.\" Man page generated from reStructuredText. . . .nr rst2man-indent-level 0 . .de1 rstReportMargin \\$1 \\n[an-margin] level \\n[rst2man-indent-level] level margin: \\n[rst2man-indent\\n[rst2man-indent-level]] - \\n[rst2man-indent0] \\n[rst2man-indent1] \\n[rst2man-indent2] .. .de1 INDENT .\" .rstReportMargin pre: . RS \\$1 . nr rst2man-indent\\n[rst2man-indent-level] \\n[an-margin] . nr rst2man-indent-level +1 .\" .rstReportMargin post: .. .de UNINDENT . RE .\" indent \\n[an-margin] .\" old: \\n[rst2man-indent\\n[rst2man-indent-level]] .nr rst2man-indent-level -1 .\" new: \\n[rst2man-indent\\n[rst2man-indent-level]] .in \\n[rst2man-indent\\n[rst2man-indent-level]]u .. .TH "MPI4PY" "1" "November 05, 2021" "3.1" "MPI for Python" .SH NAME mpi4py \- MPI for Python .INDENT 0.0 .TP .B Author Lisandro Dalcin .TP .B Contact \fI\%dalcinl@gmail.com\fP .TP .B Date November 05, 2021 .UNINDENT .SS Abstract .sp This document describes the \fIMPI for Python\fP package. \fIMPI for Python\fP provides Python bindings for the \fIMessage Passing Interface\fP (MPI) standard, allowing Python applications to exploit multiple processors on workstations, clusters and supercomputers. .sp 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 \fIpicklable\fP 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). .SH INTRODUCTION .sp 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 \fI\%Beowulf\fP class clusters and cluster of workstations. .sp 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. .sp 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. .sp 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. .IP [1] 5 MATLAB is a registered trademark of The MathWorks, Inc. .SS What is MPI? .sp \fI\%MPI\fP, [mpi\-using] [mpi\-ref] the \fIMessage Passing Interface\fP, 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++). .sp 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 \fI\%MPICH\fP [mpi\-mpich] and \fI\%Open MPI\fP [mpi\-openmpi]\&. .SS What is Python? .sp \fI\%Python\fP 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. .sp 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. .sp 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. .SS Related Projects .sp As this work started and evolved, some ideas were borrowed from well known MPI and Python related open source projects from the Internet. .INDENT 0.0 .IP \(bu 2 \fI\%OOMPI\fP .INDENT 2.0 .IP \(bu 2 It has no relation with Python, but is an excellent object oriented approach to MPI. .IP \(bu 2 It is a C++ class library specification layered on top of the C bindings that encapsulates MPI into a functional class hierarchy. .IP \(bu 2 It provides a flexible and intuitive interface by adding some abstractions, like \fIPorts\fP and \fIMessages\fP, which enrich and simplify the syntax. .UNINDENT .IP \(bu 2 \fI\%Pypar\fP .INDENT 2.0 .IP \(bu 2 Its interface is rather minimal. There is no support for communicators or process topologies. .IP \(bu 2 It does not require the Python interpreter to be modified or recompiled, but does not permit interactive parallel runs. .IP \(bu 2 General (\fIpicklable\fP) Python objects of any type can be communicated. There is good support for numeric arrays, practically full MPI bandwidth can be achieved. .UNINDENT .IP \(bu 2 \fI\%pyMPI\fP .INDENT 2.0 .IP \(bu 2 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. .IP \(bu 2 It provides an interface suitable for basic parallel programing. There is not full support for defining new communicators or process topologies. .IP \(bu 2 General (picklable) Python objects can be messaged between processors. There is not support for numeric arrays. .UNINDENT .IP \(bu 2 \fI\%Scientific Python\fP .INDENT 2.0 .IP \(bu 2 It provides a collection of Python modules that are useful for scientific computing. .IP \(bu 2 There is an interface to MPI and BSP (\fIBulk Synchronous Parallel programming\fP). .IP \(bu 2 The interface is simple but incomplete and does not resemble the MPI specification. There is support for numeric arrays. .UNINDENT .UNINDENT .sp Additionally, we would like to mention some available tools for scientific computing and software development with Python. .INDENT 0.0 .IP \(bu 2 \fI\%NumPy\fP 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. .IP \(bu 2 \fI\%SciPy\fP 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. .IP \(bu 2 \fI\%Cython\fP 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. .IP \(bu 2 \fI\%SWIG\fP 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. .UNINDENT .IP [mpi-std1] 5 MPI Forum. MPI: A Message Passing Interface Standard. International Journal of Supercomputer Applications, volume 8, number 3\-4, pages 159\-416, 1994. .IP [mpi-std2] 5 MPI Forum. MPI: A Message Passing Interface Standard. High Performance Computing Applications, volume 12, number 1\-2, pages 1\-299, 1998. .IP [mpi-using] 5 William Gropp, Ewing Lusk, and Anthony Skjellum. Using MPI: portable parallel programming with the message\-passing interface. MIT Press, 1994. .IP [mpi-ref] 5 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. .IP [mpi-mpich] 5 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. .IP [mpi-openmpi] 5 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. .IP [Hinsen97] 5 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. .IP [Beazley97] 5 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. .SH OVERVIEW .sp 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. .SS Communicating Python Objects and Array Data .sp The Python standard library supports different mechanisms for data persistence. Many of them rely on disk storage, but \fIpickling\fP and \fImarshaling\fP can also work with memory buffers. .sp The \fBpickle\fP modules provide user\-extensible facilities to serialize general Python objects using ASCII or binary formats. The \fBmarshal\fP module provides facilities to serialize built\-in Python objects using a binary format specific to Python, but independent of machine architecture issues. .sp \fIMPI for Python\fP can communicate any built\-in or user\-defined Python object taking advantage of the features provided by the \fBpickle\fP 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). .sp Although simple and general, the serialization approach (i.e., \fIpickling\fP and \fIunpickling\fP) 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 \fIa priori\fP, 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. .sp \fIMPI for Python\fP 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. .SS Communicators .sp In \fIMPI for Python\fP, \fBComm\fP is the base class of communicators. The \fBIntracomm\fP and \fBIntercomm\fP classes are sublcasses of the \fBComm\fP class. The \fBComm.Is_inter\fP method (and \fBComm.Is_intra\fP, 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. .sp The two predefined intracommunicator instances are available: \fBCOMM_SELF\fP and \fBCOMM_WORLD\fP\&. From them, new communicators can be created as needed. .sp The number of processes in a communicator and the calling process rank can be respectively obtained with methods \fBComm.Get_size\fP and \fBComm.Get_rank\fP\&. The associated process group can be retrieved from a communicator by calling the \fBComm.Get_group\fP method, which returns an instance of the \fBGroup\fP class. Set operations with \fBGroup\fP objects like like \fBGroup.Union\fP, \fBGroup.Intersection\fP and \fBGroup.Difference\fP are fully supported, as well as the creation of new communicators from these groups using \fBComm.Create\fP and \fBComm.Create_group\fP\&. .sp New communicator instances can be obtained with the \fBComm.Clone\fP, \fBComm.Dup\fP and \fBComm.Split\fP methods, as well methods \fBIntracomm.Create_intercomm\fP and \fBIntercomm.Merge\fP\&. .sp Virtual topologies (\fBCartcomm\fP, \fBGraphcomm\fP and \fBDistgraphcomm\fP classes, which are specializations of the \fBIntracomm\fP class) are fully supported. New instances can be obtained from intracommunicator instances with factory methods \fBIntracomm.Create_cart\fP and \fBIntracomm.Create_graph\fP\&. .SS Point\-to\-Point Communications .sp 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. .sp MPI provides a set of \fIsend\fP and \fIreceive\fP functions allowing the communication of \fItyped\fP data with an associated \fItag\fP\&. 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. .SS Blocking Communications .sp MPI provides basic send and receive functions that are \fIblocking\fP\&. These functions block the caller until the data buffers involved in the communication can be safely reused by the application program. .sp In \fIMPI for Python\fP, the \fBComm.Send\fP, \fBComm.Recv\fP and \fBComm.Sendrecv\fP methods of communicator objects provide support for blocking point\-to\-point communications within \fBIntracomm\fP and \fBIntercomm\fP instances. These methods can communicate memory buffers. The variants \fBComm.send\fP, \fBComm.recv\fP and \fBComm.sendrecv\fP can communicate general Python objects. .SS Nonblocking Communications .sp 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. .sp MPI provides \fInonblocking\fP 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. .sp In \fIMPI for Python\fP, the \fBComm.Isend\fP and \fBComm.Irecv\fP methods initiate send and receive operations, respectively. These methods return a \fBRequest\fP instance, uniquely identifying the started operation. Its completion can be managed using the \fBRequest.Test\fP, \fBRequest.Wait\fP and \fBRequest.Cancel\fP methods. The management of \fBRequest\fP 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. .SS Persistent Communications .sp 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. .sp In \fIMPI for Python\fP, the \fBComm.Send_init\fP and \fBComm.Recv_init\fP methods create persistent requests for a send and receive operation, respectively. These methods return an instance of the \fBPrequest\fP class, a subclass of the \fBRequest\fP class. The actual communication can be effectively started using the \fBPrequest.Start\fP method, and its completion can be managed as previously described. .SS Collective Communications .sp 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 \fItyped\fP data, but messages are not paired with an associated \fItag\fP; selectivity of messages is implied in the calling order. Additionally, collective functions come in blocking versions only. .sp The more commonly used collective communication operations are the following. .INDENT 0.0 .IP \(bu 2 Barrier synchronization across all group members. .IP \(bu 2 Global communication functions .INDENT 2.0 .IP \(bu 2 Broadcast data from one member to all members of a group. .IP \(bu 2 Gather data from all members to one member of a group. .IP \(bu 2 Scatter data from one member to all members of a group. .UNINDENT .IP \(bu 2 Global reduction operations such as sum, maximum, minimum, etc. .UNINDENT .sp In \fIMPI for Python\fP, the \fBComm.Bcast\fP, \fBComm.Scatter\fP, \fBComm.Gather\fP, \fBComm.Allgather\fP, \fBComm.Alltoall\fP methods provide support for collective communications of memory buffers. The lower\-case variants \fBComm.bcast\fP, \fBComm.scatter\fP, \fBComm.gather\fP, \fBComm.allgather\fP and \fBComm.alltoall\fP can communicate general Python objects. The vector variants (which can communicate different amounts of data to each process) \fBComm.Scatterv\fP, \fBComm.Gatherv\fP, \fBComm.Allgatherv\fP, \fBComm.Alltoallv\fP and \fBComm.Alltoallw\fP are also supported, they can only communicate objects exposing memory buffers. .sp Global reducion operations on memory buffers are accessible through the \fBComm.Reduce\fP, \fBComm.Reduce_scatter\fP, \fBComm.Allreduce\fP, \fBIntracomm.Scan\fP and \fBIntracomm.Exscan\fP methods. The lower\-case variants \fBComm.reduce\fP, \fBComm.allreduce\fP, \fBIntracomm.scan\fP and \fBIntracomm.exscan\fP can communicate general Python objects; however, the actual required reduction computations are performed sequentially at some process. All the predefined (i.e., \fBSUM\fP, \fBPROD\fP, \fBMAX\fP, etc.) reduction operations can be applied. .SS Support for GPU\-aware MPI .sp Several MPI implementations, including Open MPI and MVAPICH, support passing GPU pointers to MPI calls to avoid explict data movement between the host and the device. On the Python side, GPU arrays have been implemented by many libraries that need GPU computation, such as CuPy, Numba, PyTorch, and PyArrow. In order to increase library interoperability, two kinds of zero\-copy data exchange protocols are defined and agreed upon: \fI\%DLPack\fP and \fI\%CUDA Array Interface\fP\&. For example, a CuPy array can be passed to a Numba CUDA\-jit kernel. .sp \fIMPI for Python\fP provides an experimental support for GPU\-aware MPI. This feature requires: .INDENT 0.0 .IP 1. 3 mpi4py is built against a GPU\-aware MPI library. .IP 2. 3 The Python GPU arrays are compliant with either of the protocols. .UNINDENT .sp See the tutorial section for further information. We note that .INDENT 0.0 .IP \(bu 2 Whether or not a MPI call can work for GPU arrays depends on the underlying MPI implementation, not on mpi4py. .IP \(bu 2 This support is currently experimental and subject to change in the future. .UNINDENT .SS Dynamic Process Management .sp 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. .sp 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 \fIstart\fP the other. .sp In \fIMPI for Python\fP, new independent process groups can be created by calling the \fBIntracomm.Spawn\fP method within an intracommunicator. This call returns a new intercommunicator (i.e., an \fBIntercomm\fP instance) at the parent process group. The child process group can retrieve the matching intercommunicator by calling the \fBComm.Get_parent\fP 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. .sp Alternatively, disjoint groups of processes can establish communication using a client/server approach. Any server application must first call the \fBOpen_port\fP function to open a \fIport\fP and the \fBPublish_name\fP function to publish a provided \fIservice\fP, and next call the \fBIntracomm.Accept\fP method. Any client applications can first find a published \fIservice\fP by calling the \fBLookup_name\fP function, which returns the \fIport\fP where a server can be contacted; and next call the \fBIntracomm.Connect\fP method. Both \fBIntracomm.Accept\fP and \fBIntracomm.Connect\fP methods return an \fBIntercomm\fP instance. When connection between client/server processes is no longer needed, all of them must cooperatively call the \fBComm.Disconnect\fP method. Additionally, server applications should release resources by calling the \fBUnpublish_name\fP and \fBClose_port\fP functions. .SS One\-Sided Communications .sp One\-sided communications (also called \fIRemote Memory Access\fP, \fIRMA\fP) 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. .sp The MPI specification revolves around the use of objects called \fIwindows\fP; 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. .sp In \fIMPI for Python\fP, one\-sided operations are available by using instances of the \fBWin\fP class. New window objects are created by calling the \fBWin.Create\fP method at all processes within a communicator and specifying a memory buffer . When a window instance is no longer needed, the \fBWin.Free\fP method should be called. .sp The three one\-sided MPI operations for remote write, read and reduction are available through calling the methods \fBWin.Put\fP, \fBWin.Get\fP, and \fBWin.Accumulate\fP respectively within a \fBWin\fP 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. .sp 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 \fBWin.Start\fP and \fBWin.Complete\fP methods at the origin process, and target process cooperates by calling the \fBWin.Post\fP and \fBWin.Wait\fP methods. There is also a collective variant provided by the \fBWin.Fence\fP method. Passive target synchronization is more lenient, only the origin process calls the \fBWin.Lock\fP and \fBWin.Unlock\fP 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. .SS Parallel Input/Output .sp 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. .sp The MPI specification for parallel input/output revolves around the use objects called \fIfiles\fP\&. As defined by MPI, files are not just contiguous byte streams. Instead, they are regarded as ordered collections of \fItyped\fP 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. .sp 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). .sp In \fIMPI for Python\fP, all MPI input/output operations are performed through instances of the \fBFile\fP class. File handles are obtained by calling the \fBFile.Open\fP 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 \fBFile.Close\fP method. Files even can be deleted by calling method \fBFile.Delete\fP\&. .sp After creation, files are typically associated with a per\-process \fIview\fP\&. 