup() Duplicate an existing info object.
Free() Free an info object.
Get(key) Retrieve the value associated with a key.
Get_nkeys() Return the number of currently defined keys in info.
Get_nthkey(n) Return the n-th defined key in info.
Set(key, value) Store a value associated with a key.
clear() Clear contents.
copy() Copy contents.
f2py(arg)
free() Call Free if not null or predefined.
fromhandle(handle) Create object from MPI handle.
get(key[, default]) Retrieve value by key.
items() Return list of items.
keys() Return list of keys.
pop(key, *default) Pop value by key.
popitem() Pop first item.
py2f()
update([items]) Update contents.
values() Return list of values.

Attributes Summary

handle MPI handle.

Methods Documentation

Create a new info object.
items (Info | Mapping[str, str] | Iterable[tuple[str, str]] | None)
Self


Create a new environment info object.
args (Sequence[str] | None)
Self


Remove a (key, value) pair from info.
key (str)
None


Duplicate an existing info object.


Free an info object.


Retrieve the value associated with a key.
key (str)
str | None


Return the number of currently defined keys in info.


Return the n-th defined key in info.
n (int)
str


Store a value associated with a key.
  • key (str)
  • value (str)

None


Clear contents.


Copy contents.



Call Free if not null or predefined.


Create object from MPI handle.
handle (int)
Info


Retrieve value by key.
  • key (str)
  • default (str | None)

str | None


Return list of items.
list[tuple[str, str]]


Return list of keys.
list[str]


Pop value by key.
  • key (str)
  • default (str)

str


Pop first item.
tuple[str, str]



Update contents.
  • items (Info | Mapping[str, str] | Iterable[tuple[str, str]])
  • kwds (str)

None


Return list of values.
list[str]


Attributes Documentation

MPI handle.


mpi4py.MPI.Intercomm

Bases: Comm

Intercommunicator.


Methods Summary

Create_from_groups(local_group, ...[, ...]) Create communicator from group.
Get_remote_group() Access the remote group associated with the inter-communicator.
Get_remote_size() Intercommunicator remote size.
Merge([high]) Merge intercommunicator into an intracommunicator.

Attributes Summary

remote_group Remote group.
remote_size Number of remote processes.

Methods Documentation

Create communicator from group.
  • local_group (Group)
  • local_leader (int)
  • remote_group (Group)
  • remote_leader (int)
  • stringtag (str)
  • info (Info)
  • errhandler (Errhandler | None)

Intracomm


Access the remote group associated with the inter-communicator.
Group


Intercommunicator remote size.


Merge intercommunicator into an intracommunicator.
high (bool)
Intracomm


Attributes Documentation

Remote group.

Number of remote processes.


mpi4py.MPI.Intracomm

Bases: Comm

Intracommunicator.


Methods Summary

Accept(port_name[, info, root]) Accept a request to form a new intercommunicator.
Cart_map(dims[, periods]) Determine optimal process placement on a Cartesian topology.
Connect(port_name[, info, root]) Make a request to form a new intercommunicator.
Create_cart(dims[, periods, reorder]) Create cartesian communicator.
Create_dist_graph(sources, degrees, destinations) Create distributed graph communicator.
Create_dist_graph_adjacent(sources, destinations) Create distributed graph communicator.
Create_from_group(group[, stringtag, info, ...]) Create communicator from group.
Create_graph(index, edges[, reorder]) Create graph communicator.
Create_group(group[, tag]) Create communicator from group.
Create_intercomm(local_leader, peer_comm, ...) Create intercommunicator.
Exscan(sendbuf, recvbuf[, op]) Exclusive Scan.
Exscan_init(sendbuf, recvbuf[, op, info]) Persistent Exclusive Scan.
Graph_map(index, edges) Determine optimal process placement on a graph topology.
Iexscan(sendbuf, recvbuf[, op]) Inclusive Scan.
Iscan(sendbuf, recvbuf[, op]) Inclusive Scan.
Scan(sendbuf, recvbuf[, op]) Inclusive Scan.
Scan_init(sendbuf, recvbuf[, op, info]) Persistent Inclusive Scan.
Spawn(command[, args, maxprocs, info, root, ...]) Spawn instances of a single MPI application.
Spawn_multiple(command[, args, maxprocs, ...]) Spawn instances of multiple MPI applications.
exscan(sendobj[, op]) Exclusive Scan.
scan(sendobj[, op]) Inclusive Scan.

Methods Documentation

Accept a request to form a new intercommunicator.
  • port_name (str)
  • info (Info)
  • root (int)

Intercomm


Determine optimal process placement on a Cartesian topology.
  • dims (Sequence[int])
  • periods (Sequence[bool] | None)

int


Make a request to form a new intercommunicator.
  • port_name (str)
  • info (Info)
  • root (int)

Intercomm


Create cartesian communicator.
  • dims (Sequence[int])
  • periods (Sequence[bool] | None)
  • reorder (bool)

Cartcomm


Create distributed graph communicator.
  • sources (Sequence[int])
  • degrees (Sequence[int])
  • destinations (Sequence[int])
  • weights (Sequence[int] | None)
  • info (Info)
  • reorder (bool)

Distgraphcomm


Create distributed graph communicator.
  • sources (Sequence[int])
  • destinations (Sequence[int])
  • sourceweights (Sequence[int] | None)
  • destweights (Sequence[int] | None)
  • info (Info)
  • reorder (bool)

Distgraphcomm


Create communicator from group.
  • group (Group)
  • stringtag (str)
  • info (Info)
  • errhandler (Errhandler | None)

Intracomm


Create graph communicator.
  • index (Sequence[int])
  • edges (Sequence[int])
  • reorder (bool)

Graphcomm


Create communicator from group.
  • group (Group)
  • tag (int)

Intracomm


Create intercommunicator.
  • local_leader (int)
  • peer_comm (Intracomm)
  • remote_leader (int)
  • tag (int)

Intercomm


Exclusive Scan.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)

None


Persistent Exclusive Scan.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)
  • info (Info)

Prequest


Determine optimal process placement on a graph topology.
  • index (Sequence[int])
  • edges (Sequence[int])

int


Inclusive Scan.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)

Request


Inclusive Scan.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)

Request


Inclusive Scan.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)

None


Persistent Inclusive Scan.
  • sendbuf (BufSpec | InPlace)
  • recvbuf (BufSpec)
  • op (Op)
  • info (Info)

Prequest


Spawn instances of a single MPI application.
  • command (str)
  • args (Sequence[str] | None)
  • maxprocs (int)
  • info (Info)
  • root (int)
  • errcodes (list[int] | None)

Intercomm


Spawn instances of multiple MPI applications.
  • command (Sequence[str])
  • args (Sequence[Sequence[str]] | None)
  • maxprocs (Sequence[int] | None)
  • info (Sequence[Info] | Info)
  • root (int)
  • errcodes (list[list[int]] | None)

Intercomm


Exclusive Scan.
  • sendobj (Any)
  • op (Op | Callable[[Any, Any], Any])

Any


Inclusive Scan.
  • sendobj (Any)
  • op (Op | Callable[[Any, Any], Any])

Any



mpi4py.MPI.Message

Bases: object

Matched message.


Methods Summary

Iprobe(comm[, source, tag, status]) Nonblocking test for a matched message.
Irecv(buf) Nonblocking receive of matched message.
Probe(comm[, source, tag, status]) Blocking test for a matched message.
Recv(buf[, status]) Blocking receive of matched message.
f2py(arg)
free() Do nothing.
fromhandle(handle) Create object from MPI handle.
iprobe(comm[, source, tag, status]) Nonblocking test for a matched message.
irecv() Nonblocking receive of matched message.
probe(comm[, source, tag, status]) Blocking test for a matched message.
py2f()
recv([status]) Blocking receive of matched message.

Attributes Summary

handle MPI handle.

Methods Documentation

Nonblocking test for a matched message.
  • comm (Comm)
  • source (int)
  • tag (int)
  • status (Status | None)

Self | None


Nonblocking receive of matched message.
buf (BufSpec)
Request


Blocking test for a matched message.
  • comm (Comm)
  • source (int)
  • tag (int)
  • status (Status | None)

Self


Blocking receive of matched message.
  • buf (BufSpec)
  • status (Status | None)

None



Do nothing.


Create object from MPI handle.
handle (int)
Message


Nonblocking test for a matched message.
  • comm (Comm)
  • source (int)
  • tag (int)
  • status (Status | None)

Self | None


Nonblocking receive of matched message.
Request


Blocking test for a matched message.
  • comm (Comm)
  • source (int)
  • tag (int)
  • status (Status | None)

Self



Blocking receive of matched message.
status (Status | None)
Any


Attributes Documentation

MPI handle.


mpi4py.MPI.Op

Bases: object

Reduction operation.


Methods Summary

Create(function[, commute]) Create a user-defined reduction operation.
Free() Free a user-defined reduction operation.
Is_commutative() Query reduction operations for their commutativity.
Reduce_local(inbuf, inoutbuf) Apply a reduction operation to local data.
f2py(arg)
free() Call Free if not null or predefined.
fromhandle(handle) Create object from MPI handle.
py2f()

Attributes Summary

handle MPI handle.
is_commutative Is a commutative operation.
is_predefined Is a predefined operation.

Methods Documentation

Create a user-defined reduction operation.
  • function (Callable[[Buffer, Buffer, Datatype], None])
  • commute (bool)

Self


Free a user-defined reduction operation.


Query reduction operations for their commutativity.


Apply a reduction operation to local data.
  • inbuf (BufSpec)
  • inoutbuf (BufSpec)

None



Call Free if not null or predefined.


Create object from MPI handle.
handle (int)
Op



Attributes Documentation

MPI handle.

Is a commutative operation.

Is a predefined operation.


mpi4py.MPI.Pickle

Bases: object

Pickle/unpickle Python objects.


Methods Summary

dumps(obj) Serialize object to pickle data stream.
dumps_oob(obj) Serialize object to pickle data stream and out-of-band buffers.
loads(data) Deserialize object from pickle data stream.
loads_oob(data, buffers) Deserialize object from pickle data stream and out-of-band buffers.

Attributes Summary

PROTOCOL Protocol version.
THRESHOLD Out-of-band threshold.

Methods Documentation

Serialize object to pickle data stream.
obj (Any)
bytes


Serialize object to pickle data stream and out-of-band buffers.
obj (Any)
tuple[bytes, list[buffer]]


Deserialize object from pickle data stream.
data (Buffer)
Any


Deserialize object from pickle data stream and out-of-band buffers.
  • data (Buffer)
  • buffers (Iterable[Buffer])

Any


Attributes Documentation

Protocol version.

Out-of-band threshold.


mpi4py.MPI.Prequest

Bases: Request

Persistent request handler.


Methods Summary

Parrived(partition) Test partial completion of a partitioned receive operation.
Pready(partition) Mark a given partition as ready.
Pready_list(partitions) Mark a sequence of partitions as ready.
Pready_range(partition_low, partition_high) Mark a range of partitions as ready.
Start() Initiate a communication with a persistent request.
Startall(requests) Start a collection of persistent requests.

Methods Documentation

Test partial completion of a partitioned receive operation.
partition (int)
bool


Mark a given partition as ready.
partition (int)
None


Mark a sequence of partitions as ready.
partitions (Sequence[int])
None


Mark a range of partitions as ready.
  • partition_low (int)
  • partition_high (int)

None


Initiate a communication with a persistent request.


Start a collection of persistent requests.
requests (list[Prequest])
None



mpi4py.MPI.Request

Bases: object

Request handler.


Methods Summary

Cancel() Cancel a request.
Free() Free a communication request.
Get_status([status]) Non-destructive test for the completion of a request.
Get_status_all(requests[, statuses]) Non-destructive test for the completion of all requests.
Get_status_any(requests[, status]) Non-destructive test for the completion of any requests.
Get_status_some(requests[, statuses]) Non-destructive test for completion of some requests.
Test([status]) Test for the completion of a non-blocking operation.
Testall(requests[, statuses]) Test for completion of all previously initiated requests.
Testany(requests[, status]) Test for completion of any previously initiated request.
Testsome(requests[, statuses]) Test for completion of some previously initiated requests.
Wait([status]) Wait for a non-blocking operation to complete.
Waitall(requests[, statuses]) Wait for all previously initiated requests to complete.
Waitany(requests[, status]) Wait for any previously initiated request to complete.
Waitsome(requests[, statuses]) Wait for some previously initiated requests to complete.
cancel() Cancel a request.
f2py(arg)
free() Call Free if not null.
fromhandle(handle) Create object from MPI handle.
get_status([status]) Non-destructive test for the completion of a request.
get_status_all(requests[, statuses]) Non-destructive test for the completion of all requests.
get_status_any(requests[, status]) Non-destructive test for the completion of any requests.
get_status_some(requests[, statuses]) Non-destructive test for completion of some requests.
py2f()
test([status]) Test for the completion of a non-blocking operation.
testall(requests[, statuses]) Test for completion of all previously initiated requests.
testany(requests[, status]) Test for completion of any previously initiated request.
testsome(requests[, statuses]) Test for completion of some previously initiated requests.
wait([status]) Wait for a non-blocking operation to complete.
waitall(requests[, statuses]) Wait for all previously initiated requests to complete.
waitany(requests[, status]) Wait for any previously initiated request to complete.
waitsome(requests[, statuses]) Wait for some previously initiated requests to complete.

Attributes Summary

handle MPI handle.

Methods Documentation

Cancel a request.


Free a communication request.


Non-destructive test for the completion of a request.
status (Status | None)
bool


Non-destructive test for the completion of all requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

bool


Non-destructive test for the completion of any requests.
  • requests (Sequence[Request])
  • status (Status | None)

tuple[int, bool]


Non-destructive test for completion of some requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

list[int] | None


Test for the completion of a non-blocking operation.
status (Status | None)
bool


Test for completion of all previously initiated requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

bool


Test for completion of any previously initiated request.
  • requests (Sequence[Request])
  • status (Status | None)

tuple[int, bool]


Test for completion of some previously initiated requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

list[int] | None


Wait for a non-blocking operation to complete.
status (Status | None)
Literal[True]


Wait for all previously initiated requests to complete.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

Literal[True]


Wait for any previously initiated request to complete.
  • requests (Sequence[Request])
  • status (Status | None)

int


Wait for some previously initiated requests to complete.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

list[int] | None


Cancel a request.



Call Free if not null.