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 \fBFile.Set_view\fP and \fBFile.Get_view\fP methods respectively. .sp Actual input/output operations are achieved by many methods combining read and write calls with different behavior regarding positioning, coordination, and synchronism. Summing up, \fIMPI for Python\fP 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. .SS Environmental Management .SS Initialization and Exit .sp Module functions \fBInit\fP or \fBInit_thread\fP and \fBFinalize\fP provide MPI initialization and finalization respectively. Module functions \fBIs_initialized\fP and \fBIs_finalized\fP provide the respective tests for initialization and finalization. .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 \fBMPI_Init()\fP or \fBMPI_Init_thread()\fP is actually called when you import the \fBMPI\fP module from the \fBmpi4py\fP package, but only if MPI is not already initialized. In such case, calling \fBInit\fP or \fBInit_thread\fP from Python is expected to generate an MPI error, and in turn an exception will be raised. .UNINDENT .UNINDENT .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 \fBMPI_Finalize()\fP is registered (by using Python C/API function \fBPy_AtExit()\fP) for being automatically called when Python processes exit, but only if \fBmpi4py\fP actually initialized MPI. Therefore, there is no need to call \fBFinalize\fP from Python to ensure MPI finalization. .UNINDENT .UNINDENT .SS Implementation Information .INDENT 0.0 .IP \(bu 2 The MPI version number can be retrieved from module function \fBGet_version\fP\&. It returns a two\-integer tuple \fB(version, subversion)\fP\&. .IP \(bu 2 The \fBGet_processor_name\fP function can be used to access the processor name. .IP \(bu 2 The values of predefined attributes attached to the world communicator can be obtained by calling the \fBComm.Get_attr\fP method within the \fBCOMM_WORLD\fP instance. .UNINDENT .SS Timers .sp MPI timer functionalities are available through the \fBWtime\fP and \fBWtick\fP functions. .SS Error Handling .sp In order facilitate handle sharing with other Python modules interfacing MPI\-based parallel libraries, the predefined MPI error handlers \fBERRORS_RETURN\fP and \fBERRORS_ARE_FATAL\fP can be assigned to and retrieved from communicators using methods \fBComm.Set_errhandler\fP and \fBComm.Get_errhandler\fP, and similarly for windows and files. .sp When the predefined error handler \fBERRORS_RETURN\fP is set, errors returned from MPI calls within Python code will raise an instance of the exception class \fBException\fP, which is a subclass of the standard Python exception \fBpython:RuntimeError\fP\&. .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 After import, mpi4py overrides the default MPI rules governing inheritance of error handlers. The \fBERRORS_RETURN\fP error handler is set in the predefined \fBCOMM_SELF\fP and \fBCOMM_WORLD\fP communicators, as well as any new \fBComm\fP, \fBWin\fP, or \fBFile\fP instance created through mpi4py. If you ever pass such handles to C/C++/Fortran library code, it is recommended to set the \fBERRORS_ARE_FATAL\fP error handler on them to ensure MPI errors do not pass silently. .UNINDENT .UNINDENT .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 Importing with \fBfrom mpi4py.MPI import *\fP will cause a name clashing with the standard Python \fBpython:Exception\fP base class. .UNINDENT .UNINDENT .SH TUTORIAL .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 Under construction. Contributions very welcome! .UNINDENT .UNINDENT .sp \fIMPI for Python\fP supports convenient, \fIpickle\fP\-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). .INDENT 0.0 .IP \(bu 2 Communication of generic Python objects .sp You have to use methods with \fBall\-lowercase\fP names, like \fBComm.send\fP, \fBComm.recv\fP, \fBComm.bcast\fP, \fBComm.scatter\fP, \fBComm.gather\fP . An object to be sent is passed as a parameter to the communication call, and the received object is simply the return value. .sp The \fBComm.isend\fP and \fBComm.irecv\fP methods return \fBRequest\fP instances; completion of these methods can be managed using the \fBRequest.test\fP and \fBRequest.wait\fP methods. .sp The \fBComm.recv\fP and \fBComm.irecv\fP 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 \fBComm.recv\fP or \fBComm.irecv\fP must be at least as long as the \fIpickled\fP data transmitted to the receiver. .sp Collective calls like \fBComm.scatter\fP, \fBComm.gather\fP, \fBComm.allgather\fP, \fBComm.alltoall\fP expect a single value or a sequence of \fBComm.size\fP elements at the root or all process. They return a single value, a list of \fBComm.size\fP elements, or \fBNone\fP\&. .sp \fBNOTE:\fP .INDENT 2.0 .INDENT 3.5 \fIMPI for Python\fP uses the \fBhighest\fP protocol version available in the Python runtime (see the \fBHIGHEST_PROTOCOL\fP constant in the \fBpickle\fP module). The default protocol can be changed at import time by setting the \fBMPI4PY_PICKLE_PROTOCOL\fP environment variable, or at runtime by assigning a different value to the \fBPROTOCOL\fP attribute of the \fBpickle\fP object within the \fBMPI\fP module. .UNINDENT .UNINDENT .IP \(bu 2 Communication of buffer\-like objects .sp You have to use method names starting with an \fBupper\-case\fP letter, like \fBComm.Send\fP, \fBComm.Recv\fP, \fBComm.Bcast\fP, \fBComm.Scatter\fP, \fBComm.Gather\fP\&. .sp In general, buffer arguments to these calls must be explicitly specified by using a 2/3\-list/tuple like \fB[data, MPI.DOUBLE]\fP, or \fB[data, count, MPI.DOUBLE]\fP (the former one uses the byte\-size of \fBdata\fP and the extent of the MPI datatype to define \fBcount\fP). .sp For vector collectives communication operations like \fBComm.Scatterv\fP and \fBComm.Gatherv\fP, buffer arguments are specified as \fB[data, count, displ, datatype]\fP, where \fBcount\fP and \fBdispl\fP are sequences of integral values. .sp 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. .sp 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 \fB__dlpack__\fP and \fB__dlpack_device__\fP methods or the \fB__cuda_array_interface__\fP 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. .UNINDENT .SS Running Python scripts with MPI .sp Most MPI programs can be run with the command \fBmpiexec\fP\&. In practice, running Python programs looks like: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 4 python script.py .ft P .fi .UNINDENT .UNINDENT .sp to run the program with 4 processors. .SS Point\-to\-Point Communication .INDENT 0.0 .IP \(bu 2 Python objects (\fBpickle\fP under the hood): .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C from mpi4py import MPI comm = MPI.COMM_WORLD rank = comm.Get_rank() if rank == 0: data = {\(aqa\(aq: 7, \(aqb\(aq: 3.14} comm.send(data, dest=1, tag=11) elif rank == 1: data = comm.recv(source=0, tag=11) .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Python objects with non\-blocking communication: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C from mpi4py import MPI comm = MPI.COMM_WORLD rank = comm.Get_rank() if rank == 0: data = {\(aqa\(aq: 7, \(aqb\(aq: 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() .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 NumPy arrays (the fast way!): .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi\(aq) comm.Send([data, MPI.INT], dest=1, tag=77) elif rank == 1: data = numpy.empty(1000, dtype=\(aqi\(aq) 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) .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS Collective Communication .INDENT 0.0 .IP \(bu 2 Broadcasting a Python dictionary: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C from mpi4py import MPI comm = MPI.COMM_WORLD rank = comm.Get_rank() if rank == 0: data = {\(aqkey1\(aq : [7, 2.72, 2+3j], \(aqkey2\(aq : ( \(aqabc\(aq, \(aqxyz\(aq)} else: data = None data = comm.bcast(data, root=0) .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Scattering Python objects: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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 .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Gathering Python objects: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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 .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Broadcasting a NumPy array: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C from mpi4py import MPI import numpy as np comm = MPI.COMM_WORLD rank = comm.Get_rank() if rank == 0: data = np.arange(100, dtype=\(aqi\(aq) else: data = np.empty(100, dtype=\(aqi\(aq) comm.Bcast(data, root=0) for i in range(100): assert data[i] == i .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Scattering NumPy arrays: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi\(aq) sendbuf.T[:,:] = range(size) recvbuf = np.empty(100, dtype=\(aqi\(aq) comm.Scatter(sendbuf, recvbuf, root=0) assert np.allclose(recvbuf, rank) .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Gathering NumPy arrays: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi\(aq) + rank recvbuf = None if rank == 0: recvbuf = np.empty([size, 100], dtype=\(aqi\(aq) comm.Gather(sendbuf, recvbuf, root=0) if rank == 0: for i in range(size): assert np.allclose(recvbuf[i,:], i) .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Parallel matrix\-vector product: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqd\(aq) comm.Allgather([x, MPI.DOUBLE], [xg, MPI.DOUBLE]) y = numpy.dot(A, xg) return y .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS MPI\-IO .INDENT 0.0 .IP \(bu 2 Collective I/O with NumPy arrays: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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() .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Non\-contiguous Collective I/O with NumPy arrays and datatypes: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi\(aq) 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() .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS Dynamic Process Management .INDENT 0.0 .IP \(bu 2 Compute Pi \- Master (or parent, or client) side: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C #!/usr/bin/env python from mpi4py import MPI import numpy import sys comm = MPI.COMM_SELF.Spawn(sys.executable, args=[\(aqcpi.py\(aq], maxprocs=5) N = numpy.array(100, \(aqi\(aq) comm.Bcast([N, MPI.INT], root=MPI.ROOT) PI = numpy.array(0.0, \(aqd\(aq) comm.Reduce(None, [PI, MPI.DOUBLE], op=MPI.SUM, root=MPI.ROOT) print(PI) comm.Disconnect() .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Compute Pi \- Worker (or child, or server) side: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C #!/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=\(aqi\(aq) 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=\(aqd\(aq) comm.Reduce([PI, MPI.DOUBLE], None, op=MPI.SUM, root=0) comm.Disconnect() .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS CUDA\-aware MPI + Python GPU arrays .INDENT 0.0 .IP \(bu 2 Reduce\-to\-all CuPy arrays: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi\(aq) recvbuf = cp.empty_like(sendbuf) assert hasattr(sendbuf, \(aq__cuda_array_interface__\(aq) assert hasattr(recvbuf, \(aq__cuda_array_interface__\(aq) cp.cuda.get_current_stream().synchronize() comm.Allreduce(sendbuf, recvbuf) assert cp.allclose(recvbuf, sendbuf*size) .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS One\-Sided Communications .INDENT 0.0 .IP \(bu 2 Read from (write to) the entire RMA window: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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) .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Accessing a part of the RMA window using the \fBtarget\fP argument, which is defined as \fB(offset, count, datatype)\fP: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C 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]) .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS Wrapping with SWIG .INDENT 0.0 .IP \(bu 2 C source: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C /* 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.\en", rank, size); } .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 SWIG interface file: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C // file: helloworld.i %module helloworld %{ #include #include "helloworld.c" }% %include mpi4py/mpi4py.i %mpi4py_typemap(Comm, MPI_Comm); void sayhello(MPI_Comm comm); .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Try it in the Python prompt: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C >>> from mpi4py import MPI >>> import helloworld >>> helloworld.sayhello(MPI.COMM_WORLD) Hello, World! I am process 0 of 1. .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SS Wrapping with F2Py .INDENT 0.0 .IP \(bu 2 Fortran 90 source: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C ! 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 *, \(aqHello, World! I am process \(aq,rank,\(aq of \(aq,size,\(aq.\(aq end subroutine sayhello .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Compiling example using f2py .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C $ f2py \-c \-\-f90exec=mpif90 helloworld.f90 \-m helloworld .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Try it in the Python prompt: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C >>> from mpi4py import MPI >>> import helloworld >>> fcomm = MPI.COMM_WORLD.py2f() >>> helloworld.sayhello(fcomm) Hello, World! I am process 0 of 1. .ft P .fi .UNINDENT .UNINDENT .UNINDENT .SH MPI4PY .sp This is the \fBMPI for Python\fP package. .sp The \fIMessage Passing Interface\fP (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. .sp \fIMPI for Python\fP 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. .SS Runtime configuration options .INDENT 0.0 .TP .B mpi4py.rc This object has attributes exposing runtime configuration options that become effective at import time of the \fBMPI\fP module. .UNINDENT .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%initialize\fP T} T{ Automatic MPI initialization at import T} _ T{ \fI\%threads\fP T} T{ Request initialization with thread support T} _ T{ \fI\%thread_level\fP T} T{ Level of thread support to request T} _ T{ \fI\%finalize\fP T} T{ Automatic MPI finalization at exit T} _ T{ \fI\%fast_reduce\fP T} T{ Use tree\-based reductions for objects T} _ T{ \fI\%recv_mprobe\fP T} T{ Use matched probes to receive objects T} _ T{ \fI\%errors\fP T} T{ Error handling policy T} _ .TE .sp Attributes Documentation .INDENT 0.0 .TP .B mpi4py.rc.initialize Automatic MPI initialization at import. .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_INITIALIZE\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.rc.threads Request initialization with thread support. .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_THREADS\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.rc.thread_level Level of thread support to request. .INDENT 7.0 .TP .B Type \fBstr\fP .TP .B Default \fB"multiple"\fP .TP .B Choices \fB"multiple"\fP, \fB"serialized"\fP, \fB"funneled"\fP, \fB"single"\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_THREAD_LEVEL\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.rc.finalize Automatic MPI finalization at exit. .INDENT 7.0 .TP .B Type \fBNone\fP or \fBbool\fP .TP .B Default \fBNone\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_FINALIZE\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.rc.fast_reduce Use tree\-based reductions for objects. .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_FAST_REDUCE\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.rc.recv_mprobe Use matched probes to receive objects. .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_RECV_MPROBE\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.rc.errors Error handling policy. .INDENT 7.0 .TP .B Type \fBstr\fP .TP .B Default \fB"exception"\fP .TP .B Choices \fB"exception"\fP, \fB"default"\fP, \fB"fatal"\fP .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%MPI4PY_RC_ERRORS\fP .UNINDENT .UNINDENT .UNINDENT .sp Example .sp MPI for Python features automatic initialization and finalization of the MPI execution environment. By using the \fI\%mpi4py.rc\fP object, MPI initialization and finalization can be handled programatically: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C 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 \(aqMPI\(aq module MPI.Init() # manual initialization of the MPI environment \&... # your finest code here ... MPI.Finalize() # manual finalization of the MPI environment .ft P .fi .UNINDENT .UNINDENT .SS Environment variables .sp The following environment variables override the corresponding attributes of the \fI\%mpi4py.rc\fP and \fBMPI.pickle\fP objects at import time of the \fBMPI\fP module. .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 For variables of boolean type, accepted values are \fB0\fP and \fB1\fP (interpreted as \fBFalse\fP and \fBTrue\fP, respectively), and strings specifying a \fI\%YAML boolean\fP value (case\-insensitive). .UNINDENT .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_INITIALIZE .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp Whether to automatically initialize MPI at import time of the \fBmpi4py.MPI\fP module. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.initialize\fP .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_FINALIZE .INDENT 7.0 .TP .B Type \fBNone\fP | \fBbool\fP .TP .B Default \fBNone\fP .TP .B Choices \fBNone\fP, \fBTrue\fP, \fBFalse\fP .UNINDENT .sp Whether to automatically finalize MPI at exit time of the Python process. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.finalize\fP .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_THREADS .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp Whether to initialize MPI with thread support. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.threads\fP .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_THREAD_LEVEL .INDENT 7.0 .TP .B Default \fB"multiple"\fP .TP .B Choices \fB"single"\fP, \fB"funneled"\fP, \fB"serialized"\fP, \fB"multiple"\fP .UNINDENT .sp The level of required thread support. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.thread_level\fP .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_FAST_REDUCE .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp Whether to use tree\-based reductions for objects. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.fast_reduce\fP .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_RECV_MPROBE .INDENT 7.0 .TP .B Type \fBbool\fP .TP .B Default \fBTrue\fP .UNINDENT .sp Whether to use matched probes to receive objects. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.recv_mprobe\fP .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B MPI4PY_RC_ERRORS .INDENT 7.0 .TP .B Default \fB"exception"\fP .TP .B Choices \fB"exception"\fP, \fB"default"\fP, \fB"fatal"\fP .UNINDENT .sp Controls default MPI error handling policy. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fI\%mpi4py.rc.errors\fP .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_PICKLE_PROTOCOL .INDENT 7.0 .TP .B Type \fBint\fP .TP .B Default \fBpickle.HIGHEST_PROTOCOL\fP .UNINDENT .sp Controls the default pickle protocol to use when communicating Python objects. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 \fBPROTOCOL\fP attribute of the \fBMPI.pickle\fP object within the \fBMPI\fP module. .UNINDENT .UNINDENT .sp New in version 3.1.0. .UNINDENT .INDENT 0.0 .TP .B MPI4PY_PICKLE_THRESHOLD .INDENT 7.0 .TP .B Type \fBint\fP .TP .B Default \fB262144\fP .UNINDENT .sp 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. .sp \fBSEE ALSO:\fP .INDENT 7.0 .INDENT 3.5 Module \fBmpi4py.util.pkl5\fP\&. .UNINDENT .UNINDENT .sp New in version 3.1.2. .UNINDENT .SS Miscellaneous functions .INDENT 0.0 .TP .B mpi4py.profile(name, *, path=None, logfile=None) Support for the MPI profiling interface. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBname\fP (\fBstr\fP) – Name of the profiler library to load. .IP \(bu 2 \fBpath\fP (\fBsequence\fP of \fBstr\fP, \fIoptional\fP) – Additional paths to search for the profiler. .IP \(bu 2 \fBlogfile\fP (\fBstr\fP, \fIoptional\fP) – Filename prefix for dumping profiler output. .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.get_config() Return a dictionary with information about MPI. .INDENT 7.0 .TP .B Return type Dict[str, str] .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.get_include() Return the directory in the package that contains header files. .sp 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): .INDENT 7.0 .INDENT 3.5 .sp .nf .ft C import mpi4py Extension(\(aqextension_name\(aq, ... include_dirs=[..., mpi4py.get_include()]) .ft P .fi .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .SH MPI4PY.MPI .SS Classes .sp Ancillary .TS center; |l|l|. _ T{ \fBDatatype\fP([datatype]) T} T{ Datatype object T} _ T{ \fBStatus\fP([status]) T} T{ Status object T} _ T{ \fBRequest\fP([request]) T} T{ Request handle T} _ T{ \fBPrequest\fP([request]) T} T{ Persistent request handle T} _ T{ \fBGrequest\fP([request]) T} T{ Generalized request handle T} _ T{ \fBOp\fP([op]) T} T{ Operation object T} _ T{ \fBGroup\fP([group]) T} T{ Group of processes T} _ T{ \fBInfo\fP([info]) T} T{ Info object T} _ .TE .sp Communication .TS center; |l|l|. _ T{ \fBComm\fP([comm]) T} T{ Communicator T} _ T{ \fBIntracomm\fP([comm]) T} T{ Intracommunicator T} _ T{ \fBTopocomm\fP([comm]) T} T{ Topology intracommunicator T} _ T{ \fBCartcomm\fP([comm]) T} T{ Cartesian topology intracommunicator T} _ T{ \fBGraphcomm\fP([comm]) T} T{ General graph topology intracommunicator T} _ T{ \fBDistgraphcomm\fP([comm]) T} T{ Distributed graph topology intracommunicator T} _ T{ \fBIntercomm\fP([comm]) T} T{ Intercommunicator T} _ T{ \fBMessage\fP([message]) T} T{ Matched message handle T} _ .TE .sp One\-sided operations .TS center; |l|l|. _ T{ \fBWin\fP([win]) T} T{ Window handle T} _ .TE .sp Input/Output .TS center; |l|l|. _ T{ \fBFile\fP([file]) T} T{ File handle T} _ .TE .sp Error handling .TS center; |l|l|. _ T{ \fBErrhandler\fP([errhandler]) T} T{ Error handler T} _ T{ \fBException\fP([ierr]) T} T{ Exception class T} _ .TE .sp Auxiliary .TS center; |l|l|. _ T{ \fBPickle\fP([dumps, loads, protocol]) T} T{ Pickle/unpickle Python objects T} _ T{ \fBmemory\fP(buf) T} T{ Memory buffer T} _ .TE .SS Functions .sp Version inquiry .TS center; |l|l|. _ T{ \fBGet_version\fP() T} T{ Obtain the version number of the MPI standard supported by the implementation as a tuple \fB(version, subversion)\fP T} _ T{ \fBGet_library_version\fP() T} T{ Obtain the version string of the MPI library T} _ .TE .sp Initialization and finalization .TS center; |l|l|. _ T{ \fBInit\fP() T} T{ Initialize the MPI execution environment T} _ T{ \fBInit_thread\fP([required]) T} T{ Initialize the MPI execution environment T} _ T{ \fBFinalize\fP() T} T{ Terminate the MPI execution environment T} _ T{ \fBIs_initialized\fP() T} T{ Indicates whether \fBInit\fP has been called T} _ T{ \fBIs_finalized\fP() T} T{ Indicates whether \fBFinalize\fP has completed T} _ T{ \fBQuery_thread\fP() T} T{ Return the level of thread support provided by the MPI library T} _ T{ \fBIs_thread_main\fP() T} T{ Indicate whether this thread called \fBInit\fP or \fBInit_thread\fP T} _ .TE .sp Memory allocation .TS center; |l|l|. _ T{ \fBAlloc_mem\fP(size[, info]) T} T{ Allocate memory for message passing and RMA T} _ T{ \fBFree_mem\fP(mem) T} T{ Free memory allocated with \fBAlloc_mem()\fP T} _ .TE .sp Address manipulation .TS center; |l|l|. _ T{ \fBGet_address\fP(location) T} T{ Get the address of a location in memory T} _ T{ \fBAint_add\fP(base, disp) T} T{ Return the sum of base address and displacement T} _ T{ \fBAint_diff\fP(addr1, addr2) T} T{ Return the difference between absolute addresses T} _ .TE .sp Timer .TS center; |l|l|. _ T{ \fBWtick\fP() T} T{ Return the resolution of \fBWtime\fP T} _ T{ \fBWtime\fP() T} T{ Return an elapsed time on the calling processor T} _ .TE .sp Error handling .TS center; |l|l|. _ T{ \fBGet_error_class\fP(errorcode) T} T{ Convert an \fIerror code\fP into an \fIerror class\fP T} _ T{ \fBGet_error_string\fP(errorcode) T} T{ Return the \fIerror string\fP for a given \fIerror class\fP or \fIerror code\fP T} _ T{ \fBAdd_error_class\fP() T} T{ Add an \fIerror class\fP to the known error classes T} _ T{ \fBAdd_error_code\fP(errorclass) T} T{ Add an \fIerror code\fP to an \fIerror class\fP T} _ T{ \fBAdd_error_string\fP(errorcode, string) T} T{ Associate an \fIerror string\fP with an \fIerror class\fP or \fIerrorcode\fP T} _ .TE .sp Dynamic process management .TS center; |l|l|. _ T{ \fBOpen_port\fP([info]) T} T{ Return an address that can be used to establish connections between groups of MPI processes T} _ T{ \fBClose_port\fP(port_name) T} T{ Close a port T} _ T{ \fBPublish_name\fP(service_name, port_name[, info]) T} T{ Publish a service name T} _ T{ \fBUnpublish_name\fP(service_name, port_name[, info]) T} T{ Unpublish a service name T} _ T{ \fBLookup_name\fP(service_name[, info]) T} T{ Lookup a port name given a service name T} _ .TE .sp Miscellanea .TS center; |l|l|. _ T{ \fBAttach_buffer\fP(buf) T} T{ Attach a user\-provided buffer for sending in buffered mode T} _ T{ \fBDetach_buffer\fP() T} T{ Remove an existing attached buffer T} _ T{ \fBCompute_dims\fP(nnodes, dims) T} T{ Return a balanced distribution of processes per coordinate direction T} _ T{ \fBGet_processor_name\fP() T} T{ Obtain the name of the calling processor T} _ T{ \fBRegister_datarep\fP(datarep, read_fn, write_fn, ...) T} T{ Register user\-defined data representations T} _ T{ \fBPcontrol\fP(level) T} T{ Control profiling T} _ .TE .sp Utilities .TS center; |l|l|. _ T{ \fBget_vendor\fP() T} T{ Infomation about the underlying MPI implementation T} _ .TE .SS Attributes .TS center; |l|l|. _ T{ \fBUNDEFINED\fP T} T{ \fBint\fP \fBUNDEFINED\fP T} _ T{ \fBANY_SOURCE\fP T} T{ \fBint\fP \fBANY_SOURCE\fP T} _ T{ \fBANY_TAG\fP T} T{ \fBint\fP \fBANY_TAG\fP T} _ T{ \fBPROC_NULL\fP T} T{ \fBint\fP \fBPROC_NULL\fP T} _ T{ \fBROOT\fP T} T{ \fBint\fP \fBROOT\fP T} _ T{ \fBBOTTOM\fP T} T{ \fBBottom\fP \fBBOTTOM\fP T} _ T{ \fBIN_PLACE\fP T} T{ \fBInPlace\fP \fBIN_PLACE\fP T} _ T{ \fBKEYVAL_INVALID\fP T} T{ \fBint\fP \fBKEYVAL_INVALID\fP T} _ T{ \fBTAG_UB\fP T} T{ \fBint\fP \fBTAG_UB\fP T} _ T{ \fBHOST\fP T} T{ \fBint\fP \fBHOST\fP T} _ T{ \fBIO\fP T} T{ \fBint\fP \fBIO\fP T} _ T{ \fBWTIME_IS_GLOBAL\fP T} T{ \fBint\fP \fBWTIME_IS_GLOBAL\fP T} _ T{ \fBUNIVERSE_SIZE\fP T} T{ \fBint\fP \fBUNIVERSE_SIZE\fP T} _ T{ \fBAPPNUM\fP T} T{ \fBint\fP \fBAPPNUM\fP T} _ T{ \fBLASTUSEDCODE\fP T} T{ \fBint\fP \fBLASTUSEDCODE\fP T} _ T{ \fBWIN_BASE\fP T} T{ \fBint\fP \fBWIN_BASE\fP T} _ T{ \fBWIN_SIZE\fP T} T{ \fBint\fP \fBWIN_SIZE\fP T} _ T{ \fBWIN_DISP_UNIT\fP T} T{ \fBint\fP \fBWIN_DISP_UNIT\fP T} _ T{ \fBWIN_CREATE_FLAVOR\fP T} T{ \fBint\fP \fBWIN_CREATE_FLAVOR\fP T} _ T{ \fBWIN_FLAVOR\fP T} T{ \fBint\fP \fBWIN_FLAVOR\fP T} _ T{ \fBWIN_MODEL\fP T} T{ \fBint\fP \fBWIN_MODEL\fP T} _ T{ \fBSUCCESS\fP T} T{ \fBint\fP \fBSUCCESS\fP T} _ T{ \fBERR_LASTCODE\fP T} T{ \fBint\fP \fBERR_LASTCODE\fP T} _ T{ \fBERR_COMM\fP T} T{ \fBint\fP \fBERR_COMM\fP T} _ T{ \fBERR_GROUP\fP T} T{ \fBint\fP \fBERR_GROUP\fP T} _ T{ \fBERR_TYPE\fP T} T{ \fBint\fP \fBERR_TYPE\fP T} _ T{ \fBERR_REQUEST\fP T} T{ \fBint\fP \fBERR_REQUEST\fP T} _ T{ \fBERR_OP\fP T} T{ \fBint\fP \fBERR_OP\fP T} _ T{ \fBERR_BUFFER\fP T} T{ \fBint\fP \fBERR_BUFFER\fP T} _ T{ \fBERR_COUNT\fP T} T{ \fBint\fP \fBERR_COUNT\fP T} _ T{ \fBERR_TAG\fP T} T{ \fBint\fP \fBERR_TAG\fP T} _ T{ \fBERR_RANK\fP T} T{ \fBint\fP \fBERR_RANK\fP T} _ T{ \fBERR_ROOT\fP T} T{ \fBint\fP \fBERR_ROOT\fP T} _ T{ \fBERR_TRUNCATE\fP T} T{ \fBint\fP \fBERR_TRUNCATE\fP T} _ T{ \fBERR_IN_STATUS\fP T} T{ \fBint\fP \fBERR_IN_STATUS\fP T} _ T{ \fBERR_PENDING\fP T} T{ \fBint\fP \fBERR_PENDING\fP T} _ T{ \fBERR_TOPOLOGY\fP T} T{ \fBint\fP \fBERR_TOPOLOGY\fP T} _ T{ \fBERR_DIMS\fP T} T{ \fBint\fP \fBERR_DIMS\fP T} _ T{ \fBERR_ARG\fP T} T{ \fBint\fP \fBERR_ARG\fP T} _ T{ \fBERR_OTHER\fP T} T{ \fBint\fP \fBERR_OTHER\fP T} _ T{ \fBERR_UNKNOWN\fP T} T{ \fBint\fP \fBERR_UNKNOWN\fP T} _ T{ \fBERR_INTERN\fP T} T{ \fBint\fP \fBERR_INTERN\fP T} _ T{ \fBERR_INFO\fP T} T{ \fBint\fP \fBERR_INFO\fP T} _ T{ \fBERR_FILE\fP T} T{ \fBint\fP \fBERR_FILE\fP T} _ T{ \fBERR_WIN\fP T} T{ \fBint\fP \fBERR_WIN\fP T} _ T{ \fBERR_KEYVAL\fP T} T{ \fBint\fP \fBERR_KEYVAL\fP T} _ T{ \fBERR_INFO_KEY\fP T} T{ \fBint\fP \fBERR_INFO_KEY\fP T} _ T{ \fBERR_INFO_VALUE\fP T} T{ \fBint\fP \fBERR_INFO_VALUE\fP T} _ T{ \fBERR_INFO_NOKEY\fP T} T{ \fBint\fP \fBERR_INFO_NOKEY\fP T} _ T{ \fBERR_ACCESS\fP T} T{ \fBint\fP \fBERR_ACCESS\fP T} _ T{ \fBERR_AMODE\fP T} T{ \fBint\fP \fBERR_AMODE\fP T} _ T{ \fBERR_BAD_FILE\fP T} T{ \fBint\fP \fBERR_BAD_FILE\fP T} _ T{ \fBERR_FILE_EXISTS\fP T} T{ \fBint\fP \fBERR_FILE_EXISTS\fP T} _ T{ \fBERR_FILE_IN_USE\fP T} T{ \fBint\fP \fBERR_FILE_IN_USE\fP T} _ T{ \fBERR_NO_SPACE\fP T} T{ \fBint\fP \fBERR_NO_SPACE\fP T} _ T{ \fBERR_NO_SUCH_FILE\fP T} T{ \fBint\fP \fBERR_NO_SUCH_FILE\fP T} _ T{ \fBERR_IO\fP T} T{ \fBint\fP \fBERR_IO\fP T} _ T{ \fBERR_READ_ONLY\fP T} T{ \fBint\fP \fBERR_READ_ONLY\fP T} _ T{ \fBERR_CONVERSION\fP T} T{ \fBint\fP \fBERR_CONVERSION\fP T} _ T{ \fBERR_DUP_DATAREP\fP T} T{ \fBint\fP \fBERR_DUP_DATAREP\fP T} _ T{ \fBERR_UNSUPPORTED_DATAREP\fP T} T{ \fBint\fP \fBERR_UNSUPPORTED_DATAREP\fP T} _ T{ \fBERR_UNSUPPORTED_OPERATION\fP T} T{ \fBint\fP \fBERR_UNSUPPORTED_OPERATION\fP T} _ T{ \fBERR_NAME\fP T} T{ \fBint\fP \fBERR_NAME\fP T} _ T{ \fBERR_NO_MEM\fP T} T{ \fBint\fP \fBERR_NO_MEM\fP T} _ T{ \fBERR_NOT_SAME\fP T} T{ \fBint\fP \fBERR_NOT_SAME\fP T} _ T{ \fBERR_PORT\fP T} T{ \fBint\fP \fBERR_PORT\fP T} _ T{ \fBERR_QUOTA\fP T} T{ \fBint\fP \fBERR_QUOTA\fP T} _ T{ \fBERR_SERVICE\fP T} T{ \fBint\fP \fBERR_SERVICE\fP T} _ T{ \fBERR_SPAWN\fP T} T{ \fBint\fP \fBERR_SPAWN\fP T} _ T{ \fBERR_BASE\fP T} T{ \fBint\fP \fBERR_BASE\fP T} _ T{ \fBERR_SIZE\fP T} T{ \fBint\fP \fBERR_SIZE\fP T} _ T{ \fBERR_DISP\fP T} T{ \fBint\fP \fBERR_DISP\fP T} _ T{ \fBERR_ASSERT\fP T} T{ \fBint\fP \fBERR_ASSERT\fP T} _ T{ \fBERR_LOCKTYPE\fP T} T{ \fBint\fP \fBERR_LOCKTYPE\fP T} _ T{ \fBERR_RMA_CONFLICT\fP T} T{ \fBint\fP \fBERR_RMA_CONFLICT\fP T} _ T{ \fBERR_RMA_SYNC\fP T} T{ \fBint\fP \fBERR_RMA_SYNC\fP T} _ T{ \fBERR_RMA_RANGE\fP T} T{ \fBint\fP \fBERR_RMA_RANGE\fP T} _ T{ \fBERR_RMA_ATTACH\fP T} T{ \fBint\fP \fBERR_RMA_ATTACH\fP T} _ T{ \fBERR_RMA_SHARED\fP T} T{ \fBint\fP \fBERR_RMA_SHARED\fP T} _ T{ \fBERR_RMA_FLAVOR\fP T} T{ \fBint\fP \fBERR_RMA_FLAVOR\fP T} _ T{ \fBORDER_C\fP T} T{ \fBint\fP \fBORDER_C\fP T} _ T{ \fBORDER_F\fP T} T{ \fBint\fP \fBORDER_F\fP T} _ T{ \fBORDER_FORTRAN\fP T} T{ \fBint\fP \fBORDER_FORTRAN\fP T} _ T{ \fBTYPECLASS_INTEGER\fP T} T{ \fBint\fP \fBTYPECLASS_INTEGER\fP T} _ T{ \fBTYPECLASS_REAL\fP T} T{ \fBint\fP \fBTYPECLASS_REAL\fP T} _ T{ \fBTYPECLASS_COMPLEX\fP T} T{ \fBint\fP \fBTYPECLASS_COMPLEX\fP T} _ T{ \fBDISTRIBUTE_NONE\fP T} T{ \fBint\fP \fBDISTRIBUTE_NONE\fP T} _ T{ \fBDISTRIBUTE_BLOCK\fP T} T{ \fBint\fP \fBDISTRIBUTE_BLOCK\fP T} _ T{ \fBDISTRIBUTE_CYCLIC\fP T} T{ \fBint\fP \fBDISTRIBUTE_CYCLIC\fP T} _ T{ \fBDISTRIBUTE_DFLT_DARG\fP T} T{ \fBint\fP \fBDISTRIBUTE_DFLT_DARG\fP T} _ T{ \fBCOMBINER_NAMED\fP T} T{ \fBint\fP \fBCOMBINER_NAMED\fP T} _ T{ \fBCOMBINER_DUP\fP T} T{ \fBint\fP \fBCOMBINER_DUP\fP T} _ T{ \fBCOMBINER_CONTIGUOUS\fP T} T{ \fBint\fP \fBCOMBINER_CONTIGUOUS\fP T} _ T{ \fBCOMBINER_VECTOR\fP T} T{ \fBint\fP \fBCOMBINER_VECTOR\fP T} _ T{ \fBCOMBINER_HVECTOR\fP T} T{ \fBint\fP \fBCOMBINER_HVECTOR\fP T} _ T{ \fBCOMBINER_INDEXED\fP T} T{ \fBint\fP \fBCOMBINER_INDEXED\fP T} _ T{ \fBCOMBINER_HINDEXED\fP T} T{ \fBint\fP \fBCOMBINER_HINDEXED\fP T} _ T{ \fBCOMBINER_INDEXED_BLOCK\fP T} T{ \fBint\fP \fBCOMBINER_INDEXED_BLOCK\fP T} _ T{ \fBCOMBINER_HINDEXED_BLOCK\fP T} T{ \fBint\fP \fBCOMBINER_HINDEXED_BLOCK\fP T} _ T{ \fBCOMBINER_STRUCT\fP T} T{ \fBint\fP \fBCOMBINER_STRUCT\fP T} _ T{ \fBCOMBINER_SUBARRAY\fP T} T{ \fBint\fP \fBCOMBINER_SUBARRAY\fP T} _ T{ \fBCOMBINER_DARRAY\fP T} T{ \fBint\fP \fBCOMBINER_DARRAY\fP T} _ T{ \fBCOMBINER_RESIZED\fP T} T{ \fBint\fP \fBCOMBINER_RESIZED\fP T} _ T{ \fBCOMBINER_F90_REAL\fP T} T{ \fBint\fP \fBCOMBINER_F90_REAL\fP T} _ T{ \fBCOMBINER_F90_COMPLEX\fP T} T{ \fBint\fP \fBCOMBINER_F90_COMPLEX\fP T} _ T{ \fBCOMBINER_F90_INTEGER\fP T} T{ \fBint\fP \fBCOMBINER_F90_INTEGER\fP T} _ T{ \fBIDENT\fP T} T{ \fBint\fP \fBIDENT\fP T} _ T{ \fBCONGRUENT\fP T} T{ \fBint\fP \fBCONGRUENT\fP T} _ T{ \fBSIMILAR\fP T} T{ \fBint\fP \fBSIMILAR\fP T} _ T{ \fBUNEQUAL\fP T} T{ \fBint\fP \fBUNEQUAL\fP T} _ T{ \fBCART\fP T} T{ \fBint\fP \fBCART\fP T} _ T{ \fBGRAPH\fP T} T{ \fBint\fP \fBGRAPH\fP T} _ T{ \fBDIST_GRAPH\fP T} T{ \fBint\fP \fBDIST_GRAPH\fP T} _ T{ \fBUNWEIGHTED\fP T} T{ \fBint\fP \fBUNWEIGHTED\fP T} _ T{ \fBWEIGHTS_EMPTY\fP T} T{ \fBint\fP \fBWEIGHTS_EMPTY\fP T} _ T{ \fBCOMM_TYPE_SHARED\fP T} T{ \fBint\fP \fBCOMM_TYPE_SHARED\fP T} _ T{ \fBBSEND_OVERHEAD\fP T} T{ \fBint\fP \fBBSEND_OVERHEAD\fP T} _ T{ \fBWIN_FLAVOR_CREATE\fP T} T{ \fBint\fP \fBWIN_FLAVOR_CREATE\fP T} _ T{ \fBWIN_FLAVOR_ALLOCATE\fP T} T{ \fBint\fP \fBWIN_FLAVOR_ALLOCATE\fP T} _ T{ \fBWIN_FLAVOR_DYNAMIC\fP T} T{ \fBint\fP \fBWIN_FLAVOR_DYNAMIC\fP T} _ T{ \fBWIN_FLAVOR_SHARED\fP T} T{ \fBint\fP \fBWIN_FLAVOR_SHARED\fP T} _ T{ \fBWIN_SEPARATE\fP T} T{ \fBint\fP \fBWIN_SEPARATE\fP T} _ T{ \fBWIN_UNIFIED\fP T} T{ \fBint\fP \fBWIN_UNIFIED\fP T} _ T{ \fBMODE_NOCHECK\fP T} T{ \fBint\fP \fBMODE_NOCHECK\fP T} _ T{ \fBMODE_NOSTORE\fP T} T{ \fBint\fP \fBMODE_NOSTORE\fP T} _ T{ \fBMODE_NOPUT\fP T} T{ \fBint\fP \fBMODE_NOPUT\fP T} _ T{ \fBMODE_NOPRECEDE\fP T} T{ \fBint\fP \fBMODE_NOPRECEDE\fP T} _ T{ \fBMODE_NOSUCCEED\fP T} T{ \fBint\fP \fBMODE_NOSUCCEED\fP T} _ T{ \fBLOCK_EXCLUSIVE\fP T} T{ \fBint\fP \fBLOCK_EXCLUSIVE\fP T} _ T{ \fBLOCK_SHARED\fP T} T{ \fBint\fP \fBLOCK_SHARED\fP T} _ T{ \fBMODE_RDONLY\fP T} T{ \fBint\fP \fBMODE_RDONLY\fP T} _ T{ \fBMODE_WRONLY\fP T} T{ \fBint\fP \fBMODE_WRONLY\fP T} _ T{ \fBMODE_RDWR\fP T} T{ \fBint\fP \fBMODE_RDWR\fP T} _ T{ \fBMODE_CREATE\fP T} T{ \fBint\fP \fBMODE_CREATE\fP T} _ T{ \fBMODE_EXCL\fP T} T{ \fBint\fP \fBMODE_EXCL\fP T} _ T{ \fBMODE_DELETE_ON_CLOSE\fP T} T{ \fBint\fP \fBMODE_DELETE_ON_CLOSE\fP T} _ T{ \fBMODE_UNIQUE_OPEN\fP T} T{ \fBint\fP \fBMODE_UNIQUE_OPEN\fP T} _ T{ \fBMODE_SEQUENTIAL\fP T} T{ \fBint\fP \fBMODE_SEQUENTIAL\fP T} _ T{ \fBMODE_APPEND\fP T} T{ \fBint\fP \fBMODE_APPEND\fP T} _ T{ \fBSEEK_SET\fP T} T{ \fBint\fP \fBSEEK_SET\fP T} _ T{ \fBSEEK_CUR\fP T} T{ \fBint\fP \fBSEEK_CUR\fP T} _ T{ \fBSEEK_END\fP T} T{ \fBint\fP \fBSEEK_END\fP T} _ T{ \fBDISPLACEMENT_CURRENT\fP T} T{ \fBint\fP \fBDISPLACEMENT_CURRENT\fP T} _ T{ \fBDISP_CUR\fP T} T{ \fBint\fP \fBDISP_CUR\fP T} _ T{ \fBTHREAD_SINGLE\fP T} T{ \fBint\fP \fBTHREAD_SINGLE\fP T} _ T{ \fBTHREAD_FUNNELED\fP T} T{ \fBint\fP \fBTHREAD_FUNNELED\fP T} _ T{ \fBTHREAD_SERIALIZED\fP T} T{ \fBint\fP \fBTHREAD_SERIALIZED\fP T} _ T{ \fBTHREAD_MULTIPLE\fP T} T{ \fBint\fP \fBTHREAD_MULTIPLE\fP T} _ T{ \fBVERSION\fP T} T{ \fBint\fP \fBVERSION\fP T} _ T{ \fBSUBVERSION\fP T} T{ \fBint\fP \fBSUBVERSION\fP T} _ T{ \fBMAX_PROCESSOR_NAME\fP T} T{ \fBint\fP \fBMAX_PROCESSOR_NAME\fP T} _ T{ \fBMAX_ERROR_STRING\fP T} T{ \fBint\fP \fBMAX_ERROR_STRING\fP T} _ T{ \fBMAX_PORT_NAME\fP T} T{ \fBint\fP \fBMAX_PORT_NAME\fP T} _ T{ \fBMAX_INFO_KEY\fP T} T{ \fBint\fP \fBMAX_INFO_KEY\fP T} _ T{ \fBMAX_INFO_VAL\fP T} T{ \fBint\fP \fBMAX_INFO_VAL\fP T} _ T{ \fBMAX_OBJECT_NAME\fP T} T{ \fBint\fP \fBMAX_OBJECT_NAME\fP T} _ T{ \fBMAX_DATAREP_STRING\fP T} T{ \fBint\fP \fBMAX_DATAREP_STRING\fP T} _ T{ \fBMAX_LIBRARY_VERSION_STRING\fP T} T{ \fBint\fP \fBMAX_LIBRARY_VERSION_STRING\fP T} _ T{ \fBDATATYPE_NULL\fP T} T{ \fBDatatype\fP \fBDATATYPE_NULL\fP T} _ T{ \fBUB\fP T} T{ \fBDatatype\fP \fBUB\fP T} _ T{ \fBLB\fP T} T{ \fBDatatype\fP \fBLB\fP T} _ T{ \fBPACKED\fP T} T{ \fBDatatype\fP \fBPACKED\fP T} _ T{ \fBBYTE\fP T} T{ \fBDatatype\fP \fBBYTE\fP T} _ T{ \fBAINT\fP T} T{ \fBDatatype\fP \fBAINT\fP T} _ T{ \fBOFFSET\fP T} T{ \fBDatatype\fP \fBOFFSET\fP T} _ T{ \fBCOUNT\fP T} T{ \fBDatatype\fP \fBCOUNT\fP T} _ T{ \fBCHAR\fP T} T{ \fBDatatype\fP \fBCHAR\fP T} _ T{ \fBWCHAR\fP T} T{ \fBDatatype\fP \fBWCHAR\fP T} _ T{ \fBSIGNED_CHAR\fP T} T{ \fBDatatype\fP \fBSIGNED_CHAR\fP T} _ T{ \fBSHORT\fP T} T{ \fBDatatype\fP \fBSHORT\fP T} _ T{ \fBINT\fP T} T{ \fBDatatype\fP \fBINT\fP T} _ T{ \fBLONG\fP T} T{ \fBDatatype\fP \fBLONG\fP T} _ T{ \fBLONG_LONG\fP T} T{ \fBDatatype\fP \fBLONG_LONG\fP T} _ T{ \fBUNSIGNED_CHAR\fP T} T{ \fBDatatype\fP \fBUNSIGNED_CHAR\fP T} _ T{ \fBUNSIGNED_SHORT\fP T} T{ \fBDatatype\fP \fBUNSIGNED_SHORT\fP T} _ T{ \fBUNSIGNED\fP T} T{ \fBDatatype\fP \fBUNSIGNED\fP T} _ T{ \fBUNSIGNED_LONG\fP T} T{ \fBDatatype\fP \fBUNSIGNED_LONG\fP T} _ T{ \fBUNSIGNED_LONG_LONG\fP T} T{ \fBDatatype\fP \fBUNSIGNED_LONG_LONG\fP T} _ T{ \fBFLOAT\fP T} T{ \fBDatatype\fP \fBFLOAT\fP T} _ T{ \fBDOUBLE\fP T} T{ \fBDatatype\fP \fBDOUBLE\fP T} _ T{ \fBLONG_DOUBLE\fP T} T{ \fBDatatype\fP \fBLONG_DOUBLE\fP T} _ T{ \fBC_BOOL\fP T} T{ \fBDatatype\fP \fBC_BOOL\fP T} _ T{ \fBINT8_T\fP T} T{ \fBDatatype\fP \fBINT8_T\fP T} _ T{ \fBINT16_T\fP T} T{ \fBDatatype\fP \fBINT16_T\fP T} _ T{ \fBINT32_T\fP T} T{ \fBDatatype\fP \fBINT32_T\fP T} _ T{ \fBINT64_T\fP T} T{ \fBDatatype\fP \fBINT64_T\fP T} _ T{ \fBUINT8_T\fP T} T{ \fBDatatype\fP \fBUINT8_T\fP T} _ T{ \fBUINT16_T\fP T} T{ \fBDatatype\fP \fBUINT16_T\fP T} _ T{ \fBUINT32_T\fP T} T{ \fBDatatype\fP \fBUINT32_T\fP T} _ T{ \fBUINT64_T\fP T} T{ \fBDatatype\fP \fBUINT64_T\fP T} _ T{ \fBC_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_COMPLEX\fP T} _ T{ \fBC_FLOAT_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_FLOAT_COMPLEX\fP T} _ T{ \fBC_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_DOUBLE_COMPLEX\fP T} _ T{ \fBC_LONG_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_LONG_DOUBLE_COMPLEX\fP T} _ T{ \fBCXX_BOOL\fP T} T{ \fBDatatype\fP \fBCXX_BOOL\fP T} _ T{ \fBCXX_FLOAT_COMPLEX\fP T} T{ \fBDatatype\fP \fBCXX_FLOAT_COMPLEX\fP T} _ T{ \fBCXX_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBCXX_DOUBLE_COMPLEX\fP T} _ T{ \fBCXX_LONG_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBCXX_LONG_DOUBLE_COMPLEX\fP T} _ T{ \fBSHORT_INT\fP T} T{ \fBDatatype\fP \fBSHORT_INT\fP T} _ T{ \fBINT_INT\fP T} T{ \fBDatatype\fP \fBINT_INT\fP T} _ T{ \fBTWOINT\fP T} T{ \fBDatatype\fP \fBTWOINT\fP T} _ T{ \fBLONG_INT\fP T} T{ \fBDatatype\fP \fBLONG_INT\fP T} _ T{ \fBFLOAT_INT\fP T} T{ \fBDatatype\fP \fBFLOAT_INT\fP T} _ T{ \fBDOUBLE_INT\fP T} T{ \fBDatatype\fP \fBDOUBLE_INT\fP T} _ T{ \fBLONG_DOUBLE_INT\fP T} T{ \fBDatatype\fP \fBLONG_DOUBLE_INT\fP T} _ T{ \fBCHARACTER\fP T} T{ \fBDatatype\fP \fBCHARACTER\fP T} _ T{ \fBLOGICAL\fP T} T{ \fBDatatype\fP \fBLOGICAL\fP T} _ T{ \fBINTEGER\fP T} T{ \fBDatatype\fP \fBINTEGER\fP T} _ T{ \fBREAL\fP T} T{ \fBDatatype\fP \fBREAL\fP T} _ T{ \fBDOUBLE_PRECISION\fP T} T{ \fBDatatype\fP \fBDOUBLE_PRECISION\fP T} _ T{ \fBCOMPLEX\fP T} T{ \fBDatatype\fP \fBCOMPLEX\fP T} _ T{ \fBDOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBDOUBLE_COMPLEX\fP T} _ T{ \fBLOGICAL1\fP T} T{ \fBDatatype\fP \fBLOGICAL1\fP T} _ T{ \fBLOGICAL2\fP T} T{ \fBDatatype\fP \fBLOGICAL2\fP T} _ T{ \fBLOGICAL4\fP T} T{ \fBDatatype\fP \fBLOGICAL4\fP T} _ T{ \fBLOGICAL8\fP T} T{ \fBDatatype\fP \fBLOGICAL8\fP T} _ T{ \fBINTEGER1\fP T} T{ \fBDatatype\fP \fBINTEGER1\fP T} _ T{ \fBINTEGER2\fP T} T{ \fBDatatype\fP \fBINTEGER2\fP T} _ T{ \fBINTEGER4\fP T} T{ \fBDatatype\fP \fBINTEGER4\fP T} _ T{ \fBINTEGER8\fP T} T{ \fBDatatype\fP \fBINTEGER8\fP T} _ T{ \fBINTEGER16\fP T} T{ \fBDatatype\fP \fBINTEGER16\fP T} _ T{ \fBREAL2\fP T} T{ \fBDatatype\fP \fBREAL2\fP T} _ T{ \fBREAL4\fP T} T{ \fBDatatype\fP \fBREAL4\fP T} _ T{ \fBREAL8\fP T} T{ \fBDatatype\fP \fBREAL8\fP T} _ T{ \fBREAL16\fP T} T{ \fBDatatype\fP \fBREAL16\fP T} _ T{ \fBCOMPLEX4\fP T} T{ \fBDatatype\fP \fBCOMPLEX4\fP T} _ T{ \fBCOMPLEX8\fP T} T{ \fBDatatype\fP \fBCOMPLEX8\fP T} _ T{ \fBCOMPLEX16\fP T} T{ \fBDatatype\fP \fBCOMPLEX16\fP T} _ T{ \fBCOMPLEX32\fP T} T{ \fBDatatype\fP \fBCOMPLEX32\fP T} _ T{ \fBUNSIGNED_INT\fP T} T{ \fBDatatype\fP \fBUNSIGNED_INT\fP T} _ T{ \fBSIGNED_SHORT\fP T} T{ \fBDatatype\fP \fBSIGNED_SHORT\fP T} _ T{ \fBSIGNED_INT\fP T} T{ \fBDatatype\fP \fBSIGNED_INT\fP T} _ T{ \fBSIGNED_LONG\fP T} T{ \fBDatatype\fP \fBSIGNED_LONG\fP T} _ T{ \fBSIGNED_LONG_LONG\fP T} T{ \fBDatatype\fP \fBSIGNED_LONG_LONG\fP T} _ T{ \fBBOOL\fP T} T{ \fBDatatype\fP \fBBOOL\fP T} _ T{ \fBSINT8_T\fP T} T{ \fBDatatype\fP \fBSINT8_T\fP T} _ T{ \fBSINT16_T\fP T} T{ \fBDatatype\fP \fBSINT16_T\fP T} _ T{ \fBSINT32_T\fP T} T{ \fBDatatype\fP \fBSINT32_T\fP T} _ T{ \fBSINT64_T\fP T} T{ \fBDatatype\fP \fBSINT64_T\fP T} _ T{ \fBF_BOOL\fP T} T{ \fBDatatype\fP \fBF_BOOL\fP T} _ T{ \fBF_INT\fP T} T{ \fBDatatype\fP \fBF_INT\fP T} _ T{ \fBF_FLOAT\fP T} T{ \fBDatatype\fP \fBF_FLOAT\fP T} _ T{ \fBF_DOUBLE\fP T} T{ \fBDatatype\fP \fBF_DOUBLE\fP T} _ T{ \fBF_COMPLEX\fP T} T{ \fBDatatype\fP \fBF_COMPLEX\fP T} _ T{ \fBF_FLOAT_COMPLEX\fP T} T{ \fBDatatype\fP \fBF_FLOAT_COMPLEX\fP T} _ T{ \fBF_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBF_DOUBLE_COMPLEX\fP T} _ T{ \fBREQUEST_NULL\fP T} T{ \fBRequest\fP \fBREQUEST_NULL\fP T} _ T{ \fBMESSAGE_NULL\fP T} T{ \fBMessage\fP \fBMESSAGE_NULL\fP T} _ T{ \fBMESSAGE_NO_PROC\fP T} T{ \fBMessage\fP \fBMESSAGE_NO_PROC\fP T} _ T{ \fBOP_NULL\fP T} T{ \fBOp\fP \fBOP_NULL\fP T} _ T{ \fBMAX\fP T} T{ \fBOp\fP \fBMAX\fP T} _ T{ \fBMIN\fP T} T{ \fBOp\fP \fBMIN\fP T} _ T{ \fBSUM\fP T} T{ \fBOp\fP \fBSUM\fP T} _ T{ \fBPROD\fP T} T{ \fBOp\fP \fBPROD\fP T} _ T{ \fBLAND\fP T} T{ \fBOp\fP \fBLAND\fP T} _ T{ \fBBAND\fP T} T{ \fBOp\fP \fBBAND\fP T} _ T{ \fBLOR\fP T} T{ \fBOp\fP \fBLOR\fP T} _ T{ \fBBOR\fP T} T{ \fBOp\fP \fBBOR\fP T} _ T{ \fBLXOR\fP T} T{ \fBOp\fP \fBLXOR\fP T} _ T{ \fBBXOR\fP T} T{ \fBOp\fP \fBBXOR\fP T} _ T{ \fBMAXLOC\fP T} T{ \fBOp\fP \fBMAXLOC\fP T} _ T{ \fBMINLOC\fP T} T{ \fBOp\fP \fBMINLOC\fP T} _ T{ \fBREPLACE\fP T} T{ \fBOp\fP \fBREPLACE\fP T} _ T{ \fBNO_OP\fP T} T{ \fBOp\fP \fBNO_OP\fP T} _ T{ \fBGROUP_NULL\fP T} T{ \fBGroup\fP \fBGROUP_NULL\fP T} _ T{ \fBGROUP_EMPTY\fP T} T{ \fBGroup\fP \fBGROUP_EMPTY\fP T} _ T{ \fBINFO_NULL\fP T} T{ \fBInfo\fP \fBINFO_NULL\fP T} _ T{ \fBINFO_ENV\fP T} T{ \fBInfo\fP \fBINFO_ENV\fP T} _ T{ \fBERRHANDLER_NULL\fP T} T{ \fBErrhandler\fP \fBERRHANDLER_NULL\fP T} _ T{ \fBERRORS_RETURN\fP T} T{ \fBErrhandler\fP \fBERRORS_RETURN\fP T} _ T{ \fBERRORS_ARE_FATAL\fP T} T{ \fBErrhandler\fP \fBERRORS_ARE_FATAL\fP T} _ T{ \fBCOMM_NULL\fP T} T{ \fBComm\fP \fBCOMM_NULL\fP T} _ T{ \fBCOMM_SELF\fP T} T{ \fBIntracomm\fP \fBCOMM_SELF\fP T} _ T{ \fBCOMM_WORLD\fP T} T{ \fBIntracomm\fP \fBCOMM_WORLD\fP T} _ T{ \fBWIN_NULL\fP T} T{ \fBWin\fP \fBWIN_NULL\fP T} _ T{ \fBFILE_NULL\fP T} T{ \fBFile\fP \fBFILE_NULL\fP T} _ T{ \fBpickle\fP T} T{ \fBPickle\fP \fBpickle\fP T} _ .TE .SH MPI4PY.FUTURES .sp New in version 3.0.0. .sp This package provides a high\-level interface for asynchronously executing callables on a pool of worker processes using MPI for inter\-process communication. .SS concurrent.futures .sp The \fI\%mpi4py.futures\fP package is based on \fBconcurrent.futures\fP from the Python standard library. More precisely, \fI\%mpi4py.futures\fP provides the \fI\%MPIPoolExecutor\fP class as a concrete implementation of the abstract class \fBExecutor\fP\&. The \fBsubmit()\fP interface schedules a callable to be executed asynchronously and returns a \fBFuture\fP object representing the execution of the callable. \fBFuture\fP instances can be queried for the call result or exception. Sets of \fBFuture\fP instances can be passed to the \fBwait()\fP and \fBas_completed()\fP functions. .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 The \fBconcurrent.futures\fP package was introduced in Python 3.2. A \fI\%backport\fP targeting Python 2.7 is available on \fI\%PyPI\fP\&. The \fI\%mpi4py.futures\fP package uses \fBconcurrent.futures\fP if available, either from the Python 3 standard library or the Python 2.7 backport if installed. Otherwise, \fI\%mpi4py.futures\fP uses a bundled copy of core functionality backported from Python 3.5 to work with Python 2.7. .UNINDENT .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 0.0 .INDENT 3.5 .INDENT 0.0 .TP .B Module \fBconcurrent.futures\fP Documentation of the \fBconcurrent.futures\fP standard module. .UNINDENT .UNINDENT .UNINDENT .SS MPIPoolExecutor .sp The \fI\%MPIPoolExecutor\fP 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 \fB__main__\fP module must be importable by worker processes, thus \fI\%MPIPoolExecutor\fP instances may not work in the interactive interpreter. .sp \fI\%MPIPoolExecutor\fP takes advantage of the dynamic process management features introduced in the MPI\-2 standard. In particular, the \fBMPI.Intracomm.Spawn\fP method of \fBMPI.COMM_SELF\fP 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 \fI\%MPIPoolExecutor\fP 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. .