Create object from MPI handle.
handle (int)
Request


Non-destructive test for the completion of a request.
status (Status | None)
bool


Non-destructive test for the completion of all requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

bool


Non-destructive test for the completion of any requests.
  • requests (Sequence[Request])
  • status (Status | None)

tuple[int, bool]


Non-destructive test for completion of some requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

list[int] | None



Test for the completion of a non-blocking operation.
status (Status | None)
tuple[bool, Any | None]


Test for completion of all previously initiated requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

tuple[bool, list[Any] | None]


Test for completion of any previously initiated request.
  • requests (Sequence[Request])
  • status (Status | None)

tuple[int, bool, Any | None]


Test for completion of some previously initiated requests.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

tuple[list[int] | None, list[Any] | None]


Wait for a non-blocking operation to complete.
status (Status | None)
Any


Wait for all previously initiated requests to complete.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

list[Any]


Wait for any previously initiated request to complete.
  • requests (Sequence[Request])
  • status (Status | None)

tuple[int, Any]


Wait for some previously initiated requests to complete.
  • requests (Sequence[Request])
  • statuses (list[Status] | None)

tuple[list[int] | None, list[Any] | None]


Attributes Documentation

MPI handle.


mpi4py.MPI.Session

Bases: object

Session context.


Methods Summary

Attach_buffer(buf) Attach a user-provided buffer for sending in buffered mode.
Call_errhandler(errorcode) Call the error handler installed on a session.
Create_errhandler(errhandler_fn) Create a new error handler for sessions.
Create_group(pset_name) Create a new group from session and process set.
Detach_buffer() Remove an existing attached buffer.
Finalize() Finalize a session.
Flush_buffer() Block until all buffered messages have been transmitted.
Get_errhandler() Get the error handler for a session.
Get_info() Return the current hints for a session.
Get_nth_pset(n[, info]) Name of the n-th process set.
Get_num_psets([info]) Number of available process sets.
Get_pset_info(pset_name) Return the current hints for a session and process set.
Iflush_buffer() Nonblocking flush for buffered messages.
Init([info, errhandler]) Create a new session.
Set_errhandler(errhandler) Set the error handler for a session.
f2py(arg)
free() Call Finalize if not null.
fromhandle(handle) Create object from MPI handle.
py2f()

Attributes Summary

handle MPI handle.

Methods Documentation

Attach a user-provided buffer for sending in buffered mode.
buf (Buffer | None)
None


Call the error handler installed on a session.
errorcode (int)
None


Create a new error handler for sessions.
errhandler_fn (Callable[[Session, int], None])
Errhandler


Create a new group from session and process set.
pset_name (str)
Group


Remove an existing attached buffer.
Buffer | None


Finalize a session.


Block until all buffered messages have been transmitted.


Get the error handler for a session.
Errhandler


Return the current hints for a session.


Name of the n-th process set.
  • n (int)
  • info (Info)

str


Number of available process sets.
info (Info)
int


Return the current hints for a session and process set.
pset_name (str)
Info


Nonblocking flush for buffered messages.
Request


Create a new session.
  • info (Info)
  • errhandler (Errhandler | None)

Self


Set the error handler for a session.
errhandler (Errhandler)
None



Call Finalize if not null.


Create object from MPI handle.
handle (int)
Session



Attributes Documentation

MPI handle.


mpi4py.MPI.Status

Bases: object

Status object.


Methods Summary

Get_count([datatype]) Get the number of top level elements.
Get_elements(datatype) Get the number of basic elements in a datatype.
Get_error() Get message error.
Get_source() Get message source.
Get_tag() Get message tag.
Is_cancelled() Test to see if a request was cancelled.
Set_cancelled(flag) Set the cancelled state associated with a status.
Set_elements(datatype, count) Set the number of elements in a status.
Set_error(error) Set message error.
Set_source(source) Set message source.
Set_tag(tag) Set message tag.
f2py(arg)
py2f()

Attributes Summary

cancelled Cancelled state.
count Byte count.
error Message error.
source Message source.
tag Message tag.

Methods Documentation

Get the number of top level elements.
datatype (Datatype)
int


Get the number of basic elements in a datatype.
datatype (Datatype)
int


Get message error.


Get message source.


Get message tag.


Test to see if a request was cancelled.


Set the cancelled state associated with a status.

NOTE:

This method should be used only when implementing query callback functions for generalized requests.


flag (bool)
None


Set the number of elements in a status.

NOTE:

This method should be only used when implementing query callback functions for generalized requests.


  • datatype (Datatype)
  • count (int)

None


Set message error.
error (int)
None


Set message source.
source (int)
None


Set message tag.
tag (int)
None



list[int]


Attributes Documentation

Cancelled state.

Byte count.

Message error.

Message source.

Message tag.


mpi4py.MPI.Topocomm

Bases: Intracomm

Topology intracommunicator.


Methods Summary

Ineighbor_allgather(sendbuf, recvbuf) Nonblocking Neighbor Gather to All.
Ineighbor_allgatherv(sendbuf, recvbuf) Nonblocking Neighbor Gather to All Vector.
Ineighbor_alltoall(sendbuf, recvbuf) Nonblocking Neighbor All to All.
Ineighbor_alltoallv(sendbuf, recvbuf) Nonblocking Neighbor All to All Vector.
Ineighbor_alltoallw(sendbuf, recvbuf) Nonblocking Neighbor All to All General.
Neighbor_allgather(sendbuf, recvbuf) Neighbor Gather to All.
Neighbor_allgather_init(sendbuf, recvbuf[, info]) Persistent Neighbor Gather to All.
Neighbor_allgatherv(sendbuf, recvbuf) Neighbor Gather to All Vector.
Neighbor_allgatherv_init(sendbuf, recvbuf[, ...]) Persistent Neighbor Gather to All Vector.
Neighbor_alltoall(sendbuf, recvbuf) Neighbor All to All.
Neighbor_alltoall_init(sendbuf, recvbuf[, info]) Persistent Neighbor All to All.
Neighbor_alltoallv(sendbuf, recvbuf) Neighbor All to All Vector.
Neighbor_alltoallv_init(sendbuf, recvbuf[, info]) Persistent Neighbor All to All Vector.
Neighbor_alltoallw(sendbuf, recvbuf) Neighbor All to All General.
Neighbor_alltoallw_init(sendbuf, recvbuf[, info]) Persistent Neighbor All to All General.
neighbor_allgather(sendobj) Neighbor Gather to All.
neighbor_alltoall(sendobj) Neighbor All to All.

Attributes Summary

degrees Number of incoming and outgoing neighbors.
indegree Number of incoming neighbors.
inedges Incoming neighbors.
inoutedges Incoming and outgoing neighbors.
outdegree Number of outgoing neighbors.
outedges Outgoing neighbors.

Methods Documentation

Nonblocking Neighbor Gather to All.
  • sendbuf (BufSpec)
  • recvbuf (BufSpecB)

Request


Nonblocking Neighbor Gather to All Vector.
  • sendbuf (BufSpec)
  • recvbuf (BufSpecV)

Request


Nonblocking Neighbor All to All.
  • sendbuf (BufSpecB)
  • recvbuf (BufSpecB)

Request


Nonblocking Neighbor All to All Vector.
  • sendbuf (BufSpecV)
  • recvbuf (BufSpecV)

Request


Nonblocking Neighbor All to All General.
  • sendbuf (BufSpecW)
  • recvbuf (BufSpecW)

Request


Neighbor Gather to All.
  • sendbuf (BufSpec)
  • recvbuf (BufSpecB)

None


Persistent Neighbor Gather to All.
  • sendbuf (BufSpec)
  • recvbuf (BufSpecB)
  • info (Info)

Prequest


Neighbor Gather to All Vector.
  • sendbuf (BufSpec)
  • recvbuf (BufSpecV)

None


Persistent Neighbor Gather to All Vector.
  • sendbuf (BufSpec)
  • recvbuf (BufSpecV)
  • info (Info)

Prequest


Neighbor All to All.
  • sendbuf (BufSpecB)
  • recvbuf (BufSpecB)

None


Persistent Neighbor All to All.
  • sendbuf (BufSpecB)
  • recvbuf (BufSpecB)
  • info (Info)

Prequest


Neighbor All to All Vector.
  • sendbuf (BufSpecV)
  • recvbuf (BufSpecV)

None


Persistent Neighbor All to All Vector.
  • sendbuf (BufSpecV)
  • recvbuf (BufSpecV)
  • info (Info)

Prequest


Neighbor All to All General.
  • sendbuf (BufSpecW)
  • recvbuf (BufSpecW)

None


Persistent Neighbor All to All General.
  • sendbuf (BufSpecW)
  • recvbuf (BufSpecW)
  • info (Info)

Prequest


Neighbor Gather to All.
sendobj (Any)
list[Any]


Neighbor All to All.
sendobj (list[Any])
list[Any]


Attributes Documentation

Number of incoming and outgoing neighbors.

Number of incoming neighbors.

Incoming neighbors.

Incoming and outgoing neighbors.

Number of outgoing neighbors.

Outgoing neighbors.


mpi4py.MPI.Win

Bases: object

Remote memory access context.


Methods Summary

Accumulate(origin, target_rank[, target, op]) Accumulate data into the target process.
Allocate(size[, disp_unit, info, comm]) Create an window object for one-sided communication.
Allocate_shared(size[, disp_unit, info, comm]) Create an window object for one-sided communication.
Attach(memory) Attach a local memory region.
Call_errhandler(errorcode) Call the error handler installed on a window.
Compare_and_swap(origin, compare, result, ...) Perform one-sided atomic compare-and-swap.
Complete() Complete an RMA operation begun after an Start.
Create(memory[, disp_unit, info, comm]) Create an window object for one-sided communication.
Create_dynamic([info, comm]) Create an window object for one-sided communication.
Create_errhandler(errhandler_fn) Create a new error handler for windows.
Create_keyval([copy_fn, delete_fn, nopython]) Create a new attribute key for windows.
Delete_attr(keyval) Delete attribute value associated with a key.
Detach(memory) Detach a local memory region.
Fence([assertion]) Perform an MPI fence synchronization on a window.
Fetch_and_op(origin, result, target_rank[, ...]) Perform one-sided read-modify-write.
Flush(rank) Complete all outstanding RMA operations at a target.
Flush_all() Complete all outstanding RMA operations at all targets.
Flush_local(rank) Complete locally all outstanding RMA operations at a target.
Flush_local_all() Complete locally all outstanding RMA operations at all targets.
Free() Free a window.
Free_keyval(keyval) Free an attribute key for windows.
Get(origin, target_rank[, target]) Get data from a memory window on a remote process.
Get_accumulate(origin, result, target_rank) Fetch-and-accumulate data into the target process.
Get_attr(keyval) Retrieve attribute value by key.
Get_errhandler() Get the error handler for a window.
Get_group() Access the group of processes that created the window.
Get_info() Return the current hints for a window.
Get_name() Get the print name for this window.
Lock(rank[, lock_type, assertion]) Begin an RMA access epoch at the target process.
Lock_all([assertion]) Begin an RMA access epoch at all processes.
Post(group[, assertion]) Start an RMA exposure epoch.
Put(origin, target_rank[, target]) Put data into a memory window on a remote process.
Raccumulate(origin, target_rank[, target, op]) Fetch-and-accumulate data into the target process.
Rget(origin, target_rank[, target]) Get data from a memory window on a remote process.
Rget_accumulate(origin, result, target_rank) Accumulate data into the target process using remote memory access.
Rput(origin, target_rank[, target]) Put data into a memory window on a remote process.
Set_attr(keyval, attrval) Store attribute value associated with a key.
Set_errhandler(errhandler) Set the error handler for a window.
Set_info(info) Set new values for the hints associated with a window.
Set_name(name) Set the print name for this window.
Shared_query(rank) Query the process-local address for remote memory segments.
Start(group[, assertion]) Start an RMA access epoch for MPI.
Sync() Synchronize public and private copies of the window.
Test() Test whether an RMA exposure epoch has completed.
Unlock(rank) Complete an RMA access epoch at the target process.
Unlock_all() Complete an RMA access epoch at all processes.
Wait() Complete an RMA exposure epoch begun with Post.
f2py(arg)
free() Call Free if not null.
fromhandle(handle) Create object from MPI handle.
py2f()
tomemory() Return window memory buffer.

Attributes Summary

attrs Attributes.
flavor Create flavor.
group Group.
group_rank Group rank.
group_size Group size.
handle MPI handle.
info Info hints.
model Memory model.
name Print name.

Methods Documentation

Accumulate data into the target process.
  • origin (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)
  • op (Op)

None


Create an window object for one-sided communication.
  • size (int)
  • disp_unit (int)
  • info (Info)
  • comm (Intracomm)

Self


Create an window object for one-sided communication.
  • size (int)
  • disp_unit (int)
  • info (Info)
  • comm (Intracomm)

Self


Attach a local memory region.
memory (Buffer)
None


Call the error handler installed on a window.
errorcode (int)
None


Perform one-sided atomic compare-and-swap.
  • origin (BufSpec)
  • compare (BufSpec)
  • result (BufSpec)
  • target_rank (int)
  • target_disp (int)

None


Complete an RMA operation begun after an Start.


Create an window object for one-sided communication.
  • memory (Buffer | Bottom)
  • disp_unit (int)
  • info (Info)
  • comm (Intracomm)

Self


Create an window object for one-sided communication.
  • info (Info)
  • comm (Intracomm)

Self


Create a new error handler for windows.
errhandler_fn (Callable[[Win, int], None])
Errhandler


Create a new attribute key for windows.
  • copy_fn (Callable[[Win, int, Any], Any] | None)
  • delete_fn (Callable[[Win, int, Any], None] | None)
  • nopython (bool)

int


Delete attribute value associated with a key.
keyval (int)
None


Detach a local memory region.
memory (Buffer)
None


Perform an MPI fence synchronization on a window.
assertion (int)
None


Perform one-sided read-modify-write.
  • origin (BufSpec)
  • result (BufSpec)
  • target_rank (int)
  • target_disp (int)
  • op (Op)

None


Complete all outstanding RMA operations at a target.
rank (int)
None


Complete all outstanding RMA operations at all targets.


Complete locally all outstanding RMA operations at a target.
rank (int)
None


Complete locally all outstanding RMA operations at all targets.


Free a window.


Free an attribute key for windows.
keyval (int)
int


Get data from a memory window on a remote process.
  • origin (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)

None


Fetch-and-accumulate data into the target process.
  • origin (BufSpec)
  • result (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)
  • op (Op)

None


Retrieve attribute value by key.
keyval (int)
int | Any | None


Get the error handler for a window.
Errhandler


Access the group of processes that created the window.
Group


Return the current hints for a window.


Get the print name for this window.


Begin an RMA access epoch at the target process.
  • rank (int)
  • lock_type (int)
  • assertion (int)

None


Begin an RMA access epoch at all processes.
assertion (int)
None


Start an RMA exposure epoch.
  • group (Group)
  • assertion (int)

None


Put data into a memory window on a remote process.
  • origin (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)

None


Fetch-and-accumulate data into the target process.
  • origin (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)
  • op (Op)

Request


Get data from a memory window on a remote process.
  • origin (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)

Request


Accumulate data into the target process using remote memory access.
  • origin (BufSpec)
  • result (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)
  • op (Op)

Request


Put data into a memory window on a remote process.
  • origin (BufSpec)
  • target_rank (int)
  • target (TargetSpec | None)

Request


Store attribute value associated with a key.
  • keyval (int)
  • attrval (Any)

None


Set the error handler for a window.
errhandler (Errhandler)
None


Set new values for the hints associated with a window.
info (Info)
None


Set the print name for this window.
name (str)
None


Query the process-local address for remote memory segments.
rank (int)
tuple[buffer, int]


Start an RMA access epoch for MPI.
  • group (Group)
  • assertion (int)

None


Synchronize public and private copies of the window.


Test whether an RMA exposure epoch has completed.


Complete an RMA access epoch at the target process.
rank (int)
None


Complete an RMA access epoch at all processes.


Complete an RMA exposure epoch begun with Post.
Literal[True]



Call Free if not null.


Create object from MPI handle.
handle (int)
Win



Return window memory buffer.
buffer


Attributes Documentation

Attributes.

Create flavor.

Group.

Group rank.

Group size.

MPI handle.

Info hints.

Memory model.

Print name.


mpi4py.MPI.buffer

Bases: object

Buffer.


Methods Summary

allocate(nbytes[, clear]) Buffer allocation.
cast(format[, shape]) Cast to a memoryview with new format or shape.
fromaddress(address, nbytes[, readonly]) Buffer from address and size in bytes.
frombuffer(obj[, readonly]) Buffer from buffer-like object.
release() Release the underlying buffer exposed by the buffer object.
tobytes([order]) Return the data in the buffer as a byte string.
toreadonly() Return a readonly version of the buffer object.

Attributes Summary

address Buffer address.
format Format of each element.
itemsize Size (in bytes) of each element.
nbytes Buffer size (in bytes).
obj Object exposing buffer.
readonly Buffer is read-only.

Methods Documentation

Buffer allocation.
  • nbytes (int)
  • clear (bool)

buffer


Cast to a memoryview with new format or shape.
  • format (str)
  • shape (list[int] | tuple[int, ...])

memoryview


Buffer from address and size in bytes.
  • address (int)
  • nbytes (int)
  • readonly (bool)

buffer


Buffer from buffer-like object.
  • obj (Buffer)
  • readonly (bool)

buffer


Release the underlying buffer exposed by the buffer object.


Return the data in the buffer as a byte string.
order (str | None)
bytes


Return a readonly version of the buffer object.
buffer


Attributes Documentation

Buffer address.

Format of each element.

Size (in bytes) of each element.

Buffer size (in bytes).

Object exposing buffer.

Buffer is read-only.


mpi4py.MPI.memory

alias of buffer

Exceptions

Exception Exception class.

mpi4py.MPI.Exception

Bases: RuntimeError

Exception class.


Methods Summary

Get_error_class() Error class.
Get_error_code() Error code.
Get_error_string() Error string.

Attributes Summary

error_class Error class.
error_code Error code.
error_string Error string.

Methods Documentation

Error class.


Error code.


Error string.


Attributes Documentation

Error class.

Error code.

Error string.


Functions

Add_error_class() Add an error class to the known error classes.
Add_error_code(errorclass) Add an error code to an error class.
Add_error_string(errorcode, string) Associate an error string with an error class or error code.
Aint_add(base, disp) Return the sum of base address and displacement.
Aint_diff(addr1, addr2) Return the difference between absolute addresses.
Alloc_mem(size[, info]) Allocate memory for message passing and remote memory access.
Attach_buffer(buf) Attach a user-provided buffer for sending in buffered mode.
Close_port(port_name) Close a port.
Compute_dims(nnodes, dims) Return a balanced distribution of processes per coordinate direction.
Detach_buffer() Remove an existing attached buffer.
Finalize() Terminate the MPI execution environment.
Flush_buffer() Block until all buffered messages have been transmitted.
Free_mem(mem) Free memory allocated with Alloc_mem.
Get_address(location) Get the address of a location in memory.
Get_error_class(errorcode) Convert an error code into an error class.
Get_error_string(errorcode) Return the error string for a given error class or error code.
Get_hw_resource_info() Obtain information about the hardware platform of the calling processor.
Get_library_version() Obtain the version string of the MPI library.
Get_processor_name() Obtain the name of the calling processor.
Get_version() Obtain the version number of the MPI standard.
Iflush_buffer() Nonblocking flush for buffered messages.
Init() Initialize the MPI execution environment.
Init_thread([required]) Initialize the MPI execution environment.
Is_finalized() Indicate whether Finalize has completed.
Is_initialized() Indicate whether Init has been called.
Is_thread_main() Indicate whether this thread called Init or Init_thread.
Lookup_name(service_name[, info]) Lookup a port name given a service name.
Open_port([info]) Return an address used to connect group of processes.
Pcontrol(level) Control profiling.
Publish_name(service_name, port_name[, info]) Publish a service name.
Query_thread() Return the level of thread support provided by the MPI library.
Register_datarep(datarep, read_fn, write_fn, ...) Register user-defined data representations.
Remove_error_class(errorclass) Remove an error class from the known error classes.
Remove_error_code(errorcode) Remove an error code from the known error codes.
Remove_error_string(errorcode) Remove error string association from error class or error code.
Unpublish_name(service_name, port_name[, info]) Unpublish a service name.
Wtick() Return the resolution of Wtime.
Wtime() Return an elapsed time on the calling processor.
get_vendor() Information about the underlying MPI implementation.

mpi4py.MPI.Add_error_class

Add an error class to the known error classes.


mpi4py.MPI.Add_error_code

Add an error code to an error class.
errorclass (int)
int


mpi4py.MPI.Add_error_string

Associate an error string with an error class or error code.
  • errorcode (int)
  • string (str)

None


mpi4py.MPI.Aint_add

Return the sum of base address and displacement.
  • base (int)
  • disp (int)

int


mpi4py.MPI.Aint_diff

Return the difference between absolute addresses.
  • addr1 (int)
  • addr2 (int)

int


mpi4py.MPI.Alloc_mem

Allocate memory for message passing and remote memory access.
  • size (int)
  • info (Info)

buffer


mpi4py.MPI.Attach_buffer

Attach a user-provided buffer for sending in buffered mode.
buf (Buffer | None)
None


mpi4py.MPI.Close_port

Close a port.
port_name (str)
None


mpi4py.MPI.Compute_dims

Return a balanced distribution of processes per coordinate direction.
  • nnodes (int)
  • dims (int | Sequence[int])

list[int]


mpi4py.MPI.Detach_buffer

Remove an existing attached buffer.
Buffer | None


mpi4py.MPI.Finalize

Terminate the MPI execution environment.


mpi4py.MPI.Flush_buffer

Block until all buffered messages have been transmitted.


mpi4py.MPI.Free_mem

Free memory allocated with Alloc_mem.
mem (buffer)
None


mpi4py.MPI.Get_address

Get the address of a location in memory.
location (Buffer | Bottom)
int


mpi4py.MPI.Get_error_class

Convert an error code into an error class.
errorcode (int)
int


mpi4py.MPI.Get_error_string

Return the error string for a given error class or error code.
errorcode (int)
str


mpi4py.MPI.Get_hw_resource_info

Obtain information about the hardware platform of the calling processor.


mpi4py.MPI.Get_library_version

Obtain the version string of the MPI library.


mpi4py.MPI.Get_processor_name

Obtain the name of the calling processor.


mpi4py.MPI.Get_version

Obtain the version number of the MPI standard.
tuple[int, int]


mpi4py.MPI.Iflush_buffer

Nonblocking flush for buffered messages.
Request


mpi4py.MPI.Init

Initialize the MPI execution environment.


mpi4py.MPI.Init_thread

Initialize the MPI execution environment.
required (int)
int


mpi4py.MPI.Is_finalized

Indicate whether Finalize has completed.


mpi4py.MPI.Is_initialized

Indicate whether Init has been called.


mpi4py.MPI.Is_thread_main

Indicate whether this thread called Init or Init_thread.


mpi4py.MPI.Lookup_name

Lookup a port name given a service name.
  • service_name (str)
  • info (Info)

str


mpi4py.MPI.Open_port

Return an address used to connect group of processes.
info (Info)
str


mpi4py.MPI.Pcontrol

Control profiling.
level (int)
None


mpi4py.MPI.Publish_name

Publish a service name.
  • service_name (str)
  • port_name (str)
  • info (Info)

None


mpi4py.MPI.Query_thread

Return the level of thread support provided by the MPI library.


mpi4py.MPI.Register_datarep

Register user-defined data representations.
  • datarep (str)
  • read_fn (Callable[[Buffer, Datatype, int, Buffer, int], None])
  • write_fn (Callable[[Buffer, Datatype, int, Buffer, int], None])
  • extent_fn (Callable[[Datatype], int])

None


mpi4py.MPI.Remove_error_class

Remove an error class from the known error classes.
errorclass (int)
None


mpi4py.MPI.Remove_error_code

Remove an error code from the known error codes.
errorcode (int)
None


mpi4py.MPI.Remove_error_string

Remove error string association from error class or error code.
errorcode (int)
None


mpi4py.MPI.Unpublish_name

Unpublish a service name.
  • service_name (str)
  • port_name (str)
  • info (Info)

None


mpi4py.MPI.Wtick

Return the resolution of Wtime.
float


mpi4py.MPI.Wtime

Return an elapsed time on the calling processor.
float


mpi4py.MPI.get_vendor

Information about the underlying MPI implementation.
  • string with the name of the MPI implementation.
  • integer 3-tuple version number (major, minor, micro).

tuple[str, tuple[int, int, int]]