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 The worker processes must import the main script in order to \fIunpickle\fP any callable defined in the \fB__main__\fP module and submitted from the master process. Furthermore, the callables may need access to other global variables. At the worker processes, \fI\%mpi4py.futures\fP executes the main script code (using the \fBrunpy\fP module) under the \fB__worker__\fP namespace to define the \fB__main__\fP module. The \fB__main__\fP and \fB__worker__\fP modules are added to \fBsys.modules\fP (both at the master and worker processes) to ensure proper \fIpickling\fP and \fIunpickling\fP\&. .UNINDENT .UNINDENT .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 During the initial import phase at the workers, the main script cannot create and use new \fI\%MPIPoolExecutor\fP instances. Otherwise, each worker would attempt to spawn a new pool of workers, leading to infinite recursion. \fI\%mpi4py.futures\fP detects such recursive attempts to spawn new workers and aborts the MPI execution environment. As the main script code is run under the \fB__worker__\fP namespace, the easiest way to avoid spawn recursion is using the idiom \fBif __name__ == \(aq__main__\(aq: ...\fP in the main script. .UNINDENT .UNINDENT .INDENT 0.0 .TP .B class mpi4py.futures.MPIPoolExecutor(max_workers=None, initializer=None, initargs=(), **kwargs) An \fBExecutor\fP subclass that executes calls asynchronously using a pool of at most \fImax_workers\fP processes. If \fImax_workers\fP is \fBNone\fP or not given, its value is determined from the \fI\%MPI4PY_FUTURES_MAX_WORKERS\fP environment variable if set, or the MPI universe size if set, otherwise a single worker process is spawned. If \fImax_workers\fP is lower than or equal to \fB0\fP, then a \fBValueError\fP will be raised. .sp \fIinitializer\fP is an optional callable that is called at the start of each worker process before executing any tasks; \fIinitargs\fP is a tuple of arguments passed to the initializer. If \fIinitializer\fP raises an exception, all pending tasks and any attempt to submit new tasks to the pool will raise a \fBBrokenExecutor\fP exception. .sp Other parameters: .INDENT 7.0 .IP \(bu 2 \fIpython_exe\fP: Path to the Python interpreter executable used to spawn worker processes, otherwise \fBsys.executable\fP is used. .IP \(bu 2 \fIpython_args\fP: \fBlist\fP or iterable with additional command line flags to pass to the Python executable. Command line flags determined from inspection of \fBsys.flags\fP, \fBsys.warnoptions\fP and \fBsys._xoptions\fP in are passed unconditionally. .IP \(bu 2 \fImpi_info\fP: \fBdict\fP or iterable yielding \fB(key, value)\fP pairs. These \fB(key, value)\fP pairs are passed (through an \fBMPI.Info\fP object) to the \fBMPI.Intracomm.Spawn\fP 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. .IP \(bu 2 \fIglobals\fP: \fBdict\fP or iterable yielding \fB(name, value)\fP pairs to initialize the main module namespace in worker processes. .IP \(bu 2 \fImain\fP: If set to \fBFalse\fP, do not import the \fB__main__\fP module in worker processes. Setting \fImain\fP to \fBFalse\fP prevents worker processes from accessing definitions in the parent \fB__main__\fP namespace. .IP \(bu 2 \fIpath\fP: \fBlist\fP or iterable with paths to append to \fBsys.path\fP in worker processes to extend the module search path\&. .IP \(bu 2 \fIwdir\fP: Path to set the current working directory in worker processes using \fBos.chdir()\fP\&. The initial working directory is set by the MPI implementation. Quality MPI implementations should honor a \fBwdir\fP info key passed through \fImpi_info\fP, although such feature is not mandatory. .IP \(bu 2 \fIenv\fP: \fBdict\fP or iterable yielding \fB(name, value)\fP pairs with environment variables to update \fBos.environ\fP in worker processes. The initial environment is set by the MPI implementation. MPI implementations may allow setting the initial environment through \fImpi_info\fP, however such feature is not required nor recommended by the MPI standard. .UNINDENT .INDENT 7.0 .TP .B submit(func, *args, **kwargs) Schedule the callable, \fIfunc\fP, to be executed as \fBfunc(*args, **kwargs)\fP and returns a \fBFuture\fP object representing the execution of the callable. .INDENT 7.0 .INDENT 3.5 .sp .nf .ft C executor = MPIPoolExecutor(max_workers=1) future = executor.submit(pow, 321, 1234) print(future.result()) .ft P .fi .UNINDENT .UNINDENT .UNINDENT .INDENT 7.0 .TP .B map(func, *iterables, timeout=None, chunksize=1, **kwargs) Equivalent to \fBmap(func, *iterables)\fP except \fIfunc\fP is executed asynchronously and several calls to \fIfunc\fP may be made concurrently, out\-of\-order, in separate processes. The returned iterator raises a \fBTimeoutError\fP if \fB__next__()\fP is called and the result isn’t available after \fItimeout\fP seconds from the original call to \fI\%map()\fP\&. \fItimeout\fP can be an int or a float. If \fItimeout\fP is not specified or \fBNone\fP, 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 \fIiterables\fP 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 \fIchunksize\fP to a positive integer. For very long iterables, using a large value for \fIchunksize\fP 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 \fIunordered\fP as \fBTrue\fP, then the result iterator will yield a result as soon as any of the tasks complete. .INDENT 7.0 .INDENT 3.5 .sp .nf .ft C executor = MPIPoolExecutor(max_workers=3) for result in executor.map(pow, [2]*32, range(32)): print(result) .ft P .fi .UNINDENT .UNINDENT .UNINDENT .INDENT 7.0 .TP .B starmap(func, iterable, timeout=None, chunksize=1, **kwargs) Equivalent to \fBitertools.starmap(func, iterable)\fP\&. Used instead of \fI\%map()\fP when argument parameters are already grouped in tuples from a single iterable (the data has been “pre\-zipped”). \fI\%map(func, *iterable)\fP is equivalent to \fI\%starmap(func, zip(*iterable))\fP\&. .INDENT 7.0 .INDENT 3.5 .sp .nf .ft C executor = MPIPoolExecutor(max_workers=3) iterable = ((2, n) for n in range(32)) for result in executor.starmap(pow, iterable): print(result) .ft P .fi .UNINDENT .UNINDENT .UNINDENT .INDENT 7.0 .TP .B shutdown(wait=True, cancel_futures=False) Signal the executor that it should free any resources that it is using when the currently pending futures are done executing. Calls to \fI\%submit()\fP and \fI\%map()\fP made after \fI\%shutdown()\fP will raise \fBRuntimeError\fP\&. .sp If \fIwait\fP is \fBTrue\fP 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 \fIwait\fP is \fBFalse\fP 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 \fIwait\fP, the entire Python program will not exit until all pending futures are done executing. .sp If \fIcancel_futures\fP is \fBTrue\fP, 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 \fIcancel_futures\fP\&. .sp You can avoid having to call this method explicitly if you use the \fBwith\fP statement, which will shutdown the executor instance (waiting as if \fI\%shutdown()\fP were called with \fIwait\fP set to \fBTrue\fP). .INDENT 7.0 .INDENT 3.5 .sp .nf .ft C import time with MPIPoolExecutor(max_workers=1) as executor: future = executor.submit(time.sleep, 2) assert future.done() .ft P .fi .UNINDENT .UNINDENT .UNINDENT .INDENT 7.0 .TP .B bootup(wait=True) Signal the executor that it should allocate eagerly any required resources (in particular, MPI worker processes). If \fIwait\fP is \fBTrue\fP, then \fI\%bootup()\fP will not return until the executor resources are ready to process submissions. Resources are automatically allocated in the first call to \fI\%submit()\fP, thus calling \fI\%bootup()\fP explicitly is seldom needed. .UNINDENT .UNINDENT .INDENT 0.0 .TP .B MPI4PY_FUTURES_MAX_WORKERS If the \fImax_workers\fP parameter to \fI\%MPIPoolExecutor\fP is \fBNone\fP or not given, the \fI\%MPI4PY_FUTURES_MAX_WORKERS\fP environment variable provides fallback value for the maximum number of MPI worker processes to spawn. .UNINDENT .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 As the master process uses a separate thread to perform MPI communication with the workers, the backend MPI implementation should provide support for \fBMPI.THREAD_MULTIPLE\fP\&. 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 \fBMPI.THREAD_MULTIPLE\fP, \fI\%mpi4py.futures\fP will use a global lock to serialize MPI calls. If the level of thread support is less than \fBMPI.THREAD_SERIALIZED\fP, \fI\%mpi4py.futures\fP will emit a \fBRuntimeWarning\fP\&. .UNINDENT .UNINDENT .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 If the level of thread support in the backend MPI is less than \fBMPI.THREAD_SERIALIZED\fP (i.e, it is either \fBMPI.THREAD_SINGLE\fP or \fBMPI.THREAD_FUNNELED\fP), in theory \fI\%mpi4py.futures\fP cannot be used. Rather than raising an exception, \fI\%mpi4py.futures\fP 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. .UNINDENT .UNINDENT .SS MPICommExecutor .sp 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 \fBMPI_Comm_spawn()\fP routine. .sp With these issues in mind, \fI\%mpi4py.futures\fP supports an additonal, more traditional, SPMD\-like usage pattern requiring MPI\-1 calls only. Python applications are started the usual way, e.g., using the \fBmpiexec\fP command. Python code should make a collective call to the \fI\%MPICommExecutor\fP 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 \fI\%MPIPoolExecutor\fP instance to submit tasks. Meanwhile, the worker process follow a different execution path and team\-up to execute the tasks submitted from the master. .sp Besides alleviating the lack of dynamic process managment features in legacy MPI\-1 or partial MPI\-2 implementations, the \fI\%MPICommExecutor\fP 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 \fI\%mpi4py.futures\fP package. .INDENT 0.0 .TP .B class mpi4py.futures.MPICommExecutor(comm=None, root=0) Context manager for \fI\%MPIPoolExecutor\fP\&. This context manager splits a MPI (intra)communicator \fIcomm\fP (defaults to \fBMPI.COMM_WORLD\fP if not provided or \fBNone\fP) in two disjoint sets: a single master process (with rank \fIroot\fP in \fIcomm\fP) and the remaining worker processes. These sets are then connected through an intercommunicator. The target of the \fBwith\fP statement is assigned either an \fI\%MPIPoolExecutor\fP instance (at the master) or \fBNone\fP (at the workers). .INDENT 7.0 .INDENT 3.5 .sp .nf .ft C 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} .ft P .fi .UNINDENT .UNINDENT .UNINDENT .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 If \fI\%MPICommExecutor\fP is passed a communicator of size one (e.g., \fBMPI.COMM_SELF\fP), then the executor instace assigned to the target of the \fBwith\fP statement will execute all submitted tasks in a single worker thread, thus ensuring that task execution still progress asynchronously. However, the \fI\%GIL\fP 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 \fI\%GIL\fP is not usually an issue, however, as a single worker thread is used, it progress one task at a time. We advice against using \fI\%MPICommExecutor\fP with communicators of size one and suggest refactoring your code to use instead a \fBThreadPoolExecutor\fP\&. .UNINDENT .UNINDENT .SS Command line .sp Recalling the issues related to the lack of support for dynamic process managment features in MPI implementations, \fI\%mpi4py.futures\fP supports an alternative usage pattern where Python code (either from scripts, modules, or zip files) is run under command line control of the \fI\%mpi4py.futures\fP package by passing \fB\-m mpi4py.futures\fP to the \fBpython\fP executable. The \fBmpi4py.futures\fP invocation should be passed a \fIpyfile\fP path to a script (or a zipfile/directory containing a \fB__main__.py\fP file). Additionally, \fBmpi4py.futures\fP accepts \fB\-m \fP\fImod\fP to execute a module named \fImod\fP, \fB\-c \fP\fIcmd\fP to execute a command string \fIcmd\fP, or even \fB\-\fP to read commands from standard input (\fBsys.stdin\fP). Summarizing, \fBmpi4py.futures\fP can be invoked in the following ways: .INDENT 0.0 .IP \(bu 2 \fB$ mpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py.futures \fP\fIpyfile\fP\fB [arg] ...\fP .IP \(bu 2 \fB$ mpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py.futures \-m \fP\fImod\fP\fB [arg] ...\fP .IP \(bu 2 \fB$ mpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py.futures \-c \fP\fIcmd\fP\fB [arg] ...\fP .IP \(bu 2 \fB$ mpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py.futures \- [arg] ...\fP .UNINDENT .sp Before starting the main script execution, \fI\%mpi4py.futures\fP splits \fBMPI.COMM_WORLD\fP in one master (the process with rank 0 in \fBMPI.COMM_WORLD\fP) and \fInumprocs \- 1\fP workers and connects them through an MPI intercommunicator. Afterwards, the master process proceeds with the execution of the user script code, which eventually creates \fI\%MPIPoolExecutor\fP 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 \fBMPI.COMM_WORLD.Abort(1)\fP to prevent deadlocks and force termination of entire MPI execution environment. .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 Running scripts under command line control of \fI\%mpi4py.futures\fP 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 \fI\%MPIPoolExecutor\fP 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 \fI\%MPIPoolExecutor.map()\fP with long iterables). If that ever happens, submissions from other executors will not be serviced until free workers are available. .UNINDENT .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 0.0 .INDENT 3.5 .INDENT 0.0 .TP .B python:using\-on\-cmdline Documentation on Python command line interface. .UNINDENT .UNINDENT .UNINDENT .SS Examples .sp The following \fBjulia.py\fP script computes the \fI\%Julia set\fP and dumps an image to disk in binary \fI\%PGM\fP format. The code starts by importing \fI\%MPIPoolExecutor\fP from the \fI\%mpi4py.futures\fP package. Next, some global constants and functions implement the computation of the Julia set. The computations are protected with the standard \fBif __name__ == \(aq__main__\(aq:...\fP idiom. The image is computed by whole scanlines submitting all these tasks at once using the \fI\%map\fP method. The result iterator yields scanlines in\-order as the tasks complete. Finally, each scanline is dumped to disk. .sp \fBjulia.py\fP .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C 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__ == \(aq__main__\(aq: with MPIPoolExecutor() as executor: image = executor.map(julia_line, range(h)) with open(\(aqjulia.pgm\(aq, \(aqwb\(aq) as f: f.write(b\(aqP5 %d %d %d\en\(aq % (w, h, 255)) for line in image: f.write(line) .ft P .fi .UNINDENT .UNINDENT .sp The recommended way to execute the script is by using the \fBmpiexec\fP 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 \fBmpiexec\fP or environment variables. Below we provide examples for MPICH and Open MPI implementations [1]\&. In all of these examples, the \fBmpiexec\fP command launches a single master process running the Python interpreter and executing the main script. When required, \fI\%mpi4py.futures\fP 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. .sp When using MPICH implementation or its derivatives based on the Hydra process manager, users can set the MPI universe size via the \fB\-usize\fP argument to \fBmpiexec\fP: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 1 \-usize 17 python julia.py .ft P .fi .UNINDENT .UNINDENT .sp or, alternatively, by setting the \fBMPIEXEC_UNIVERSE_SIZE\fP environment variable: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ MPIEXEC_UNIVERSE_SIZE=17 mpiexec \-n 1 python julia.py .ft P .fi .UNINDENT .UNINDENT .sp In the Open MPI implementation, the MPI universe size can be set via the \fB\-host\fP argument to \fBmpiexec\fP: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 1 \-host :17 python julia.py .ft P .fi .UNINDENT .UNINDENT .sp Another way to specify the number of workers is to use the \fI\%mpi4py.futures\fP\-specific environment variable \fI\%MPI4PY_FUTURES_MAX_WORKERS\fP: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ MPI4PY_FUTURES_MAX_WORKERS=16 mpiexec \-n 1 python julia.py .ft P .fi .UNINDENT .UNINDENT .sp Note that in this case, the MPI universe size is ignored. .sp 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 \fI\%mpi4py.futures\fP passing the \-m flag to the \fBpython\fP executable: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 17 python \-m mpi4py.futures julia.py .ft P .fi .UNINDENT .UNINDENT .sp 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. .IP [1] 5 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. .INDENT 0.0 .TP .B GIL See global interpreter lock\&. .UNINDENT .SH MPI4PY.UTIL .sp New in version 3.1.0. .sp The \fI\%mpi4py.util\fP package collects miscellaneous utilities within the intersection of Python and MPI. .SS mpi4py.util.pkl5 .sp New in version 3.1.0. .sp \fBpickle\fP protocol 5 (see \fI\%PEP 574\fP) introduced support for out\-of\-band buffers, allowing for more efficient handling of certain object types with large memory footprints. .sp 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. .sp The \fI\%mpi4py.util.pkl5\fP module provides communicator wrapper classes reimplementing pickle\-based point\-to\-point communication methods using pickle protocol 5. Handling out\-of\-band buffers necessarily involve 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). .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 Support for pickle protocol 5 is available in the \fBpickle\fP module within the Python standard library since Python 3.8. Previous Python 3 releases can use the \fBpickle5\fP backport, which is available on \fI\%PyPI\fP and can be installed with: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C python \-m pip install pickle5 .ft P .fi .UNINDENT .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B class mpi4py.util.pkl5.Request(request=None) Request. .sp Custom request class for nonblocking communications. .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 \fI\%Request\fP is not a subclass of \fBmpi4py.MPI.Request\fP .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIIterable\fP\fI[\fP\fIMPI.Request\fP\fI]\fP) – .TP .B Return type Request .UNINDENT .INDENT 7.0 .TP .B Free() Free a communication request. .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B cancel() Cancel a communication request. .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B get_status(status=None) Non\-destructive test for the completion of a request. .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B test(status=None) Test for the completion of a request. .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Tuple[bool, Optional[Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B wait(status=None) Wait for a request to complete. .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod testall(requests, statuses=None) Test for the completion of all requests. .INDENT 7.0 .TP .B Classmethod .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod waitall(requests, statuses=None) Wait for all requests to complete. .INDENT 7.0 .TP .B Classmethod .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B class mpi4py.util.pkl5.Message(message=None) Message. .sp Custom message class for matching probes. .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 \fI\%Message\fP is not a subclass of \fBmpi4py.MPI.Message\fP .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters \fBmessage\fP (\fIIterable\fP\fI[\fP\fIMPI.Message\fP\fI]\fP) – .TP .B Return type Message .UNINDENT .INDENT 7.0 .TP .B recv(status=None) Blocking receive of matched message. .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B irecv() Nonblocking receive of matched message. .INDENT 7.0 .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod probe(comm, source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a matched message. .INDENT 7.0 .TP .B Classmethod .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod iprobe(comm, source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a matched message. .INDENT 7.0 .TP .B Classmethod .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B class mpi4py.util.pkl5.Comm Communicator. .sp Base communicator wrapper class. .INDENT 7.0 .TP .B send(obj, dest, tag=0) Blocking send in standard mode. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B bsend(obj, dest, tag=0) Blocking send in buffered mode. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B ssend(obj, dest, tag=0) Blocking send in synchronous mode. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B isend(obj, dest, tag=0) Nonblocking send in standard mode. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B ibsend(obj, dest, tag=0) Nonblocking send in buffered mode. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B issend(obj, dest, tag=0) Nonblocking send in synchronous mode. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B recv(buf=None, source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking receive. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIOptional\fP\fI[\fP\fIBuffer\fP\fI]\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B irecv(buf=None, source=ANY_SOURCE, tag=ANY_TAG) Nonblocking receive. .sp \fBWARNING:\fP .INDENT 7.0 .INDENT 3.5 This method cannot be supported reliably and raises \fBRuntimeError\fP\&. .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIOptional\fP\fI[\fP\fIBuffer\fP\fI]\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B sendrecv(sendobj, dest, sendtag=0, recvbuf=None, source=ANY_SOURCE, recvtag=ANY_TAG, status=None) Send and receive. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBsendtag\fP (\fIint\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBuffer\fP\fI]\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBrecvtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B mprobe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a matched message. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Message .UNINDENT .UNINDENT .INDENT 7.0 .TP .B improbe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a matched message. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[Message] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B bcast(obj, root=0) Broadcast. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .UNINDENT .INDENT 0.0 .TP .B class mpi4py.util.pkl5.Intracomm Intracommunicator. .sp Intracommunicator wrapper class. .UNINDENT .INDENT 0.0 .TP .B class mpi4py.util.pkl5.Intercomm Intercommunicator. .sp Intercommunicator wrapper class. .UNINDENT .SS Examples .sp \fBtest\-pkl5\-1.py\fP .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi4\(aq) # > 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 .ft P .fi .UNINDENT .UNINDENT .sp \fBtest\-pkl5\-2.py\fP .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C 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=\(aqi4\(aq) # > 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 .ft P .fi .UNINDENT .UNINDENT .SS mpi4py.util.dtlib .sp New in version 3.1.0. .sp The \fI\%mpi4py.util.dtlib\fP module provides converter routines between NumPy and MPI datatypes. .INDENT 0.0 .TP .B mpi4py.util.dtlib.from_numpy_dtype(dtype) Convert NumPy datatype to MPI datatype. .INDENT 7.0 .TP .B Parameters \fBdtype\fP (\fInumpy.typing.DTypeLike\fP) – NumPy dtype\-like object. .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 0.0 .TP .B mpi4py.util.dtlib.to_numpy_dtype(datatype) Convert MPI datatype to NumPy datatype. .INDENT 7.0 .TP .B Parameters \fBdatatype\fP (\fIDatatype\fP) – MPI datatype. .TP .B Return type numpy.dtype .UNINDENT .UNINDENT .SH MPI4PY.RUN .sp New in version 3.0.0. .sp At import time, \fBmpi4py\fP initializes the MPI execution environment calling \fBMPI_Init_thread()\fP and installs an exit hook to automatically call \fBMPI_Finalize()\fP just before the Python process terminates. Additionally, \fBmpi4py\fP overrides the default \fBERRORS_ARE_FATAL\fP error handler in favor of \fBERRORS_RETURN\fP, 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 \fBmpi4py\fP can just \fBfrom mpi4py import MPI\fP and perform MPI calls without the tedious initialization/finalization handling. MPI errors, once translated automatically to Python exceptions, can be dealt with the common \fBtry\fP…\fBexcept\fP…\fBfinally\fP clauses; unhandled MPI exceptions will print a traceback which helps in locating problems in source code. .sp 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. .sp Consider the following snippet of Python code. Assume this code is stored in a standard Python script file and run with \fBmpiexec\fP in two or more processes. .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C 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) .ft P .fi .UNINDENT .UNINDENT .sp Process 0 raises \fBZeroDivisionError\fP 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 \fBmpi4py\fP installed a finalizer hook to call \fBMPI_Finalize()\fP before exit, process 0 will block waiting for other processes to also enter the \fBMPI_Finalize()\fP call. Meanwhile, process 1 will block waiting for a message to arrive from process 0, thus never reaching to \fBMPI_Finalize()\fP\&. The whole MPI execution environment is irremediably in a deadlock state. .sp To alleviate this issue, \fBmpi4py\fP offers a simple, alternative command line execution mechanism based on using the \-m flag and implemented with the \fBrunpy\fP module. To use this features, Python code should be run passing \fB\-m mpi4py\fP in the command line invoking the Python interpreter. In case of unhandled exceptions, the finalizer hook will call \fBMPI_Abort()\fP on the \fBMPI_COMM_WORLD\fP communicator, thus effectively aborting the MPI execution environment. .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 When a process is forced to abort, resources (e.g. open files) are not cleaned\-up and any registered finalizers (either with the \fBatexit\fP module, the Python C/API function \fBPy_AtExit()\fP, or even the C standard library function \fBatexit()\fP) 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. .UNINDENT .UNINDENT .SS Interface options .sp The use of \fB\-m mpi4py\fP to execute Python code on the command line resembles that of the Python interpreter. .INDENT 0.0 .IP \(bu 2 \fBmpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py \fP\fIpyfile\fP\fB [arg] ...\fP .IP \(bu 2 \fBmpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py \-m \fP\fImod\fP\fB [arg] ...\fP .IP \(bu 2 \fBmpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py \-c \fP\fIcmd\fP\fB [arg] ...\fP .IP \(bu 2 \fBmpiexec \-n \fP\fInumprocs\fP\fB python \-m mpi4py \- [arg] ...\fP .UNINDENT .INDENT 0.0 .TP .B Execute the Python code contained in \fIpyfile\fP, which must be a filesystem path referring to either a Python file, a directory containing a \fB__main__.py\fP file, or a zipfile containing a \fB__main__.py\fP file. .UNINDENT .INDENT 0.0 .TP .B \-m Search \fBsys.path\fP for the named module \fImod\fP and execute its contents. .UNINDENT .INDENT 0.0 .TP .B \-c Execute the Python code in the \fIcmd\fP string command. .UNINDENT .INDENT 0.0 .TP .B \- Read commands from standard input (\fBsys.stdin\fP). .UNINDENT .sp \fBSEE ALSO:\fP .INDENT 0.0 .INDENT 3.5 .INDENT 0.0 .TP .B python:using\-on\-cmdline Documentation on Python command line interface. .UNINDENT .UNINDENT .UNINDENT .SH REFERENCE .TS center; |l|l|. _ T{ \fBmpi4py.MPI\fP T} T{ Message Passing Interface. T} _ .TE .SS mpi4py.MPI .sp Message Passing Interface. .sp Classes .TS center; |l|l|. _ T{ \fBCartcomm\fP([comm]) T} T{ Cartesian topology intracommunicator T} _ T{ \fBComm\fP([comm]) T} T{ Communicator T} _ T{ \fBDatatype\fP([datatype]) T} T{ Datatype object T} _ T{ \fBDistgraphcomm\fP([comm]) T} T{ Distributed graph topology intracommunicator T} _ T{ \fBErrhandler\fP([errhandler]) T} T{ Error handler T} _ T{ \fBFile\fP([file]) T} T{ File handle T} _ T{ \fBGraphcomm\fP([comm]) T} T{ General graph topology intracommunicator T} _ T{ \fBGrequest\fP([request]) T} T{ Generalized request handle T} _ T{ \fBGroup\fP([group]) T} T{ Group of processes T} _ T{ \fBInfo\fP([info]) T} T{ Info object T} _ T{ \fBIntercomm\fP([comm]) T} T{ Intercommunicator T} _ T{ \fBIntracomm\fP([comm]) T} T{ Intracommunicator T} _ T{ \fBMessage\fP([message]) T} T{ Matched message handle T} _ T{ \fBOp\fP([op]) T} T{ Operation object T} _ T{ \fBPickle\fP([dumps, loads, protocol]) T} T{ Pickle/unpickle Python objects T} _ T{ \fBPrequest\fP([request]) T} T{ Persistent request handle T} _ T{ \fBRequest\fP([request]) T} T{ Request handle T} _ T{ \fBStatus\fP([status]) T} T{ Status object T} _ T{ \fBTopocomm\fP([comm]) T} T{ Topology intracommunicator T} _ T{ \fBWin\fP([win]) T} T{ Window handle T} _ T{ \fBmemory\fP(buf) T} T{ Memory buffer T} _ .TE .SS mpi4py.MPI.Cartcomm .INDENT 0.0 .TP .B class mpi4py.MPI.Cartcomm(comm=None) Bases: \fBmpi4py.MPI.Topocomm\fP .sp Cartesian topology intracommunicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fICartcomm\fP\fI]\fP) – .TP .B Return type Cartcomm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fICartcomm\fP\fI]\fP) – .TP .B Return type Cartcomm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Get_cart_rank\fP(coords) T} T{ Translate logical coordinates to ranks T} _ T{ \fI\%Get_coords\fP(rank) T} T{ Translate ranks to logical coordinates T} _ T{ \fI\%Get_dim\fP() T} T{ Return number of dimensions T} _ T{ \fI\%Get_topo\fP() T} T{ Return information on the cartesian topology T} _ T{ \fI\%Shift\fP(direction, disp) T} T{ Return a tuple (source, dest) of process ranks for data shifting with Comm.Sendrecv() T} _ T{ \fI\%Sub\fP(remain_dims) T} T{ Return cartesian communicators that form lower\-dimensional subgrids T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%coords\fP T} T{ coordinates T} _ T{ \fI\%dim\fP T} T{ number of dimensions T} _ T{ \fI\%dims\fP T} T{ dimensions T} _ T{ \fI\%ndim\fP T} T{ number of dimensions T} _ T{ \fI\%periods\fP T} T{ periodicity T} _ T{ \fI\%topo\fP T} T{ topology information T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Get_cart_rank(coords) Translate logical coordinates to ranks .INDENT 7.0 .TP .B Parameters \fBcoords\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_coords(rank) Translate ranks to logical coordinates .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type List[int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_dim() Return number of dimensions .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_topo() Return information on the cartesian topology .INDENT 7.0 .TP .B Return type Tuple[List[int], List[int], List[int]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Shift(direction, disp) Return a tuple (source, dest) of process ranks for data shifting with Comm.Sendrecv() .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdirection\fP (\fIint\fP) – .IP \(bu 2 \fBdisp\fP (\fIint\fP) – .UNINDENT .TP .B Return type Tuple[int, int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Sub(remain_dims) Return cartesian communicators that form lower\-dimensional subgrids .INDENT 7.0 .TP .B Parameters \fBremain_dims\fP (\fISequence\fP\fI[\fP\fIbool\fP\fI]\fP) – .TP .B Return type Cartcomm .