Attributes

UNDEFINED Constant UNDEFINED of type int
ANY_SOURCE Constant ANY_SOURCE of type int
ANY_TAG Constant ANY_TAG of type int
PROC_NULL Constant PROC_NULL of type int
ROOT Constant ROOT of type int
BOTTOM Constant BOTTOM of type BottomType
IN_PLACE Constant IN_PLACE of type InPlaceType
KEYVAL_INVALID Constant KEYVAL_INVALID of type int
TAG_UB Constant TAG_UB of type int
IO Constant IO of type int
WTIME_IS_GLOBAL Constant WTIME_IS_GLOBAL of type int
UNIVERSE_SIZE Constant UNIVERSE_SIZE of type int
APPNUM Constant APPNUM of type int
LASTUSEDCODE Constant LASTUSEDCODE of type int
WIN_BASE Constant WIN_BASE of type int
WIN_SIZE Constant WIN_SIZE of type int
WIN_DISP_UNIT Constant WIN_DISP_UNIT of type int
WIN_CREATE_FLAVOR Constant WIN_CREATE_FLAVOR of type int
WIN_FLAVOR Constant WIN_FLAVOR of type int
WIN_MODEL Constant WIN_MODEL of type int
SUCCESS Constant SUCCESS of type int
ERR_LASTCODE Constant ERR_LASTCODE of type int
ERR_TYPE Constant ERR_TYPE of type int
ERR_REQUEST Constant ERR_REQUEST of type int
ERR_OP Constant ERR_OP of type int
ERR_GROUP Constant ERR_GROUP of type int
ERR_INFO Constant ERR_INFO of type int
ERR_ERRHANDLER Constant ERR_ERRHANDLER of type int
ERR_SESSION Constant ERR_SESSION of type int
ERR_COMM Constant ERR_COMM of type int
ERR_WIN Constant ERR_WIN of type int
ERR_FILE Constant ERR_FILE of type int
ERR_BUFFER Constant ERR_BUFFER of type int
ERR_COUNT Constant ERR_COUNT of type int
ERR_TAG Constant ERR_TAG of type int
ERR_RANK Constant ERR_RANK of type int
ERR_ROOT Constant ERR_ROOT of type int
ERR_TRUNCATE Constant ERR_TRUNCATE of type int
ERR_IN_STATUS Constant ERR_IN_STATUS of type int
ERR_PENDING Constant ERR_PENDING of type int
ERR_TOPOLOGY Constant ERR_TOPOLOGY of type int
ERR_DIMS Constant ERR_DIMS of type int
ERR_ARG Constant ERR_ARG of type int
ERR_OTHER Constant ERR_OTHER of type int
ERR_UNKNOWN Constant ERR_UNKNOWN of type int
ERR_INTERN Constant ERR_INTERN of type int
ERR_KEYVAL Constant ERR_KEYVAL of type int
ERR_NO_MEM Constant ERR_NO_MEM of type int
ERR_INFO_KEY Constant ERR_INFO_KEY of type int
ERR_INFO_VALUE Constant ERR_INFO_VALUE of type int
ERR_INFO_NOKEY Constant ERR_INFO_NOKEY of type int
ERR_SPAWN Constant ERR_SPAWN of type int
ERR_PORT Constant ERR_PORT of type int
ERR_SERVICE Constant ERR_SERVICE of type int
ERR_NAME Constant ERR_NAME of type int
ERR_PROC_ABORTED Constant ERR_PROC_ABORTED of type int
ERR_BASE Constant ERR_BASE of type int
ERR_SIZE Constant ERR_SIZE of type int
ERR_DISP Constant ERR_DISP of type int
ERR_ASSERT Constant ERR_ASSERT of type int
ERR_LOCKTYPE Constant ERR_LOCKTYPE of type int
ERR_RMA_CONFLICT Constant ERR_RMA_CONFLICT of type int
ERR_RMA_SYNC Constant ERR_RMA_SYNC of type int
ERR_RMA_RANGE Constant ERR_RMA_RANGE of type int
ERR_RMA_ATTACH Constant ERR_RMA_ATTACH of type int
ERR_RMA_SHARED Constant ERR_RMA_SHARED of type int
ERR_RMA_FLAVOR Constant ERR_RMA_FLAVOR of type int
ERR_BAD_FILE Constant ERR_BAD_FILE of type int
ERR_NO_SUCH_FILE Constant ERR_NO_SUCH_FILE of type int
ERR_FILE_EXISTS Constant ERR_FILE_EXISTS of type int
ERR_FILE_IN_USE Constant ERR_FILE_IN_USE of type int
ERR_AMODE Constant ERR_AMODE of type int
ERR_ACCESS Constant ERR_ACCESS of type int
ERR_READ_ONLY Constant ERR_READ_ONLY of type int
ERR_NO_SPACE Constant ERR_NO_SPACE of type int
ERR_QUOTA Constant ERR_QUOTA of type int
ERR_NOT_SAME Constant ERR_NOT_SAME of type int
ERR_IO Constant ERR_IO of type int
ERR_UNSUPPORTED_OPERATION Constant ERR_UNSUPPORTED_OPERATION of type int
ERR_UNSUPPORTED_DATAREP Constant ERR_UNSUPPORTED_DATAREP of type int
ERR_CONVERSION Constant ERR_CONVERSION of type int
ERR_DUP_DATAREP Constant ERR_DUP_DATAREP of type int
ERR_VALUE_TOO_LARGE Constant ERR_VALUE_TOO_LARGE of type int
ERR_REVOKED Constant ERR_REVOKED of type int
ERR_PROC_FAILED Constant ERR_PROC_FAILED of type int
ERR_PROC_FAILED_PENDING Constant ERR_PROC_FAILED_PENDING of type int
ORDER_C Constant ORDER_C of type int
ORDER_FORTRAN Constant ORDER_FORTRAN of type int
ORDER_F Constant ORDER_F of type int
TYPECLASS_INTEGER Constant TYPECLASS_INTEGER of type int
TYPECLASS_REAL Constant TYPECLASS_REAL of type int
TYPECLASS_COMPLEX Constant TYPECLASS_COMPLEX of type int
DISTRIBUTE_NONE Constant DISTRIBUTE_NONE of type int
DISTRIBUTE_BLOCK Constant DISTRIBUTE_BLOCK of type int
DISTRIBUTE_CYCLIC Constant DISTRIBUTE_CYCLIC of type int
DISTRIBUTE_DFLT_DARG Constant DISTRIBUTE_DFLT_DARG of type int
COMBINER_NAMED Constant COMBINER_NAMED of type int
COMBINER_DUP Constant COMBINER_DUP of type int
COMBINER_CONTIGUOUS Constant COMBINER_CONTIGUOUS of type int
COMBINER_VECTOR Constant COMBINER_VECTOR of type int
COMBINER_HVECTOR Constant COMBINER_HVECTOR of type int
COMBINER_INDEXED Constant COMBINER_INDEXED of type int
COMBINER_HINDEXED Constant COMBINER_HINDEXED of type int
COMBINER_INDEXED_BLOCK Constant COMBINER_INDEXED_BLOCK of type int
COMBINER_HINDEXED_BLOCK Constant COMBINER_HINDEXED_BLOCK of type int
COMBINER_STRUCT Constant COMBINER_STRUCT of type int
COMBINER_SUBARRAY Constant COMBINER_SUBARRAY of type int
COMBINER_DARRAY Constant COMBINER_DARRAY of type int
COMBINER_RESIZED Constant COMBINER_RESIZED of type int
COMBINER_VALUE_INDEX Constant COMBINER_VALUE_INDEX of type int
COMBINER_F90_INTEGER Constant COMBINER_F90_INTEGER of type int
COMBINER_F90_REAL Constant COMBINER_F90_REAL of type int
COMBINER_F90_COMPLEX Constant COMBINER_F90_COMPLEX of type int
F_SOURCE Constant F_SOURCE of type int
F_TAG Constant F_TAG of type int
F_ERROR Constant F_ERROR of type int
F_STATUS_SIZE Constant F_STATUS_SIZE of type int
IDENT Constant IDENT of type int
CONGRUENT Constant CONGRUENT of type int
SIMILAR Constant SIMILAR of type int
UNEQUAL Constant UNEQUAL of type int
CART Constant CART of type int
GRAPH Constant GRAPH of type int
DIST_GRAPH Constant DIST_GRAPH of type int
UNWEIGHTED Constant UNWEIGHTED of type int
WEIGHTS_EMPTY Constant WEIGHTS_EMPTY of type int
COMM_TYPE_SHARED Constant COMM_TYPE_SHARED of type int
COMM_TYPE_HW_GUIDED Constant COMM_TYPE_HW_GUIDED of type int
COMM_TYPE_HW_UNGUIDED Constant COMM_TYPE_HW_UNGUIDED of type int
COMM_TYPE_RESOURCE_GUIDED Constant COMM_TYPE_RESOURCE_GUIDED of type int
BSEND_OVERHEAD Constant BSEND_OVERHEAD of type int
BUFFER_AUTOMATIC Constant BUFFER_AUTOMATIC of type BufferAutomaticType
WIN_FLAVOR_CREATE Constant WIN_FLAVOR_CREATE of type int
WIN_FLAVOR_ALLOCATE Constant WIN_FLAVOR_ALLOCATE of type int
WIN_FLAVOR_DYNAMIC Constant WIN_FLAVOR_DYNAMIC of type int
WIN_FLAVOR_SHARED Constant WIN_FLAVOR_SHARED of type int
WIN_SEPARATE Constant WIN_SEPARATE of type int
WIN_UNIFIED Constant WIN_UNIFIED of type int
MODE_NOCHECK Constant MODE_NOCHECK of type int
MODE_NOSTORE Constant MODE_NOSTORE of type int
MODE_NOPUT Constant MODE_NOPUT of type int
MODE_NOPRECEDE Constant MODE_NOPRECEDE of type int
MODE_NOSUCCEED Constant MODE_NOSUCCEED of type int
LOCK_EXCLUSIVE Constant LOCK_EXCLUSIVE of type int
LOCK_SHARED Constant LOCK_SHARED of type int
MODE_RDONLY Constant MODE_RDONLY of type int
MODE_WRONLY Constant MODE_WRONLY of type int
MODE_RDWR Constant MODE_RDWR of type int
MODE_CREATE Constant MODE_CREATE of type int
MODE_EXCL Constant MODE_EXCL of type int
MODE_DELETE_ON_CLOSE Constant MODE_DELETE_ON_CLOSE of type int
MODE_UNIQUE_OPEN Constant MODE_UNIQUE_OPEN of type int
MODE_SEQUENTIAL Constant MODE_SEQUENTIAL of type int
MODE_APPEND Constant MODE_APPEND of type int
SEEK_SET Constant SEEK_SET of type int
SEEK_CUR Constant SEEK_CUR of type int
SEEK_END Constant SEEK_END of type int
DISPLACEMENT_CURRENT Constant DISPLACEMENT_CURRENT of type int
DISP_CUR Constant DISP_CUR of type int
THREAD_SINGLE Constant THREAD_SINGLE of type int
THREAD_FUNNELED Constant THREAD_FUNNELED of type int
THREAD_SERIALIZED Constant THREAD_SERIALIZED of type int
THREAD_MULTIPLE Constant THREAD_MULTIPLE of type int
VERSION Constant VERSION of type int
SUBVERSION Constant SUBVERSION of type int
MAX_PROCESSOR_NAME Constant MAX_PROCESSOR_NAME of type int
MAX_ERROR_STRING Constant MAX_ERROR_STRING of type int
MAX_PORT_NAME Constant MAX_PORT_NAME of type int
MAX_INFO_KEY Constant MAX_INFO_KEY of type int
MAX_INFO_VAL Constant MAX_INFO_VAL of type int
MAX_OBJECT_NAME Constant MAX_OBJECT_NAME of type int
MAX_DATAREP_STRING Constant MAX_DATAREP_STRING of type int
MAX_LIBRARY_VERSION_STRING Constant MAX_LIBRARY_VERSION_STRING of type int
MAX_PSET_NAME_LEN Constant MAX_PSET_NAME_LEN of type int
MAX_STRINGTAG_LEN Constant MAX_STRINGTAG_LEN of type int
DATATYPE_NULL Object DATATYPE_NULL of type Datatype
PACKED Object PACKED of type Datatype
BYTE Object BYTE of type Datatype
AINT Object AINT of type Datatype
OFFSET Object OFFSET of type Datatype
COUNT Object COUNT of type Datatype
CHAR Object CHAR of type Datatype
WCHAR Object WCHAR of type Datatype
SIGNED_CHAR Object SIGNED_CHAR of type Datatype
SHORT Object SHORT of type Datatype
INT Object INT of type Datatype
LONG Object LONG of type Datatype
LONG_LONG Object LONG_LONG of type Datatype
UNSIGNED_CHAR Object UNSIGNED_CHAR of type Datatype
UNSIGNED_SHORT Object UNSIGNED_SHORT of type Datatype
UNSIGNED Object UNSIGNED of type Datatype
UNSIGNED_LONG Object UNSIGNED_LONG of type Datatype
UNSIGNED_LONG_LONG Object UNSIGNED_LONG_LONG of type Datatype
FLOAT Object FLOAT of type Datatype
DOUBLE Object DOUBLE of type Datatype
LONG_DOUBLE Object LONG_DOUBLE of type Datatype
C_BOOL Object C_BOOL of type Datatype
INT8_T Object INT8_T of type Datatype
INT16_T Object INT16_T of type Datatype
INT32_T Object INT32_T of type Datatype
INT64_T Object INT64_T of type Datatype
UINT8_T Object UINT8_T of type Datatype
UINT16_T Object UINT16_T of type Datatype
UINT32_T Object UINT32_T of type Datatype
UINT64_T Object UINT64_T of type Datatype
C_COMPLEX Object C_COMPLEX of type Datatype
C_FLOAT_COMPLEX Object C_FLOAT_COMPLEX of type Datatype
C_DOUBLE_COMPLEX Object C_DOUBLE_COMPLEX of type Datatype
C_LONG_DOUBLE_COMPLEX Object C_LONG_DOUBLE_COMPLEX of type Datatype
CXX_BOOL Object CXX_BOOL of type Datatype
CXX_FLOAT_COMPLEX Object CXX_FLOAT_COMPLEX of type Datatype
CXX_DOUBLE_COMPLEX Object CXX_DOUBLE_COMPLEX of type Datatype
CXX_LONG_DOUBLE_COMPLEX Object CXX_LONG_DOUBLE_COMPLEX of type Datatype
SHORT_INT Object SHORT_INT of type Datatype
INT_INT Object INT_INT of type Datatype
TWOINT Object TWOINT of type Datatype
LONG_INT Object LONG_INT of type Datatype
FLOAT_INT Object FLOAT_INT of type Datatype
DOUBLE_INT Object DOUBLE_INT of type Datatype
LONG_DOUBLE_INT Object LONG_DOUBLE_INT of type Datatype
CHARACTER Object CHARACTER of type Datatype
LOGICAL Object LOGICAL of type Datatype
INTEGER Object INTEGER of type Datatype
REAL Object REAL of type Datatype
DOUBLE_PRECISION Object DOUBLE_PRECISION of type Datatype
COMPLEX Object COMPLEX of type Datatype
DOUBLE_COMPLEX Object DOUBLE_COMPLEX of type Datatype
LOGICAL1 Object LOGICAL1 of type Datatype
LOGICAL2 Object LOGICAL2 of type Datatype
LOGICAL4 Object LOGICAL4 of type Datatype
LOGICAL8 Object LOGICAL8 of type Datatype
INTEGER1 Object INTEGER1 of type Datatype
INTEGER2 Object INTEGER2 of type Datatype
INTEGER4 Object INTEGER4 of type Datatype
INTEGER8 Object INTEGER8 of type Datatype
INTEGER16 Object INTEGER16 of type Datatype
REAL2 Object REAL2 of type Datatype
REAL4 Object REAL4 of type Datatype
REAL8 Object REAL8 of type Datatype
REAL16 Object REAL16 of type Datatype
COMPLEX4 Object COMPLEX4 of type Datatype
COMPLEX8 Object COMPLEX8 of type Datatype
COMPLEX16 Object COMPLEX16 of type Datatype
COMPLEX32 Object COMPLEX32 of type Datatype
UNSIGNED_INT Object UNSIGNED_INT of type Datatype
SIGNED_SHORT Object SIGNED_SHORT of type Datatype
SIGNED_INT Object SIGNED_INT of type Datatype
SIGNED_LONG Object SIGNED_LONG of type Datatype
SIGNED_LONG_LONG Object SIGNED_LONG_LONG of type Datatype
BOOL Object BOOL of type Datatype
SINT8_T Object SINT8_T of type Datatype
SINT16_T Object SINT16_T of type Datatype
SINT32_T Object SINT32_T of type Datatype
SINT64_T Object SINT64_T of type Datatype
F_BOOL Object F_BOOL of type Datatype
F_INT Object F_INT of type Datatype
F_FLOAT Object F_FLOAT of type Datatype
F_DOUBLE Object F_DOUBLE of type Datatype
F_COMPLEX Object F_COMPLEX of type Datatype
F_FLOAT_COMPLEX Object F_FLOAT_COMPLEX of type Datatype
F_DOUBLE_COMPLEX Object F_DOUBLE_COMPLEX of type Datatype
REQUEST_NULL Object REQUEST_NULL of type Request
MESSAGE_NULL Object MESSAGE_NULL of type Message
MESSAGE_NO_PROC Object MESSAGE_NO_PROC of type Message
OP_NULL Object OP_NULL of type Op
MAX Object MAX of type Op
MIN Object MIN of type Op
SUM Object SUM of type Op
PROD Object PROD of type Op
LAND Object LAND of type Op
BAND Object BAND of type Op
LOR Object LOR of type Op
BOR Object BOR of type Op
LXOR Object LXOR of type Op
BXOR Object BXOR of type Op
MAXLOC Object MAXLOC of type Op
MINLOC Object MINLOC of type Op
REPLACE Object REPLACE of type Op
NO_OP Object NO_OP of type Op
GROUP_NULL Object GROUP_NULL of type Group
GROUP_EMPTY Object GROUP_EMPTY of type Group
INFO_NULL Object INFO_NULL of type Info
INFO_ENV Object INFO_ENV of type Info
ERRHANDLER_NULL Object ERRHANDLER_NULL of type Errhandler
ERRORS_RETURN Object ERRORS_RETURN of type Errhandler
ERRORS_ABORT Object ERRORS_ABORT of type Errhandler
ERRORS_ARE_FATAL Object ERRORS_ARE_FATAL of type Errhandler
SESSION_NULL Object SESSION_NULL of type Session
COMM_NULL Object COMM_NULL of type Comm
COMM_SELF Object COMM_SELF of type Intracomm
COMM_WORLD Object COMM_WORLD of type Intracomm
WIN_NULL Object WIN_NULL of type Win
FILE_NULL Object FILE_NULL of type File
pickle Object pickle of type Pickle

mpi4py.MPI.UNDEFINED

Constant UNDEFINED of type int

mpi4py.MPI.ANY_SOURCE

Constant ANY_SOURCE of type int

mpi4py.MPI.ANY_TAG

Constant ANY_TAG of type int

mpi4py.MPI.PROC_NULL

Constant PROC_NULL of type int

mpi4py.MPI.ROOT

Constant ROOT of type int

mpi4py.MPI.BOTTOM

Constant BOTTOM of type BottomType

mpi4py.MPI.IN_PLACE

Constant IN_PLACE of type InPlaceType

mpi4py.MPI.KEYVAL_INVALID

Constant KEYVAL_INVALID of type int

mpi4py.MPI.TAG_UB

Constant TAG_UB of type int

mpi4py.MPI.IO

Constant IO of type int

mpi4py.MPI.WTIME_IS_GLOBAL

Constant WTIME_IS_GLOBAL of type int

mpi4py.MPI.UNIVERSE_SIZE

Constant UNIVERSE_SIZE of type int

mpi4py.MPI.APPNUM

Constant APPNUM of type int

mpi4py.MPI.LASTUSEDCODE

Constant LASTUSEDCODE of type int

mpi4py.MPI.WIN_BASE

Constant WIN_BASE of type int

mpi4py.MPI.WIN_SIZE

Constant WIN_SIZE of type int

mpi4py.MPI.WIN_DISP_UNIT

Constant WIN_DISP_UNIT of type int

mpi4py.MPI.WIN_CREATE_FLAVOR

Constant WIN_CREATE_FLAVOR of type int

mpi4py.MPI.WIN_FLAVOR

Constant WIN_FLAVOR of type int

mpi4py.MPI.WIN_MODEL

Constant WIN_MODEL of type int

mpi4py.MPI.SUCCESS

Constant SUCCESS of type int

mpi4py.MPI.ERR_LASTCODE

Constant ERR_LASTCODE of type int

mpi4py.MPI.ERR_TYPE

Constant ERR_TYPE of type int

mpi4py.MPI.ERR_REQUEST

Constant ERR_REQUEST of type int

mpi4py.MPI.ERR_OP

Constant ERR_OP of type int

mpi4py.MPI.ERR_GROUP

Constant ERR_GROUP of type int

mpi4py.MPI.ERR_INFO

Constant ERR_INFO of type int

mpi4py.MPI.ERR_ERRHANDLER

Constant ERR_ERRHANDLER of type int

mpi4py.MPI.ERR_SESSION

Constant ERR_SESSION of type int

mpi4py.MPI.ERR_COMM

Constant ERR_COMM of type int

mpi4py.MPI.ERR_WIN

Constant ERR_WIN of type int

mpi4py.MPI.ERR_FILE

Constant ERR_FILE of type int

mpi4py.MPI.ERR_BUFFER

Constant ERR_BUFFER of type int

mpi4py.MPI.ERR_COUNT

Constant ERR_COUNT of type int

mpi4py.MPI.ERR_TAG

Constant ERR_TAG of type int

mpi4py.MPI.ERR_RANK

Constant ERR_RANK of type int

mpi4py.MPI.ERR_ROOT

Constant ERR_ROOT of type int

mpi4py.MPI.ERR_TRUNCATE

Constant ERR_TRUNCATE of type int

mpi4py.MPI.ERR_IN_STATUS

Constant ERR_IN_STATUS of type int

mpi4py.MPI.ERR_PENDING

Constant ERR_PENDING of type int

mpi4py.MPI.ERR_TOPOLOGY

Constant ERR_TOPOLOGY of type int

mpi4py.MPI.ERR_DIMS

Constant ERR_DIMS of type int

mpi4py.MPI.ERR_ARG

Constant ERR_ARG of type int

mpi4py.MPI.ERR_OTHER

Constant ERR_OTHER of type int

mpi4py.MPI.ERR_UNKNOWN

Constant ERR_UNKNOWN of type int

mpi4py.MPI.ERR_INTERN

Constant ERR_INTERN of type int

mpi4py.MPI.ERR_KEYVAL

Constant ERR_KEYVAL of type int

mpi4py.MPI.ERR_NO_MEM

Constant ERR_NO_MEM of type int

mpi4py.MPI.ERR_INFO_KEY

Constant ERR_INFO_KEY of type int

mpi4py.MPI.ERR_INFO_VALUE

Constant ERR_INFO_VALUE of type int

mpi4py.MPI.ERR_INFO_NOKEY

Constant ERR_INFO_NOKEY of type int

mpi4py.MPI.ERR_SPAWN

Constant ERR_SPAWN of type int

mpi4py.MPI.ERR_PORT

Constant ERR_PORT of type int

mpi4py.MPI.ERR_SERVICE

Constant ERR_SERVICE of type int

mpi4py.MPI.ERR_NAME

Constant ERR_NAME of type int

mpi4py.MPI.ERR_PROC_ABORTED

Constant ERR_PROC_ABORTED of type int

mpi4py.MPI.ERR_BASE

Constant ERR_BASE of type int

mpi4py.MPI.ERR_SIZE

Constant ERR_SIZE of type int

mpi4py.MPI.ERR_DISP

Constant ERR_DISP of type int

mpi4py.MPI.ERR_ASSERT

Constant ERR_ASSERT of type int

mpi4py.MPI.ERR_LOCKTYPE

Constant ERR_LOCKTYPE of type int

mpi4py.MPI.ERR_RMA_CONFLICT

Constant ERR_RMA_CONFLICT of type int

mpi4py.MPI.ERR_RMA_SYNC

Constant ERR_RMA_SYNC of type int

mpi4py.MPI.ERR_RMA_RANGE

Constant ERR_RMA_RANGE of type int

mpi4py.MPI.ERR_RMA_ATTACH

Constant ERR_RMA_ATTACH of type int

mpi4py.MPI.ERR_RMA_SHARED

Constant ERR_RMA_SHARED of type int

mpi4py.MPI.ERR_RMA_FLAVOR

Constant ERR_RMA_FLAVOR of type int

mpi4py.MPI.ERR_BAD_FILE

Constant ERR_BAD_FILE of type int

mpi4py.MPI.ERR_NO_SUCH_FILE

Constant ERR_NO_SUCH_FILE of type int

mpi4py.MPI.ERR_FILE_EXISTS

Constant ERR_FILE_EXISTS of type int

mpi4py.MPI.ERR_FILE_IN_USE

Constant ERR_FILE_IN_USE of type int

mpi4py.MPI.ERR_AMODE

Constant ERR_AMODE of type int

mpi4py.MPI.ERR_ACCESS

Constant ERR_ACCESS of type int

mpi4py.MPI.ERR_READ_ONLY

Constant ERR_READ_ONLY of type int

mpi4py.MPI.ERR_NO_SPACE

Constant ERR_NO_SPACE of type int

mpi4py.MPI.ERR_QUOTA

Constant ERR_QUOTA of type int

mpi4py.MPI.ERR_NOT_SAME

Constant ERR_NOT_SAME of type int

mpi4py.MPI.ERR_IO

Constant ERR_IO of type int

mpi4py.MPI.ERR_UNSUPPORTED_OPERATION

Constant ERR_UNSUPPORTED_OPERATION of type int

mpi4py.MPI.ERR_UNSUPPORTED_DATAREP

Constant ERR_UNSUPPORTED_DATAREP of type int

mpi4py.MPI.ERR_CONVERSION

Constant ERR_CONVERSION of type int

mpi4py.MPI.ERR_DUP_DATAREP

Constant ERR_DUP_DATAREP of type int

mpi4py.MPI.ERR_VALUE_TOO_LARGE

Constant ERR_VALUE_TOO_LARGE of type int

mpi4py.MPI.ERR_REVOKED

Constant ERR_REVOKED of type int

mpi4py.MPI.ERR_PROC_FAILED

Constant ERR_PROC_FAILED of type int

mpi4py.MPI.ERR_PROC_FAILED_PENDING

Constant ERR_PROC_FAILED_PENDING of type int

mpi4py.MPI.ORDER_C

Constant ORDER_C of type int

mpi4py.MPI.ORDER_FORTRAN

Constant ORDER_FORTRAN of type int

mpi4py.MPI.ORDER_F

Constant ORDER_F of type int

mpi4py.MPI.TYPECLASS_INTEGER

Constant TYPECLASS_INTEGER of type int

mpi4py.MPI.TYPECLASS_REAL

Constant TYPECLASS_REAL of type int

mpi4py.MPI.TYPECLASS_COMPLEX

Constant TYPECLASS_COMPLEX of type int

mpi4py.MPI.DISTRIBUTE_NONE

Constant DISTRIBUTE_NONE of type int

mpi4py.MPI.DISTRIBUTE_BLOCK

Constant DISTRIBUTE_BLOCK of type int

mpi4py.MPI.DISTRIBUTE_CYCLIC

Constant DISTRIBUTE_CYCLIC of type int

mpi4py.MPI.DISTRIBUTE_DFLT_DARG

Constant DISTRIBUTE_DFLT_DARG of type int

mpi4py.MPI.COMBINER_NAMED

Constant COMBINER_NAMED of type int

mpi4py.MPI.COMBINER_DUP

Constant COMBINER_DUP of type int

mpi4py.MPI.COMBINER_CONTIGUOUS

Constant COMBINER_CONTIGUOUS of type int

mpi4py.MPI.COMBINER_VECTOR

Constant COMBINER_VECTOR of type int

mpi4py.MPI.COMBINER_HVECTOR

Constant COMBINER_HVECTOR of type int

mpi4py.MPI.COMBINER_INDEXED

Constant COMBINER_INDEXED of type int

mpi4py.MPI.COMBINER_HINDEXED

Constant COMBINER_HINDEXED of type int

mpi4py.MPI.COMBINER_INDEXED_BLOCK

Constant COMBINER_INDEXED_BLOCK of type int

mpi4py.MPI.COMBINER_HINDEXED_BLOCK

Constant COMBINER_HINDEXED_BLOCK of type int

mpi4py.MPI.COMBINER_STRUCT

Constant COMBINER_STRUCT of type int

mpi4py.MPI.COMBINER_SUBARRAY

Constant COMBINER_SUBARRAY of type int

mpi4py.MPI.COMBINER_DARRAY

Constant COMBINER_DARRAY of type int

mpi4py.MPI.COMBINER_RESIZED

Constant COMBINER_RESIZED of type int

mpi4py.MPI.COMBINER_VALUE_INDEX

Constant COMBINER_VALUE_INDEX of type int

mpi4py.MPI.COMBINER_F90_INTEGER

Constant COMBINER_F90_INTEGER of type int

mpi4py.MPI.COMBINER_F90_REAL

Constant COMBINER_F90_REAL of type int

mpi4py.MPI.COMBINER_F90_COMPLEX

Constant COMBINER_F90_COMPLEX of type int

mpi4py.MPI.F_SOURCE

Constant F_SOURCE of type int

mpi4py.MPI.F_TAG

Constant F_TAG of type int

mpi4py.MPI.F_ERROR

Constant F_ERROR of type int

mpi4py.MPI.F_STATUS_SIZE

Constant F_STATUS_SIZE of type int

mpi4py.MPI.IDENT

Constant IDENT of type int

mpi4py.MPI.CONGRUENT

Constant CONGRUENT of type int

mpi4py.MPI.SIMILAR

Constant SIMILAR of type int

mpi4py.MPI.UNEQUAL

Constant UNEQUAL of type int

mpi4py.MPI.CART

Constant CART of type int

mpi4py.MPI.GRAPH

Constant GRAPH of type int

mpi4py.MPI.DIST_GRAPH

Constant DIST_GRAPH of type int

mpi4py.MPI.UNWEIGHTED

Constant UNWEIGHTED of type int

mpi4py.MPI.WEIGHTS_EMPTY

Constant WEIGHTS_EMPTY of type int

mpi4py.MPI.COMM_TYPE_SHARED

Constant COMM_TYPE_SHARED of type int

mpi4py.MPI.COMM_TYPE_HW_GUIDED

Constant COMM_TYPE_HW_GUIDED of type int

mpi4py.MPI.COMM_TYPE_HW_UNGUIDED

Constant COMM_TYPE_HW_UNGUIDED of type int

mpi4py.MPI.COMM_TYPE_RESOURCE_GUIDED

Constant COMM_TYPE_RESOURCE_GUIDED of type int

mpi4py.MPI.BSEND_OVERHEAD

Constant BSEND_OVERHEAD of type int

mpi4py.MPI.BUFFER_AUTOMATIC

Constant BUFFER_AUTOMATIC of type BufferAutomaticType

mpi4py.MPI.WIN_FLAVOR_CREATE

Constant WIN_FLAVOR_CREATE of type int

mpi4py.MPI.WIN_FLAVOR_ALLOCATE

Constant WIN_FLAVOR_ALLOCATE of type int

mpi4py.MPI.WIN_FLAVOR_DYNAMIC

Constant WIN_FLAVOR_DYNAMIC of type int

mpi4py.MPI.WIN_FLAVOR_SHARED

Constant WIN_FLAVOR_SHARED of type int

mpi4py.MPI.WIN_SEPARATE

Constant WIN_SEPARATE of type int

mpi4py.MPI.WIN_UNIFIED

Constant WIN_UNIFIED of type int

mpi4py.MPI.MODE_NOCHECK

Constant MODE_NOCHECK of type int

mpi4py.MPI.MODE_NOSTORE

Constant MODE_NOSTORE of type int

mpi4py.MPI.MODE_NOPUT

Constant MODE_NOPUT of type int

mpi4py.MPI.MODE_NOPRECEDE

Constant MODE_NOPRECEDE of type int

mpi4py.MPI.MODE_NOSUCCEED

Constant MODE_NOSUCCEED of type int

mpi4py.MPI.LOCK_EXCLUSIVE

Constant LOCK_EXCLUSIVE of type int

mpi4py.MPI.LOCK_SHARED

Constant LOCK_SHARED of type int

mpi4py.MPI.MODE_RDONLY

Constant MODE_RDONLY of type int

mpi4py.MPI.MODE_WRONLY

Constant MODE_WRONLY of type int

mpi4py.MPI.MODE_RDWR

Constant MODE_RDWR of type int

mpi4py.MPI.MODE_CREATE

Constant MODE_CREATE of type int

mpi4py.MPI.MODE_EXCL

Constant MODE_EXCL of type int

mpi4py.MPI.MODE_DELETE_ON_CLOSE

Constant MODE_DELETE_ON_CLOSE of type int

mpi4py.MPI.MODE_UNIQUE_OPEN

Constant MODE_UNIQUE_OPEN of type int

mpi4py.MPI.MODE_SEQUENTIAL

Constant MODE_SEQUENTIAL of type int

mpi4py.MPI.MODE_APPEND

Constant MODE_APPEND of type int

mpi4py.MPI.SEEK_SET

Constant SEEK_SET of type int

mpi4py.MPI.SEEK_CUR

Constant SEEK_CUR of type int

mpi4py.MPI.SEEK_END

Constant SEEK_END of type int

mpi4py.MPI.DISPLACEMENT_CURRENT

Constant DISPLACEMENT_CURRENT of type int

mpi4py.MPI.DISP_CUR

Constant DISP_CUR of type int

mpi4py.MPI.THREAD_SINGLE

Constant THREAD_SINGLE of type int

mpi4py.MPI.THREAD_FUNNELED

Constant THREAD_FUNNELED of type int

mpi4py.MPI.THREAD_SERIALIZED

Constant THREAD_SERIALIZED of type int

mpi4py.MPI.THREAD_MULTIPLE

Constant THREAD_MULTIPLE of type int

mpi4py.MPI.VERSION

Constant VERSION of type int

mpi4py.MPI.SUBVERSION

Constant SUBVERSION of type int

mpi4py.MPI.MAX_PROCESSOR_NAME

Constant MAX_PROCESSOR_NAME of type int

mpi4py.MPI.MAX_ERROR_STRING

Constant MAX_ERROR_STRING of type int

mpi4py.MPI.MAX_PORT_NAME

Constant MAX_PORT_NAME of type int

mpi4py.MPI.MAX_INFO_KEY

Constant MAX_INFO_KEY of type int

mpi4py.MPI.MAX_INFO_VAL

Constant MAX_INFO_VAL of type int

mpi4py.MPI.MAX_OBJECT_NAME

Constant MAX_OBJECT_NAME of type int

mpi4py.MPI.MAX_DATAREP_STRING

Constant MAX_DATAREP_STRING of type int

mpi4py.MPI.MAX_LIBRARY_VERSION_STRING

Constant MAX_LIBRARY_VERSION_STRING of type int

mpi4py.MPI.MAX_PSET_NAME_LEN

Constant MAX_PSET_NAME_LEN of type int

mpi4py.MPI.MAX_STRINGTAG_LEN

Constant MAX_STRINGTAG_LEN of type int

mpi4py.MPI.DATATYPE_NULL

Object DATATYPE_NULL of type Datatype

mpi4py.MPI.PACKED

Object PACKED of type Datatype

mpi4py.MPI.BYTE

Object BYTE of type Datatype

mpi4py.MPI.AINT

Object AINT of type Datatype

mpi4py.MPI.OFFSET

Object OFFSET of type Datatype

mpi4py.MPI.COUNT

Object COUNT of type Datatype

mpi4py.MPI.CHAR

Object CHAR of type Datatype

mpi4py.MPI.WCHAR

Object WCHAR of type Datatype