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B coords coordinates .UNINDENT .INDENT 7.0 .TP .B dim number of dimensions .UNINDENT .INDENT 7.0 .TP .B dims dimensions .UNINDENT .INDENT 7.0 .TP .B ndim number of dimensions .UNINDENT .INDENT 7.0 .TP .B periods periodicity .UNINDENT .INDENT 7.0 .TP .B topo topology information .UNINDENT .UNINDENT .SS mpi4py.MPI.Comm .INDENT 0.0 .TP .B class mpi4py.MPI.Comm(comm=None) Bases: \fBobject\fP .sp Communicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIComm\fP\fI]\fP) – .TP .B Return type Comm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIComm\fP\fI]\fP) – .TP .B Return type Comm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Abort\fP([errorcode]) T} T{ Terminate MPI execution environment T} _ T{ \fI\%Allgather\fP(sendbuf, recvbuf) T} T{ Gather to All, gather data from all processes and distribute it to all other processes in a group T} _ T{ \fI\%Allgatherv\fP(sendbuf, recvbuf) T} T{ Gather to All Vector, gather data from all processes and distribute it to all other processes in a group providing different amount of data and displacements T} _ T{ \fI\%Allreduce\fP(sendbuf, recvbuf[, op]) T} T{ Reduce to All T} _ T{ \fI\%Alltoall\fP(sendbuf, recvbuf) T} T{ All to All Scatter/Gather, send data from all to all processes in a group T} _ T{ \fI\%Alltoallv\fP(sendbuf, recvbuf) T} T{ All to All Scatter/Gather Vector, send data from all to all processes in a group providing different amount of data and displacements T} _ T{ \fI\%Alltoallw\fP(sendbuf, recvbuf) T} T{ Generalized All\-to\-All communication allowing different counts, displacements and datatypes for each partner T} _ T{ \fI\%Barrier\fP() T} T{ Barrier synchronization T} _ T{ \fI\%Bcast\fP(buf[, root]) T} T{ Broadcast a message from one process to all other processes in a group T} _ T{ \fI\%Bsend\fP(buf, dest[, tag]) T} T{ Blocking send in buffered mode T} _ T{ \fI\%Bsend_init\fP(buf, dest[, tag]) T} T{ Persistent request for a send in buffered mode T} _ T{ \fI\%Call_errhandler\fP(errorcode) T} T{ Call the error handler installed on a communicator T} _ T{ \fI\%Clone\fP() T} T{ Clone an existing communicator T} _ T{ \fI\%Compare\fP(comm1, comm2) T} T{ Compare two communicators T} _ T{ \fI\%Create\fP(group) T} T{ Create communicator from group T} _ T{ \fI\%Create_group\fP(group[, tag]) T} T{ Create communicator from group T} _ T{ \fI\%Create_keyval\fP([copy_fn, delete_fn, nopython]) T} T{ Create a new attribute key for communicators T} _ T{ \fI\%Delete_attr\fP(keyval) T} T{ Delete attribute value associated with a key T} _ T{ \fI\%Disconnect\fP() T} T{ Disconnect from a communicator T} _ T{ \fI\%Dup\fP([info]) T} T{ Duplicate an existing communicator T} _ T{ \fI\%Dup_with_info\fP(info) T} T{ Duplicate an existing communicator T} _ T{ \fI\%Free\fP() T} T{ Free a communicator T} _ T{ \fI\%Free_keyval\fP(keyval) T} T{ Free an attribute key for communicators T} _ T{ \fI\%Gather\fP(sendbuf, recvbuf[, root]) T} T{ Gather together values from a group of processes T} _ T{ \fI\%Gatherv\fP(sendbuf, recvbuf[, root]) T} T{ Gather Vector, gather data to one process from all other processes in a group providing different amount of data and displacements at the receiving sides T} _ T{ \fI\%Get_attr\fP(keyval) T} T{ Retrieve attribute value by key T} _ T{ \fI\%Get_errhandler\fP() T} T{ Get the error handler for a communicator T} _ T{ \fI\%Get_group\fP() T} T{ Access the group associated with a communicator T} _ T{ \fI\%Get_info\fP() T} T{ Return the hints for a communicator that are currently in use T} _ T{ \fI\%Get_name\fP() T} T{ Get the print name for this communicator T} _ T{ \fI\%Get_parent\fP() T} T{ Return the parent intercommunicator for this process T} _ T{ \fI\%Get_rank\fP() T} T{ Return the rank of this process in a communicator T} _ T{ \fI\%Get_size\fP() T} T{ Return the number of processes in a communicator T} _ T{ \fI\%Get_topology\fP() T} T{ Determine the type of topology (if any) associated with a communicator T} _ T{ \fI\%Iallgather\fP(sendbuf, recvbuf) T} T{ Nonblocking Gather to All T} _ T{ \fI\%Iallgatherv\fP(sendbuf, recvbuf) T} T{ Nonblocking Gather to All Vector T} _ T{ \fI\%Iallreduce\fP(sendbuf, recvbuf[, op]) T} T{ Nonblocking Reduce to All T} _ T{ \fI\%Ialltoall\fP(sendbuf, recvbuf) T} T{ Nonblocking All to All Scatter/Gather T} _ T{ \fI\%Ialltoallv\fP(sendbuf, recvbuf) T} T{ Nonblocking All to All Scatter/Gather Vector T} _ T{ \fI\%Ialltoallw\fP(sendbuf, recvbuf) T} T{ Nonblocking Generalized All\-to\-All T} _ T{ \fI\%Ibarrier\fP() T} T{ Nonblocking Barrier T} _ T{ \fI\%Ibcast\fP(buf[, root]) T} T{ Nonblocking Broadcast T} _ T{ \fI\%Ibsend\fP(buf, dest[, tag]) T} T{ Nonblocking send in buffered mode T} _ T{ \fI\%Idup\fP() T} T{ Nonblocking duplicate an existing communicator T} _ T{ \fI\%Igather\fP(sendbuf, recvbuf[, root]) T} T{ Nonblocking Gather T} _ T{ \fI\%Igatherv\fP(sendbuf, recvbuf[, root]) T} T{ Nonblocking Gather Vector T} _ T{ \fI\%Improbe\fP([source, tag, status]) T} T{ Nonblocking test for a matched message T} _ T{ \fI\%Iprobe\fP([source, tag, status]) T} T{ Nonblocking test for a message T} _ T{ \fI\%Irecv\fP(buf[, source, tag]) T} T{ Nonblocking receive T} _ T{ \fI\%Ireduce\fP(sendbuf, recvbuf[, op, root]) T} T{ Nonblocking Reduce to Root T} _ T{ \fI\%Ireduce_scatter\fP(sendbuf, recvbuf[, ...]) T} T{ Nonblocking Reduce\-Scatter (vector version) T} _ T{ \fI\%Ireduce_scatter_block\fP(sendbuf, recvbuf[, op]) T} T{ Nonblocking Reduce\-Scatter Block (regular, non\-vector version) T} _ T{ \fI\%Irsend\fP(buf, dest[, tag]) T} T{ Nonblocking send in ready mode T} _ T{ \fI\%Is_inter\fP() T} T{ Test to see if a comm is an intercommunicator T} _ T{ \fI\%Is_intra\fP() T} T{ Test to see if a comm is an intracommunicator T} _ T{ \fI\%Iscatter\fP(sendbuf, recvbuf[, root]) T} T{ Nonblocking Scatter T} _ T{ \fI\%Iscatterv\fP(sendbuf, recvbuf[, root]) T} T{ Nonblocking Scatter Vector T} _ T{ \fI\%Isend\fP(buf, dest[, tag]) T} T{ Nonblocking send T} _ T{ \fI\%Issend\fP(buf, dest[, tag]) T} T{ Nonblocking send in synchronous mode T} _ T{ \fI\%Join\fP(fd) T} T{ Create a intercommunicator by joining two processes connected by a socket T} _ T{ \fI\%Mprobe\fP([source, tag, status]) T} T{ Blocking test for a matched message T} _ T{ \fI\%Probe\fP([source, tag, status]) T} T{ Blocking test for a message T} _ T{ \fI\%Recv\fP(buf[, source, tag, status]) T} T{ Blocking receive T} _ T{ \fI\%Recv_init\fP(buf[, source, tag]) T} T{ Create a persistent request for a receive T} _ T{ \fI\%Reduce\fP(sendbuf, recvbuf[, op, root]) T} T{ Reduce to Root T} _ T{ \fI\%Reduce_scatter\fP(sendbuf, recvbuf[, ...]) T} T{ Reduce\-Scatter (vector version) T} _ T{ \fI\%Reduce_scatter_block\fP(sendbuf, recvbuf[, op]) T} T{ Reduce\-Scatter Block (regular, non\-vector version) T} _ T{ \fI\%Rsend\fP(buf, dest[, tag]) T} T{ Blocking send in ready mode T} _ T{ \fI\%Rsend_init\fP(buf, dest[, tag]) T} T{ Persistent request for a send in ready mode T} _ T{ \fI\%Scatter\fP(sendbuf, recvbuf[, root]) T} T{ Scatter data from one process to all other processes in a group T} _ T{ \fI\%Scatterv\fP(sendbuf, recvbuf[, root]) T} T{ Scatter Vector, scatter data from one process to all other processes in a group providing different amount of data and displacements at the sending side T} _ T{ \fI\%Send\fP(buf, dest[, tag]) T} T{ Blocking send T} _ T{ \fI\%Send_init\fP(buf, dest[, tag]) T} T{ Create a persistent request for a standard send T} _ T{ \fI\%Sendrecv\fP(sendbuf, dest[, sendtag, recvbuf, ...]) T} T{ Send and receive a message T} _ T{ \fI\%Sendrecv_replace\fP(buf, dest[, sendtag, ...]) T} T{ Send and receive a message T} _ T{ \fI\%Set_attr\fP(keyval, attrval) T} T{ Store attribute value associated with a key T} _ T{ \fI\%Set_errhandler\fP(errhandler) T} T{ Set the error handler for a communicator T} _ T{ \fI\%Set_info\fP(info) T} T{ Set new values for the hints associated with a communicator T} _ T{ \fI\%Set_name\fP(name) T} T{ Set the print name for this communicator T} _ T{ \fI\%Split\fP([color, key]) T} T{ Split communicator by color and key T} _ T{ \fI\%Split_type\fP(split_type[, key, info]) T} T{ Split communicator by split type T} _ T{ \fI\%Ssend\fP(buf, dest[, tag]) T} T{ Blocking send in synchronous mode T} _ T{ \fI\%Ssend_init\fP(buf, dest[, tag]) T} T{ Persistent request for a send in synchronous mode T} _ T{ \fI\%allgather\fP(sendobj) T} T{ Gather to All T} _ T{ \fI\%allreduce\fP(sendobj[, op]) T} T{ Reduce to All T} _ T{ \fI\%alltoall\fP(sendobj) T} T{ All to All Scatter/Gather T} _ T{ \fI\%barrier\fP() T} T{ Barrier T} _ T{ \fI\%bcast\fP(obj[, root]) T} T{ Broadcast T} _ T{ \fI\%bsend\fP(obj, dest[, tag]) T} T{ Send in buffered mode T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%gather\fP(sendobj[, root]) T} T{ Gather T} _ T{ \fI\%ibsend\fP(obj, dest[, tag]) T} T{ Nonblocking send in buffered mode T} _ T{ \fI\%improbe\fP([source, tag, status]) T} T{ Nonblocking test for a matched message T} _ T{ \fI\%iprobe\fP([source, tag, status]) T} T{ Nonblocking test for a message T} _ T{ \fI\%irecv\fP([buf, source, tag]) T} T{ Nonblocking receive T} _ T{ \fI\%isend\fP(obj, dest[, tag]) T} T{ Nonblocking send T} _ T{ \fI\%issend\fP(obj, dest[, tag]) T} T{ Nonblocking send in synchronous mode T} _ T{ \fI\%mprobe\fP([source, tag, status]) T} T{ Blocking test for a matched message T} _ T{ \fI\%probe\fP([source, tag, status]) T} T{ Blocking test for a message T} _ T{ \fI\%py2f\fP() T} T{ T} _ T{ \fI\%recv\fP([buf, source, tag, status]) T} T{ Receive T} _ T{ \fI\%reduce\fP(sendobj[, op, root]) T} T{ Reduce to Root T} _ T{ \fI\%scatter\fP(sendobj[, root]) T} T{ Scatter T} _ T{ \fI\%send\fP(obj, dest[, tag]) T} T{ Send T} _ T{ \fI\%sendrecv\fP(sendobj, dest[, sendtag, recvbuf, ...]) T} T{ Send and Receive T} _ T{ \fI\%ssend\fP(obj, dest[, tag]) T} T{ Send in synchronous mode T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%group\fP T} T{ communicator group T} _ T{ \fI\%info\fP T} T{ communicator info T} _ T{ \fI\%is_inter\fP T} T{ is intercommunicator T} _ T{ \fI\%is_intra\fP T} T{ is intracommunicator T} _ T{ \fI\%is_topo\fP T} T{ is a topology communicator T} _ T{ \fI\%name\fP T} T{ communicator name T} _ T{ \fI\%rank\fP T} T{ rank of this process in communicator T} _ T{ \fI\%size\fP T} T{ number of processes in communicator T} _ T{ \fI\%topology\fP T} T{ communicator topology type T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Abort(errorcode=0) Terminate MPI execution environment .sp \fBWARNING:\fP .INDENT 7.0 .INDENT 3.5 This is a direct call, use it with care!!!. .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters \fBerrorcode\fP (\fIint\fP) – .TP .B Return type NoReturn .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Allgather(sendbuf, recvbuf) Gather to All, gather data from all processes and distribute it to all other processes in a group .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Allgatherv(sendbuf, recvbuf) Gather to All Vector, gather data from all processes and distribute it to all other processes in a group providing different amount of data and displacements .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Allreduce(sendbuf, recvbuf, op=SUM) Reduce to All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Alltoall(sendbuf, recvbuf) All to All Scatter/Gather, send data from all to all processes in a group .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecB\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Alltoallv(sendbuf, recvbuf) All to All Scatter/Gather Vector, send data from all to all processes in a group providing different amount of data and displacements .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecV\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Alltoallw(sendbuf, recvbuf) Generalized All\-to\-All communication allowing different counts, displacements and datatypes for each partner .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecW\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecW\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Barrier() Barrier synchronization .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Bcast(buf, root=0) Broadcast a message from one process to all other processes in a group .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Bsend(buf, dest, tag=0) Blocking send in buffered mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Bsend_init(buf, dest, tag=0) Persistent request for a send in buffered mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Call_errhandler(errorcode) Call the error handler installed on a communicator .INDENT 7.0 .TP .B Parameters \fBerrorcode\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Clone() Clone an existing communicator .INDENT 7.0 .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Compare(comm1, comm2) Compare two communicators .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcomm1\fP (\fIComm\fP) – .IP \(bu 2 \fBcomm2\fP (\fIComm\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create(group) Create communicator from group .INDENT 7.0 .TP .B Parameters \fBgroup\fP (\fIGroup\fP) – .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_group(group, tag=0) Create communicator from group .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup\fP (\fIGroup\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_keyval(copy_fn=None, delete_fn=None, nopython=False) Create a new attribute key for communicators .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcopy_fn\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIComm\fP\fI, \fP\fIint\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBdelete_fn\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIComm\fP\fI, \fP\fIint\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fINone\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBnopython\fP (\fIbool\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Delete_attr(keyval) Delete attribute value associated with a key .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Disconnect() Disconnect from a communicator .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Dup(info=None) Duplicate an existing communicator .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIOptional\fP\fI[\fP\fIInfo\fP\fI]\fP) – .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Dup_with_info(info) Duplicate an existing communicator .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIInfo\fP) – .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free a communicator .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Free_keyval(keyval) Free an attribute key for communicators .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Gather(sendbuf, recvbuf, root=0) Gather together values from a group of processes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecB\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Gatherv(sendbuf, recvbuf, root=0) Gather Vector, gather data to one process from all other processes in a group providing different amount of data and displacements at the receiving sides .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecV\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_attr(keyval) Retrieve attribute value by key .