mpi4py.MPI.SIGNED_CHAR

Object SIGNED_CHAR of type Datatype

mpi4py.MPI.SHORT

Object SHORT of type Datatype

mpi4py.MPI.INT

Object INT of type Datatype

mpi4py.MPI.LONG

Object LONG of type Datatype

mpi4py.MPI.LONG_LONG

Object LONG_LONG of type Datatype

mpi4py.MPI.UNSIGNED_CHAR

Object UNSIGNED_CHAR of type Datatype

mpi4py.MPI.UNSIGNED_SHORT

Object UNSIGNED_SHORT of type Datatype

mpi4py.MPI.UNSIGNED

Object UNSIGNED of type Datatype

mpi4py.MPI.UNSIGNED_LONG

Object UNSIGNED_LONG of type Datatype

mpi4py.MPI.UNSIGNED_LONG_LONG

Object UNSIGNED_LONG_LONG of type Datatype

mpi4py.MPI.FLOAT

Object FLOAT of type Datatype

mpi4py.MPI.DOUBLE

Object DOUBLE of type Datatype

mpi4py.MPI.LONG_DOUBLE

Object LONG_DOUBLE of type Datatype

mpi4py.MPI.C_BOOL

Object C_BOOL of type Datatype

mpi4py.MPI.INT8_T

Object INT8_T of type Datatype

mpi4py.MPI.INT16_T

Object INT16_T of type Datatype

mpi4py.MPI.INT32_T

Object INT32_T of type Datatype

mpi4py.MPI.INT64_T

Object INT64_T of type Datatype

mpi4py.MPI.UINT8_T

Object UINT8_T of type Datatype

mpi4py.MPI.UINT16_T

Object UINT16_T of type Datatype

mpi4py.MPI.UINT32_T

Object UINT32_T of type Datatype

mpi4py.MPI.UINT64_T

Object UINT64_T of type Datatype

mpi4py.MPI.C_COMPLEX

Object C_COMPLEX of type Datatype

mpi4py.MPI.C_FLOAT_COMPLEX

Object C_FLOAT_COMPLEX of type Datatype

mpi4py.MPI.C_DOUBLE_COMPLEX

Object C_DOUBLE_COMPLEX of type Datatype

mpi4py.MPI.C_LONG_DOUBLE_COMPLEX

Object C_LONG_DOUBLE_COMPLEX of type Datatype

mpi4py.MPI.CXX_BOOL

Object CXX_BOOL of type Datatype

mpi4py.MPI.CXX_FLOAT_COMPLEX

Object CXX_FLOAT_COMPLEX of type Datatype

mpi4py.MPI.CXX_DOUBLE_COMPLEX

Object CXX_DOUBLE_COMPLEX of type Datatype

mpi4py.MPI.CXX_LONG_DOUBLE_COMPLEX

Object CXX_LONG_DOUBLE_COMPLEX of type Datatype

mpi4py.MPI.SHORT_INT

Object SHORT_INT of type Datatype

mpi4py.MPI.INT_INT

Object INT_INT of type Datatype

mpi4py.MPI.TWOINT

Object TWOINT of type Datatype

mpi4py.MPI.LONG_INT

Object LONG_INT of type Datatype

mpi4py.MPI.FLOAT_INT

Object FLOAT_INT of type Datatype

mpi4py.MPI.DOUBLE_INT

Object DOUBLE_INT of type Datatype

mpi4py.MPI.LONG_DOUBLE_INT

Object LONG_DOUBLE_INT of type Datatype

mpi4py.MPI.CHARACTER

Object CHARACTER of type Datatype

mpi4py.MPI.LOGICAL

Object LOGICAL of type Datatype

mpi4py.MPI.INTEGER

Object INTEGER of type Datatype

mpi4py.MPI.REAL

Object REAL of type Datatype

mpi4py.MPI.DOUBLE_PRECISION

Object DOUBLE_PRECISION of type Datatype

mpi4py.MPI.COMPLEX

Object COMPLEX of type Datatype

mpi4py.MPI.DOUBLE_COMPLEX

Object DOUBLE_COMPLEX of type Datatype

mpi4py.MPI.LOGICAL1

Object LOGICAL1 of type Datatype

mpi4py.MPI.LOGICAL2

Object LOGICAL2 of type Datatype

mpi4py.MPI.LOGICAL4

Object LOGICAL4 of type Datatype

mpi4py.MPI.LOGICAL8

Object LOGICAL8 of type Datatype

mpi4py.MPI.INTEGER1

Object INTEGER1 of type Datatype

mpi4py.MPI.INTEGER2

Object INTEGER2 of type Datatype

mpi4py.MPI.INTEGER4

Object INTEGER4 of type Datatype

mpi4py.MPI.INTEGER8

Object INTEGER8 of type Datatype

mpi4py.MPI.INTEGER16

Object INTEGER16 of type Datatype

mpi4py.MPI.REAL2

Object REAL2 of type Datatype

mpi4py.MPI.REAL4

Object REAL4 of type Datatype

mpi4py.MPI.REAL8

Object REAL8 of type Datatype

mpi4py.MPI.REAL16

Object REAL16 of type Datatype

mpi4py.MPI.COMPLEX4

Object COMPLEX4 of type Datatype

mpi4py.MPI.COMPLEX8

Object COMPLEX8 of type Datatype

mpi4py.MPI.COMPLEX16

Object COMPLEX16 of type Datatype

mpi4py.MPI.COMPLEX32

Object COMPLEX32 of type Datatype

mpi4py.MPI.UNSIGNED_INT

Object UNSIGNED_INT of type Datatype

mpi4py.MPI.SIGNED_SHORT

Object SIGNED_SHORT of type Datatype

mpi4py.MPI.SIGNED_INT

Object SIGNED_INT of type Datatype

mpi4py.MPI.SIGNED_LONG

Object SIGNED_LONG of type Datatype

mpi4py.MPI.SIGNED_LONG_LONG

Object SIGNED_LONG_LONG of type Datatype

mpi4py.MPI.BOOL

Object BOOL of type Datatype

mpi4py.MPI.SINT8_T

Object SINT8_T of type Datatype

mpi4py.MPI.SINT16_T

Object SINT16_T of type Datatype

mpi4py.MPI.SINT32_T

Object SINT32_T of type Datatype

mpi4py.MPI.SINT64_T

Object SINT64_T of type Datatype

mpi4py.MPI.F_BOOL

Object F_BOOL of type Datatype

mpi4py.MPI.F_INT

Object F_INT of type Datatype

mpi4py.MPI.F_FLOAT

Object F_FLOAT of type Datatype

mpi4py.MPI.F_DOUBLE

Object F_DOUBLE of type Datatype

mpi4py.MPI.F_COMPLEX

Object F_COMPLEX of type Datatype

mpi4py.MPI.F_FLOAT_COMPLEX

Object F_FLOAT_COMPLEX of type Datatype

mpi4py.MPI.F_DOUBLE_COMPLEX

Object F_DOUBLE_COMPLEX of type Datatype

mpi4py.MPI.REQUEST_NULL

Object REQUEST_NULL of type Request

mpi4py.MPI.MESSAGE_NULL

Object MESSAGE_NULL of type Message

mpi4py.MPI.MESSAGE_NO_PROC

Object MESSAGE_NO_PROC of type Message

mpi4py.MPI.OP_NULL

Object OP_NULL of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.MAX

Object MAX of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.MIN

Object MIN of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.SUM

Object SUM of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.PROD

Object PROD of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.LAND

Object LAND of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.BAND

Object BAND of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.LOR

Object LOR of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.BOR

Object BOR of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.LXOR

Object LXOR of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.BXOR

Object BXOR of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.MAXLOC

Object MAXLOC of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.MINLOC

Object MINLOC of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.REPLACE

Object REPLACE of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.NO_OP

Object NO_OP of type Op
  • x (Any)
  • y (Any)

Any


mpi4py.MPI.GROUP_NULL

Object GROUP_NULL of type Group

mpi4py.MPI.GROUP_EMPTY

Object GROUP_EMPTY of type Group

mpi4py.MPI.INFO_NULL

Object INFO_NULL of type Info

mpi4py.MPI.INFO_ENV

Object INFO_ENV of type Info

mpi4py.MPI.ERRHANDLER_NULL

Object ERRHANDLER_NULL of type Errhandler

mpi4py.MPI.ERRORS_RETURN

Object ERRORS_RETURN of type Errhandler

mpi4py.MPI.ERRORS_ABORT

Object ERRORS_ABORT of type Errhandler

mpi4py.MPI.ERRORS_ARE_FATAL

Object ERRORS_ARE_FATAL of type Errhandler

mpi4py.MPI.SESSION_NULL

Object SESSION_NULL of type Session

mpi4py.MPI.COMM_NULL

Object COMM_NULL of type Comm

mpi4py.MPI.COMM_SELF

Object COMM_SELF of type Intracomm

mpi4py.MPI.COMM_WORLD

Object COMM_WORLD of type Intracomm

mpi4py.MPI.WIN_NULL

Object WIN_NULL of type Win

mpi4py.MPI.FILE_NULL

Object FILE_NULL of type File

mpi4py.MPI.pickle


CITATION

If MPI for Python been significant to a project that leads to an academic publication, please acknowledge that fact by citing the project.


INSTALLATION

Build backends

mpi4py supports two different build backends: setuptools (default), scikit-build-core (CMake-based), and meson-python (Meson-based). The build backend can be selected by setting the MPI4PY_BUILD_BACKEND environment variable.

"setuptools", "scikit-build-core", "meson-python"
"setuptools"

Request a build backend for building mpi4py from sources.


Using setuptools

TIP:

Set the MPI4PY_BUILD_BACKEND environment variable to "setuptools" to use the setuptools build backend.


When using the default setuptools build backend, mpi4py relies on the legacy Python distutils framework to build C extension modules. The following environment variables affect the build configuration.

The mpicc compiler wrapper command is searched for in the executable search path (PATH environment variable) and used to compile the mpi4py.MPI C extension module. Alternatively, use the MPI4PY_BUILD_MPICC environment variable to the full path or command corresponding to the MPI-aware C compiler.

The mpicc compiler wrapper command is also used for linking the mpi4py.MPI C extension module. Alternatively, use the MPI4PY_BUILD_MPILD environment variable to specify the full path or command corresponding to the MPI-aware C linker.

If the MPI implementation does not provide a compiler wrapper, or it is not installed in a default system location, all relevant build information like include/library locations and library lists can be provided in an ini-style configuration file under a [mpi] section. mpi4py can then be asked to use the custom build information by setting the MPI4PY_BUILD_MPICFG environment variable to the full path of the configuration file. As an example, see the mpi.cfg file located in the top level mpi4py source directory.

Some vendor MPI implementations may not provide complete coverage of the MPI standard, or may provide partial features of newer MPI standard versions while advertising support for an older version. Setting the MPI4PY_BUILD_CONFIGURE environment variable to a non-empty string will trigger the run of exhaustive checks for the availability of all MPI constants, predefined handles, and routines.

The following environment variables are aliases for the ones described above. Having shorter names, they are convenient for occasional use in the command line. Its usage is not recommended in automation scenarios like packaging recipes, deployment scripts, and container image creation.

Convenience alias for MPI4PY_BUILD_MPICC.

Convenience alias for MPI4PY_BUILD_MPILD.

Convenience alias for MPI4PY_BUILD_MPICFG.

Using scikit-build-core

TIP:

Set the MPI4PY_BUILD_BACKEND environment variable to "scikit-build-core" to use the scikit-build-core build backend.


When using the scikit-build-core build backend, mpi4py delegates all of MPI build configuration to CMake’s FindMPI module. Besides the obvious advantage of cross-platform support, this delegation to CMake may be convenient in build environments exposing vendor software stacks via intricate module systems. Note however that mpi4py will not be able to look for MPI routines available beyond the MPI standard version the MPI implementation advertises to support (via the MPI_VERSION and MPI_SUBVERSION macro constants in the mpi.h header file), any missing MPI constant or symbol will prevent a successful build.

Using meson-python

TIP:

Set the MPI4PY_BUILD_BACKEND environment variable to "meson-python" to use the meson-python build backend.


When using the meson-python build backend, mpi4py delegates build tasks to the Meson build system.

WARNING:

mpi4py support for the meson-python build backend is experimental. For the time being, users must set the CC environment variable to the command or path corresponding to the mpicc C compiler wrapper.


Using pip

You can install the latest mpi4py release from its source distribution at PyPI using pip:

$ python -m pip install mpi4py


You can also install the in-development version with:

$ python -m pip install git+https://github.com/mpi4py/mpi4py


or:


NOTE:

Installing mpi4py from its source distribution (available at PyPI) or Git source code repository (available at GitHub) requires a C compiler and a working MPI implementation with development headers and libraries.


WARNING:

pip keeps previously built wheel files on its cache for future reuse. If you want to reinstall the mpi4py package using a different or updated MPI implementation, you have to either first remove the cached wheel file with:

$ python -m pip cache remove mpi4py


or ask pip to disable the cache:

$ python -m pip install --no-cache-dir mpi4py




Using conda

The conda-forge community provides ready-to-use binary packages from an ever growing collection of software libraries built around the multi-platform conda package manager. Four MPI implementations are available on conda-forge: Open MPI (Linux and macOS), MPICH (Linux and macOS), Intel MPI (Linux and Windows) and Microsoft MPI (Windows). You can install mpi4py and your preferred MPI implementation using the conda package manager:

to use MPICH do:

$ conda install -c conda-forge mpi4py mpich


to use Open MPI do:

$ conda install -c conda-forge mpi4py openmpi


to use Intel MPI do:

$ conda install -c conda-forge mpi4py impi_rt


to use Microsoft MPI do:

$ conda install -c conda-forge mpi4py msmpi



MPICH and many of its derivatives are ABI-compatible. You can provide the package specification mpich=X.Y.*=external_* (where X and Y are the major and minor version numbers) to request the conda package manager to use system-provided MPICH (or derivative) libraries. Similarly, you can provide the package specification openmpi=X.Y.*=external_* to use system-provided Open MPI libraries.

The openmpi package on conda-forge has built-in CUDA support, but it is disabled by default. To enable it, follow the instruction outlined during conda install. Additionally, UCX support is also available once the ucx package is installed.

WARNING:

Binary conda-forge packages are built with a focus on compatibility. The MPICH and Open MPI packages are build in a constrained environment with relatively dated OS images. Therefore, they may lack support for high-performance features like cross-memory attach (XPMEM/CMA). In production scenarios, it is recommended to use external (either custom-built or system-provided) MPI installations. See the relevant conda-forge documentation about using external MPI libraries .


Linux

On Fedora Linux systems (as well as RHEL and their derivatives using the EPEL software repository), you can install binary packages with the system package manager:

using dnf and the mpich package:

$ sudo dnf install python3-mpi4py-mpich


using dnf and the openmpi package:

$ sudo dnf install python3-mpi4py-openmpi



Please remember to load the correct MPI module for your chosen MPI implementation:

for the mpich package do:

$ module load mpi/mpich-$(arch)
$ python -c "from mpi4py import MPI"


for the openmpi package do:

$ module load mpi/openmpi-$(arch)
$ python -c "from mpi4py import MPI"



On Ubuntu Linux and Debian Linux systems, binary packages are available for installation using the system package manager:

$ sudo apt install python3-mpi4py


Note that on Ubuntu/Debian systems, the mpi4py package uses Open MPI. To use MPICH, install the libmpich-dev and python3-dev packages (and any other required development tools). Afterwards, install mpi4py from sources using pip.

macOS

macOS users can install mpi4py using the Homebrew package manager:

$ brew install mpi4py


Note that the Homebrew mpi4py package uses Open MPI. Alternatively, install the mpich package and next install mpi4py from sources using pip.

Windows

Windows users can install mpi4py from binary wheels hosted on the Python Package Index (PyPI) using pip:

$ python -m pip install mpi4py


The Windows wheels available on PyPI are specially crafted to work with either the Intel MPI or the Microsoft MPI runtime, therefore requiring a separate installation of any one of these packages.

Intel MPI is under active development and supports recent version of the MPI standard. Intel MPI can be installed with pip (see the impi-rt package on PyPI), being therefore straightforward to get it up and running within a Python environment. Intel MPI can also be installed system-wide as part of the Intel HPC Toolkit for Windows or via standalone online/offline installers.

DEVELOPMENT

Prerequisites

You need to have the following software properly installed to develop MPI for Python:

  • Python 3.6 or above.
  • The Cython compiler.
  • A working MPI implementation like MPICH or Open MPI, preferably supporting MPI-4 and built with shared/dynamic libraries.

Optionally, consider installing the following packages:

  • NumPy for enabling comprehensive testing of MPI communication.
  • CuPy for enabling comprehensive testing with a GPU-aware MPI.
  • Sphinx to build the documentation.

TIP:

Most routine development tasks like building, installing in editable mode, testing, and generating documentation can be performed with the spin developer tool. Run spin at the top level source directory for a list of available subcommands.


Building

MPI for Python uses setuptools-based build system that relies on the setup.py file. Some setuptools commands (e.g., build) accept additional options:

Lets you pass a section with MPI configuration within a special configuration file. Alternatively, you can use the MPICFG environment variable.

Specify the path or name of the mpicc C compiler wrapper. Alternatively, use the MPICC environment variable.

Specify the full path or name for the MPI-aware C linker. Alternatively, use the MPILD environment variable. If not set, the mpicc C compiler wrapper is used for linking.

Runs exhaustive tests for checking about missing MPI types, constants, and functions. This option should be passed in order to build MPI for Python against old MPI-1, MPI-2, or MPI-3 implementations, possibly providing a subset of MPI-4.

If you use a MPI implementation providing a mpicc C compiler wrapper (e.g., MPICH or Open MPI), it will be used for compilation and linking. This is the preferred and easiest way to build MPI for Python.

If mpicc is found in the executable search path (PATH environment variable), simply run the build command:

$ python setup.py build


If mpicc is not in your search path or the compiler wrapper has a different name, you can run the build command specifying its location, either via the --mpicc command option or using the MPICC environment variable:

$ python setup.py build --mpicc=/path/to/mpicc
$ env MPICC=/path/to/mpicc python setup.py build


Alternatively, you can provide all the relevant information about your MPI implementation by editing the mpi.cfg file located in the top level source directory. You can use the default section [mpi] or add a new custom section, for example [vendor_mpi] (see the examples provided in the mpi.cfg file as a starting point to write your own section):

[mpi]
include_dirs         = /usr/local/mpi/include
libraries            = mpi
library_dirs         = /usr/local/mpi/lib
runtime_library_dirs = /usr/local/mpi/lib
[vendor_mpi]
include_dirs         = /opt/mpi/include ...
libraries            = mpi ...
library_dirs         = /opt/mpi/lib ...
runtime_library_dirs = /opt/mpi/lib ...
...


and then run the build command specifying you custom configuration section:

$ python setup.py build --mpi=vendor_mpi
$ env MPICFG=vendor_mpi python setup.py build


Installing

MPI for Python can be installed in editable mode:

$ python -m pip install --editable .


After modifying Cython sources, an in-place rebuild is needed:

$ python setup.py build --inplace


Testing

To quickly test the installation:

$ mpiexec -n 5 python -m mpi4py.bench helloworld
Hello, World! I am process 0 of 5 on localhost.
Hello, World! I am process 1 of 5 on localhost.
Hello, World! I am process 2 of 5 on localhost.
Hello, World! I am process 3 of 5 on localhost.
Hello, World! I am process 4 of 5 on localhost.
$ mpiexec -n 5 python -m mpi4py.bench ringtest -l 10 -n 1048576
time for 10 loops = 0.00361614 seconds (5 processes, 1048576 bytes)


If you installed from a git clone or the source distribution, issuing at the command line:

$ mpiexec -n 5 python demo/helloworld.py


will launch a five-process run of the Python interpreter and run the demo script demo/helloworld.py from the source distribution.

You can also run all the unittest scripts:

$ mpiexec -n 5 python test/main.py


or, if you have the pytest unit testing framework installed:

$ mpiexec -n 5 pytest


GUIDELINES

Fair play

Summary

This section defines Rules of Play for companies and outside developers that engage with the mpi4py project. It covers:

  • Restrictions on use of the mpi4py name.
  • How and whether to publish a modified distribution.
  • How to make us aware of patched versions.

After reading this section, companies and developers will know what kinds of behavior the mpi4py developers and contributors would like to see, and which we consider troublesome, bothersome, and unacceptable.

This document is a close adaptation of NumPy NEP 36.

Motivation

Occasionally, we learn of modified mpi4py versions and binary distributions circulated by outsiders. These patched versions can cause problems to mpi4py users (see, e.g., mpi4py/mpi4py#508). When issues like these arise, our developers waste time identifying the problematic release, locating alterations, and determining an appropriate course of action.

In addition, packages on the Python Packaging Index are sometimes named such that users assume they are sanctioned or maintained by the mpi4py developers. We wish to reduce the number of such incidents.

Scope

This document aims to define a minimal set of rules that, when followed, will be considered good-faith efforts in line with the expectations of the mpi4py developers and contributors.

Our hope is that companies and outside developers who feel they need to modify mpi4py will first consider contributing to the project, or use alternative mechanisms for patching and extending mpi4py.

When in doubt, please talk to us first. We may suggest an alternative; at minimum, we’ll be informed and we may even grant an exception if deemed appropriate.

Fair play rules

1.
Do not reuse the mpi4py name for projects not affiliated with the mpi4py project.

At time of writing, there are only a handful of mpi4py-named packages developed by the mpi4py project, including mpi4py and mpi4py-fft. We ask that outside packages not include the phrase mpi4py, i.e., avoid names such as mycompany-mpi4py or mpi4py-mycompany.

To be clear, this rule only applies to modules (package names); it is perfectly acceptable to have a submodule of your own package named mycompany.mpi4py.

2.
Do not publish binary mpi4py wheels on PyPI (https://pypi.org/).

We ask companies and outside developers to not publish binary mpi4py wheels in the main Python Package Index (https://pypi.org/) under names such mpi4py-mpich, mpi4py-openmpi, or mpi4py-vendor_mpi.

The usual approaches to build binary Python wheels involve the embedding of dependent shared libraries. While such an approach may seem convenient and often is, in the particular case of MPI and mpi4py it is ultimately harmful to end users. Embedding the MPI shared libraries would prevent the use of external, system-provided MPI installations with hardware-specific optimizations and site-specific tweaks.

The MPI Forum is currently discussing the standardization of a proposal for an Application Binary Interface (ABI) for MPI, see [mpi-abi-paper] and [mpi-abi-issue]. Such standardization will allow for any binary dependent on the MPI library to be used with multiple MPI backends. Once this proposal becomes part of the MPI standard, the mpi4py project will consider publishing on PyPI binary wheels capable of using any backend MPI implementation supporting the new MPI ABI specification. In the mean time, mpi4py is currently distributing experimental MPI and mpi4py binary wheels on https://anaconda.org/mpi4py.

[mpi-abi-paper]
J. Hammond, L. Dalcin, E. Schnetter, M. Pérache, J. B. Besnard, J. Brown, G. Brito Gadeschi, S. Byrne, J. Schuchart, and H. Zhou. MPI Application Binary Interface Standardization. EuroMPI 2023, Bristol, UK, September 2023. https://doi.org/10.1145/3615318.3615319
[mpi-abi-issue]
MPI Forum GitHub Issue: MPI needs a standard ABI. https://github.com/mpi-forum/mpi-issues/issues/751
3.
Do not republish modified versions of mpi4py.

Modified versions of mpi4py make it very difficult for the developers to address bug reports, since we typically do not know which parts of mpi4py have been modified.

If you have to break this rule (and we implore you not to!), then make it clear in the __version__ tag that you have modified mpi4py, e.g.:

>>> print(mpi4py.__version__)
'4.0.0+mycompany.13`


We understand that minor patches are often required to make a library work inside of a package ecosystem. This is totally acceptable, but we ask that no substantive changes are made.

4.
Do not extend or modify mpi4py’s API.

If you absolutely have to break the previous rule, please do not add additional functions to the namespace, or modify the API of existing functions. Having additional functions exposed in distributed versions is confusing for users and developers alike.


LICENSE

Copyright (c) 2024, Lisandro Dalcin

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

1.
Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
2.
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
3.
Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

CHANGES

Release 4.0.0 [2024-07-28]

New features:
Add support for the MPI-4.0 standard.
  • Use large count MPI-4 routines.
  • Add persistent collective communication.
  • Add partitioned point-to-point communication.
  • Add new communicator constructors.
  • Add the Session class and its methods.

Add support for the MPI-4.1 standard.
  • Add non-destructive completion test for multiple requests.
  • Add value-index datatype constructor.
  • Add communicator/session buffer attach/detach/flush.
  • Support for removal of error classes/codes/strings.
  • Support for querying hardware resource information.

Add preliminary support for the upcoming MPI-5.0 standard.
User-level failure mitigation (ULFM).

  • mpi4py.util.pool: New drop-in replacement for multiprocessing.pool.
  • mpi4py.util.sync: New synchronization utilities.
  • Add runtime check for mismatch between mpiexec and MPI library.
  • Support scikit-build-core as an alternative build backend.

Support meson-python as an alternative build backend.

Enhancements:
  • mpi4py.futures: Support for parallel tasks.
  • mpi4py.futures: Report exception tracebacks in workers.
  • mpi4py.util.pkl5: Add support for collective communication.
  • Add methods Datatype.fromcode(), Datatype.tocode() and attributes Datatype.typestr, Datatype.typechar to simplify NumPy interoperability for simple cases.
  • Add methods Comm.Create_errhandler(), Win.Create_errhandler(), and File.Create_errhandler() to create custom error handlers.
  • Add support for pickle serialization of instances of MPI types. All instances of Datatype, Info, and Status can be serialized. Instances of Op can be serialized only if created through mpi4py by calling Op.Create(). Instances of other MPI types can be serialized only if they reference predefined handles.
  • Add handle attribute and fromhandle() class method to MPI classes to ease interoperability with external code. The handle value is a unsigned integer guaranteed to fit on the platform’s uintptr_t C type.
  • Add lowercase free() method to MPI classes to ease MPI object deallocation and cleanup. This method eventually attempts to call Free(), but only if the object’s MPI handle is not a null or predefined handle, and such call is allowed within the World Model init/finalize.

Backward-incompatible changes:
  • Python 2 is no longer supported, Python 3.6+ is required, but typing stubs are supported for Python 3.8+.
  • The Intracomm.Create_group() method is no longer defined in the base Comm class.
  • Group.Compare() and Comm.Compare() are no longer class methods but instance methods. Existing codes using the former class methods are expected to continue working.
  • Group.Translate_ranks() is no longer a class method but an instance method. Existing codes using the former class method are expected to continue working.
  • The LB and UB datatypes are no longer available, use Datatype.Create_resized() instead.
  • The HOST predefined attribute key is no longer available.
  • The MPI.memory class has been renamed to MPI.buffer. The old name is still available as an alias to the new name.
  • The mpi4py.dl module is no longer available.
  • The mpi4py.get_config function returns an empty dictionary.

Miscellaneous:
  • The project is now licenced under the BSD-3-Clause license. This change is fairly inconsequential for users and distributors. It simply adds an additional clause against using contributor names for promotional purposes without their consent.
  • Add a new guidelines section to documentation laying out new fair play rules. These rules ask companies and outside developers to refrain from reusing the mpi4py name in unaffiliated projects, publishing binary mpi4py wheels on the main Python Package Index (PyPI), and distributing modified versions with incompatible or extended API changes. The primary motivation of these rules is to avoid fragmentation and end-user confusion.


Release 3.1.6 [2024-04-14]

WARNING:

This is the last release supporting Python 2.


Fix various build issues.

Release 3.1.5 [2023-10-04]

WARNING:

This is the last release supporting Python 2.


Rebuild C sources with Cython 0.29.36 to support Python 3.12.

Release 3.1.4 [2022-11-02]

WARNING:

This is the last release supporting Python 2.


  • Rebuild C sources with Cython 0.29.32 to support Python 3.11.
  • Fix contiguity check for DLPack and CAI buffers.
  • Workaround build failures with setuptools v60.

Release 3.1.3 [2021-11-25]

WARNING:

This is the last release supporting Python 2.


Add missing support for MPI.BOTTOM to generalized all-to-all collectives.

Release 3.1.2 [2021-11-04]

WARNING:

This is the last release supporting Python 2.


  • mpi4py.futures: Add _max_workers property to MPIPoolExecutor.
  • mpi4py.util.dtlib: Fix computation of alignment for predefined datatypes.
  • mpi4py.util.pkl5: Fix deadlock when using ssend() + mprobe().
  • mpi4py.util.pkl5: Add environment variable MPI4PY_PICKLE_THRESHOLD.
  • mpi4py.rc: Interpret "y" and "n" strings as boolean values.
  • Fix/add typemap/typestr for MPI.WCHAR/MPI.COUNT datatypes.
  • Minor fixes and additions to documentation.
  • Minor fixes to typing support.
  • Support for local version identifier (PEP-440).

Release 3.1.1 [2021-08-14]

WARNING:

This is the last release supporting Python 2.


  • Fix typo in Requires-Python package metadata.
  • Regenerate C sources with Cython 0.29.24.

Release 3.1.0 [2021-08-12]

WARNING:

This is the last release supporting Python 2.


New features:
mpi4py.util: New package collecting miscellaneous utilities.

Enhancements:
  • Add pickle-based Request.waitsome() and Request.testsome().
  • Add lowercase methods Request.get_status() and Request.cancel().
  • Support for passing Python GPU arrays compliant with the DLPack data interchange mechanism (link) and the __cuda_array_interface__ (CAI) standard (link) to uppercase methods. This support requires that mpi4py is built against CUDA-aware MPI implementations. This feature is currently experimental and subject to future changes.
  • mpi4py.futures: Add support for initializers and canceling futures at shutdown. Environment variables names now follow the pattern MPI4PY_FUTURES_*, the previous MPI4PY_* names are deprecated.
  • Add type annotations to Cython code. The first line of the docstring of functions and methods displays a signature including type annotations.
  • Add companion stub files to support type checkers.
  • Support for weak references.

Miscellaneous:
Add a new mpi4py publication (link) to the citation listing.


Release 3.0.3 [2019-11-04]

Regenerate Cython wrappers to support Python 3.8.

Release 3.0.2 [2019-06-11]

Bug fixes:
  • Fix handling of readonly buffers in support for Python 2 legacy buffer interface. The issue triggers only when using a buffer-like object that is readonly and does not export the new Python 3 buffer interface.
  • Fix build issues with Open MPI 4.0.x series related to removal of many MPI-1 symbols deprecated in MPI-2 and removed in MPI-3.
  • Minor documentation fixes.


Release 3.0.1 [2019-02-15]

Bug fixes:
  • Fix Comm.scatter() and other collectives corrupting input send list. Add safety measures to prevent related issues in global reduction operations.
  • Fix error-checking code for counts in Op.Reduce_local().

Enhancements:
  • Map size-specific Python/NumPy typecodes to MPI datatypes.
  • Allow partial specification of target list/tuple arguments in the various Win RMA methods.
  • Workaround for removal of MPI_{LB|UB} in Open MPI 4.0.
  • Support for Microsoft MPI v10.0.


Release 3.0.0 [2017-11-08]

New features:
  • mpi4py.futures: Execute computations asynchronously using a pool of MPI processes. This package is based on concurrent.futures from the Python standard library.
  • mpi4py.run: Run Python code and abort execution in case of unhandled exceptions to prevent deadlocks.
  • mpi4py.bench: Run basic MPI benchmarks and tests.

Enhancements:
  • Lowercase, pickle-based collective communication calls are now thread-safe through the use of fine-grained locking.
  • The MPI module now exposes a memory type which is a lightweight variant of the builtin memoryview type, but exposes both the legacy Python 2 and the modern Python 3 buffer interface under a Python 2 runtime.
  • The MPI.Comm.Alltoallw() method now uses count=1 and displ=0 as defaults, assuming that messages are specified through user-defined datatypes.
  • The Request.Wait[all]() methods now return True to match the interface of Request.Test[all]().
  • The Win class now implements the Python buffer interface.

Backward-incompatible changes:
  • The buf argument of the MPI.Comm.recv() method is deprecated, passing anything but None emits a warning.
  • The MPI.Win.memory property was removed, use the MPI.Win.tomemory() method instead.
  • Executing python -m mpi4py in the command line is now equivalent to python -m mpi4py.run. For the former behavior, use python -m mpi4py.bench.
  • Python 2.6 and 3.2 are no longer supported. The mpi4py.MPI module may still build and partially work, but other pure-Python modules under the mpi4py namespace will not.
  • Windows: Remove support for legacy MPICH2, Open MPI, and DeinoMPI.


Release 2.0.0 [2015-10-18]

Support for MPI-3 features.
  • Matched probes and receives.
  • Nonblocking collectives.
  • Neighborhood collectives.
  • New communicator constructors.
  • Request-based RMA operations.
  • New RMA communication and synchronisation calls.
  • New window constructors.
  • New datatype constructor.
  • New C++ boolean and floating complex datatypes.

Support for MPI-2 features not included in previous releases.
  • Generalized All-to-All collective (Comm.Alltoallw())
  • User-defined data representations (Register_datarep())

  • New scalable implementation of reduction operations for Python objects. This code is based on binomial tree algorithms using point-to-point communication and duplicated communicator contexts. To disable this feature, use mpi4py.rc.fast_reduce = False.
  • Backward-incompatible changes:
  • Python 2.4, 2.5, 3.0 and 3.1 are no longer supported.
  • Default MPI error handling policies are overridden. After import, mpi4py sets the ERRORS_RETURN error handler in COMM_SELF and COMM_WORLD, as well as any new Comm, Win, or File instance created through mpi4py, thus effectively ignoring the MPI rules about error handler inheritance. This way, MPI errors translate to Python exceptions. To disable this behavior and use the standard MPI error handling rules, use mpi4py.rc.errors = 'default'.
  • Change signature of all send methods, dest is a required argument.
  • Change signature of all receive and probe methods, source defaults to ANY_SOURCE, tag defaults to ANY_TAG.
  • Change signature of send lowercase-spelling methods, obj arguments are not mandatory.
  • Change signature of recv lowercase-spelling methods, renamed ‘obj’ arguments to ‘buf’.
  • Change Request.Waitsome() and Request.Testsome() to return None or list.
  • Change signature of all lowercase-spelling collectives, sendobj arguments are now mandatory, recvobj arguments were removed.
  • Reduction operations MAXLOC and MINLOC are no longer special-cased in lowercase-spelling methods Comm.[all]reduce() and Comm.[ex]scan(), the input object must be specified as a tuple (obj, location).
  • Change signature of name publishing functions. The new signatures are Publish_name(service_name, port_name, info=INFO_NULL) and Unpublish_name(service_name, port_name, info=INFO_NULL)`.
  • Win instances now cache Python objects exposing memory by keeping references instead of using MPI attribute caching.
  • Change signature of Win.Lock(). The new signature is Win.Lock(rank, lock_type=LOCK_EXCLUSIVE, assertion=0).
  • Move Cartcomm.Map() to Intracomm.Cart_map().
  • Move Graphcomm.Map() to Intracomm.Graph_map().
  • Remove the mpi4py.MPE module.
  • Rename the Cython definition file for use with cimport statement from mpi_c.pxd to libmpi.pxd.


Release 1.3.1 [2013-08-07]

  • Regenerate C wrappers with Cython 0.19.1 to support Python 3.3.
  • Install *.pxd files in <site-packages>/mpi4py to ease the support for Cython’s cimport statement in code requiring to access mpi4py internals.
  • As a side-effect of using Cython 0.19.1, ancient Python 2.3 is no longer supported. If you really need it, you can install an older Cython and run python setup.py build_src --force.

Release 1.3 [2012-01-20]

  • Now Comm.recv() accept a buffer to receive the message.
  • Add Comm.irecv() and Request.{wait|test}[any|all]().
  • Add Intracomm.Spawn_multiple().
  • Better buffer handling for PEP 3118 and legacy buffer interfaces.
  • Add support for attribute attribute caching on communicators, datatypes and windows.
  • Install MPI-enabled Python interpreter as <path>/mpi4py/bin/python-mpi.
  • Windows: Support for building with Open MPI.

Release 1.2.2 [2010-09-13]

  • Add mpi4py.get_config() to retrieve information (compiler wrappers, includes, libraries, etc) about the MPI implementation employed to build mpi4py.
  • Workaround Python libraries with missing GILState-related API calls in case of non-threaded Python builds.
  • Windows: look for MPICH2, DeinoMPI, Microsoft HPC Pack at their default install locations under %ProgramFiles.
  • MPE: fix hacks related to old API’s, these hacks are broken when MPE is built with a MPI implementations other than MPICH2.
  • HP-MPI: fix for missing Fortran datatypes, use dlopen() to load the MPI shared library before MPI_Init()
  • Many distutils-related fixes, cleanup, and enhancements, better logics to find MPI compiler wrappers.
  • Support for pip install mpi4py.

Release 1.2.1 [2010-02-26]

  • Fix declaration in Cython include file. This declaration, while valid for Cython, broke the simple-minded parsing used in conf/mpidistutils.py to implement configure-tests for availability of MPI symbols.
  • Update SWIG support and make it compatible with Python 3. Also generate an warning for SWIG < 1.3.28.
  • Fix distutils-related issues in Mac OS X. Now ARCHFLAGS environment variable is honored of all Python’s config/Makefile variables.
  • Fix issues with Open MPI < 1.4.2 related to error checking and MPI_XXX_NULL handles.

Release 1.2 [2009-12-29]

  • Automatic MPI datatype discovery for NumPy arrays and PEP-3118 buffers. Now buffer-like objects can be messaged directly, it is no longer required to explicitly pass a 2/3-list/tuple like [data, MPI.DOUBLE], or [data, count, MPI.DOUBLE]. Only basic types are supported, i.e., all C/C99-native signed/unsigned integral types and single/double precision real/complex floating types. Many thanks to Eilif Muller for the initial feedback.
  • Nonblocking send of pickled Python objects. Many thanks to Andreas Kloeckner for the initial patch and enlightening discussion about this enhancement.
  • Request instances now hold a reference to the Python object exposing the buffer involved in point-to-point communication or parallel I/O. Many thanks to Andreas Kloeckner for the initial feedback.
  • Support for logging of user-defined states and events using MPE. Runtime (i.e., without requiring a recompile!) activation of logging of all MPI calls is supported in POSIX platforms implementing dlopen().
  • Support for all the new features in MPI-2.2 (new C99 and F90 datatypes, distributed graph topology, local reduction operation, and other minor enhancements).
  • Fix the annoying issues related to Open MPI and Python dynamic loading of extension modules in platforms supporting dlopen().
  • Fix SLURM dynamic loading issues on SiCortex. Many thanks to Ian Langmore for providing me shell access.

Release 1.1.0 [2009-06-06]

  • Fix bug in Comm.Iprobe() that caused segfaults as Python C-API calls were issued with the GIL released (issue #2).
  • Add Comm.bsend() and Comm.ssend() for buffered and synchronous send semantics when communicating general Python objects.
  • Now the call Info.Get(key) return a single value (i.e, instead of a 2-tuple); this value is None if key is not in the Info object, or a string otherwise. Previously, the call redundantly returned (None, False) for missing key-value pairs; None is enough to signal a missing entry.
  • Add support for parametrized Fortran datatypes.
  • Add support for decoding user-defined datatypes.
  • Add support for user-defined reduction operations on memory buffers. However, at most 16 user-defined reduction operations can be created. Ask the author for more room if you need it.

Release 1.0.0 [2009-03-20]

This is the fist release of the all-new, Cython-based, implementation of MPI for Python. Unfortunately, this implementation is not backward-compatible with the previous one. The list below summarizes the more important changes that can impact user codes.

Some communication calls had overloaded functionality. Now there is a clear distinction between communication of general Python object with pickle, and (fast, near C-speed) communication of buffer-like objects (e.g., NumPy arrays).
  • for communicating general Python objects, you have to use all-lowercase methods, like send(), recv(), bcast(), etc.
  • for communicating array data, you have to use Send(), Recv(), Bcast(), etc. methods. Buffer arguments to these calls must be explicitly specified by using a 2/3-list/tuple like [data, MPI.DOUBLE], or [data, count, MPI.DOUBLE] (the former one uses the byte-size of data and the extent of the MPI datatype to define the count).

Indexing a communicator with an integer returned a special object associating the communication with a target rank, alleviating you from specifying source/destination/root arguments in point-to-point and collective communications. This functionality is no longer available, expressions like:

MPI.COMM_WORLD[0].Send(...)
MPI.COMM_WORLD[0].Recv(...)
MPI.COMM_WORLD[0].Bcast(...)


have to be replaced by:

MPI.COMM_WORLD.Send(..., dest=0)
MPI.COMM_WORLD.Recv(..., source=0)
MPI.COMM_WORLD.Bcast(..., root=0)


  • Automatic MPI initialization (i.e., at import time) requests the maximum level of MPI thread support (i.e., it is done by calling MPI_Init_thread() and passing MPI_THREAD_MULTIPLE). In case you need to change this behavior, you can tweak the contents of the mpi4py.rc module.
  • In order to obtain the values of predefined attributes attached to the world communicator, now you have to use the Get_attr() method on the MPI.COMM_WORLD instance:

tag_ub = MPI.COMM_WORLD.Get_attr(MPI.TAG_UB)


  • In the previous implementation, MPI.COMM_WORLD and MPI.COMM_SELF were associated to duplicates of the (C-level) MPI_COMM_WORLD and MPI_COMM_SELF predefined communicator handles. Now this is no longer the case, MPI.COMM_WORLD and MPI.COMM_SELF proxies the actual MPI_COMM_WORLD and MPI_COMM_SELF handles.
  • Convenience aliases MPI.WORLD and MPI.SELF were removed. Use instead MPI.COMM_WORLD and MPI.COMM_SELF.
  • Convenience constants MPI.WORLD_SIZE and MPI.WORLD_RANK were removed. Use instead MPI.COMM_WORLD.Get_size() and MPI.COMM_WORLD.Get_rank().

AUTHOR

Lisandro Dalcin

COPYRIGHT

2024, Lisandro Dalcin

September 29, 2024 4.0