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type Optional[Union[int, Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_errhandler() Get the error handler for a communicator .INDENT 7.0 .TP .B Return type Errhandler .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_group() Access the group associated with a communicator .INDENT 7.0 .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_info() Return the hints for a communicator that are currently in use .INDENT 7.0 .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_name() Get the print name for this communicator .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Get_parent() Return the parent intercommunicator for this process .INDENT 7.0 .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_rank() Return the rank of this process in a communicator .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_size() Return the number of processes in a communicator .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_topology() Determine the type of topology (if any) associated with a communicator .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iallgather(sendbuf, recvbuf) Nonblocking Gather to All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iallgatherv(sendbuf, recvbuf) Nonblocking Gather to All Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iallreduce(sendbuf, recvbuf, op=SUM) Nonblocking Reduce to All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ialltoall(sendbuf, recvbuf) Nonblocking All to All Scatter/Gather .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecB\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ialltoallv(sendbuf, recvbuf) Nonblocking All to All Scatter/Gather Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecV\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ialltoallw(sendbuf, recvbuf) Nonblocking Generalized All\-to\-All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecW\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecW\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ibarrier() Nonblocking Barrier .INDENT 7.0 .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ibcast(buf, root=0) Nonblocking Broadcast .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ibsend(buf, dest, tag=0) Nonblocking send in buffered mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Idup() Nonblocking duplicate an existing communicator .INDENT 7.0 .TP .B Return type Tuple[Comm, Request] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Igather(sendbuf, recvbuf, root=0) Nonblocking Gather .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecB\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Igatherv(sendbuf, recvbuf, root=0) Nonblocking Gather Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecV\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Improbe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[Message] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iprobe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Irecv(buf, source=ANY_SOURCE, tag=ANY_TAG) Nonblocking receive .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ireduce(sendbuf, recvbuf, op=SUM, root=0) Nonblocking Reduce to Root .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpec\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ireduce_scatter(sendbuf, recvbuf, recvcounts=None, op=SUM) Nonblocking Reduce\-Scatter (vector version) .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBrecvcounts\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ireduce_scatter_block(sendbuf, recvbuf, op=SUM) Nonblocking Reduce\-Scatter Block (regular, non\-vector version) .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecB\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIBufSpecB\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Irsend(buf, dest, tag=0) Nonblocking send in ready mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Is_inter() Test to see if a comm is an intercommunicator .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Is_intra() Test to see if a comm is an intracommunicator .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iscatter(sendbuf, recvbuf, root=0) Nonblocking Scatter .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecB\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iscatterv(sendbuf, recvbuf, root=0) Nonblocking Scatter Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecV\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Isend(buf, dest, tag=0) Nonblocking send .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Issend(buf, dest, tag=0) Nonblocking send in synchronous mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Join(fd) Create a intercommunicator by joining two processes connected by a socket .INDENT 7.0 .TP .B Parameters \fBfd\fP (\fIint\fP) – .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Mprobe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Message .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Probe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a message .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This function blocks until the message arrives. .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Literal[True] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Recv(buf, source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking receive .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This function blocks until the message is received .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Recv_init(buf, source=ANY_SOURCE, tag=ANY_TAG) Create a persistent request for a receive .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Prequest .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Reduce(sendbuf, recvbuf, op=SUM, root=0) Reduce to Root .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpec\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Reduce_scatter(sendbuf, recvbuf, recvcounts=None, op=SUM) Reduce\-Scatter (vector version) .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBrecvcounts\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Reduce_scatter_block(sendbuf, recvbuf, op=SUM) Reduce\-Scatter Block (regular, non\-vector version) .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpecB\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIBufSpecB\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Rsend(buf, dest, tag=0) Blocking send in ready mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Rsend_init(buf, dest, tag=0) Persistent request for a send in ready mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Scatter(sendbuf, recvbuf, root=0) Scatter data from one process to all other processes in a group .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecB\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Scatterv(sendbuf, recvbuf, root=0) Scatter Vector, scatter data from one process to all other processes in a group providing different amount of data and displacements at the sending side .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIOptional\fP\fI[\fP\fIBufSpecV\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Send(buf, dest, tag=0) Blocking send .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This function may block until the message is received. Whether or not \fI\%Send\fP blocks depends on several factors and is implementation dependent .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Send_init(buf, dest, tag=0) Create a persistent request for a standard send .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Prequest .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Sendrecv(sendbuf, dest, sendtag=0, recvbuf=None, source=ANY_SOURCE, recvtag=ANY_TAG, status=None) Send and receive a message .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This function is guaranteed not to deadlock in situations where pairs of blocking sends and receives may deadlock. .UNINDENT .UNINDENT .sp \fBCAUTION:\fP .INDENT 7.0 .INDENT 3.5 A common mistake when using this function is to mismatch the tags with the source and destination ranks, which can result in deadlock. .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBsendtag\fP (\fIint\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBrecvtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Sendrecv_replace(buf, dest, sendtag=0, source=ANY_SOURCE, recvtag=ANY_TAG, status=None) Send and receive a message .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This function is guaranteed not to deadlock in situations where pairs of blocking sends and receives may deadlock. .UNINDENT .UNINDENT .sp \fBCAUTION:\fP .INDENT 7.0 .INDENT 3.5 A common mistake when using this function is to mismatch the tags with the source and destination ranks, which can result in deadlock. .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBsendtag\fP (\fIint\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBrecvtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_attr(keyval, attrval) Store attribute value associated with a key .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkeyval\fP (\fIint\fP) – .IP \(bu 2 \fBattrval\fP (\fIAny\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_errhandler(errhandler) Set the error handler for a communicator .INDENT 7.0 .TP .B Parameters \fBerrhandler\fP (\fIErrhandler\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_info(info) Set new values for the hints associated with a communicator .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIInfo\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_name(name) Set the print name for this communicator .INDENT 7.0 .TP .B Parameters \fBname\fP (\fIstr\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Split(color=0, key=0) Split communicator by color and key .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcolor\fP (\fIint\fP) – .IP \(bu 2 \fBkey\fP (\fIint\fP) – .UNINDENT .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Split_type(split_type, key=0, info=INFO_NULL) Split communicator by split type .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsplit_type\fP (\fIint\fP) – .IP \(bu 2 \fBkey\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ssend(buf, dest, tag=0) Blocking send in synchronous mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ssend_init(buf, dest, tag=0) Persistent request for a send in synchronous mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B allgather(sendobj) Gather to All .INDENT 7.0 .TP .B Parameters \fBsendobj\fP (\fIAny\fP) – .TP .B Return type List[Any] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B allreduce(sendobj, op=SUM) Reduce to All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBop\fP (\fIUnion\fP\fI[\fP\fIOp\fP\fI, \fP\fICallable\fP\fI[\fP\fI[\fP\fIAny\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B alltoall(sendobj) All to All Scatter/Gather .INDENT 7.0 .TP .B Parameters \fBsendobj\fP (\fISequence\fP\fI[\fP\fIAny\fP\fI]\fP) – .TP .B Return type List[Any] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B barrier() Barrier .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B bcast(obj, root=0) Broadcast .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B bsend(obj, dest, tag=0) Send in buffered mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Comm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B gather(sendobj, root=0) Gather .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Optional[List[Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B ibsend(obj, dest, tag=0) Nonblocking send in buffered mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B improbe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[Message] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B iprobe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B irecv(buf=None, source=ANY_SOURCE, tag=ANY_TAG) Nonblocking receive .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIOptional\fP\fI[\fP\fIBuffer\fP\fI]\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B isend(obj, dest, tag=0) Nonblocking send .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B issend(obj, dest, tag=0) Nonblocking send in synchronous mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B mprobe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Message .UNINDENT .UNINDENT .INDENT 7.0 .TP .B probe(source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Literal[True] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B recv(buf=None, source=ANY_SOURCE, tag=ANY_TAG, status=None) Receive .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIOptional\fP\fI[\fP\fIBuffer\fP\fI]\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B reduce(sendobj, op=SUM, root=0) Reduce to Root .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBop\fP (\fIUnion\fP\fI[\fP\fIOp\fP\fI, \fP\fICallable\fP\fI[\fP\fI[\fP\fIAny\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Optional[Any] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B scatter(sendobj, root=0) Scatter .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fISequence\fP\fI[\fP\fIAny\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B send(obj, dest, tag=0) Send .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B sendrecv(sendobj, dest, sendtag=0, recvbuf=None, source=ANY_SOURCE, recvtag=ANY_TAG, status=None) Send and Receive .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBsendtag\fP (\fIint\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIOptional\fP\fI[\fP\fIBuffer\fP\fI]\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBrecvtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B ssend(obj, dest, tag=0) Send in synchronous mode .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBdest\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B group communicator group .UNINDENT .INDENT 7.0 .TP .B info communicator info .UNINDENT .INDENT 7.0 .TP .B is_inter is intercommunicator .UNINDENT .INDENT 7.0 .TP .B is_intra is intracommunicator .UNINDENT .INDENT 7.0 .TP .B is_topo is a topology communicator .UNINDENT .INDENT 7.0 .TP .B name communicator name .UNINDENT .INDENT 7.0 .TP .B rank rank of this process in communicator .UNINDENT .INDENT 7.0 .TP .B size number of processes in communicator .UNINDENT .INDENT 7.0 .TP .B topology communicator topology type .UNINDENT .UNINDENT .SS mpi4py.MPI.Datatype .INDENT 0.0 .TP .B class mpi4py.MPI.Datatype(datatype=None) Bases: \fBobject\fP .sp Datatype object .INDENT 7.0 .TP .B Parameters \fBdatatype\fP (\fIOptional\fP\fI[\fP\fIDatatype\fP\fI]\fP) – .TP .B Return type Datatype .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, datatype=None) .INDENT 7.0 .TP .B Parameters \fBdatatype\fP (\fIOptional\fP\fI[\fP\fIDatatype\fP\fI]\fP) – .TP .B Return type Datatype .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Commit\fP() T} T{ Commit the datatype T} _ T{ \fI\%Create_contiguous\fP(count) T} T{ Create a contiguous datatype T} _ T{ \fI\%Create_darray\fP(size, rank, gsizes, distribs, ...) T} T{ Create a datatype representing an HPF\-like distributed array on Cartesian process grids T} _ T{ \fI\%Create_f90_complex\fP(p, r) T} T{ Return a bounded complex datatype T} _ T{ \fI\%Create_f90_integer\fP(r) T} T{ Return a bounded integer datatype T} _ T{ \fI\%Create_f90_real\fP(p, r) T} T{ Return a bounded real datatype T} _ T{ \fI\%Create_hindexed\fP(blocklengths, displacements) T} T{ Create an indexed datatype with displacements in bytes T} _ T{ \fI\%Create_hindexed_block\fP(blocklength, displacements) T} T{ Create an indexed datatype with constant\-sized blocks and displacements in bytes T} _ T{ \fI\%Create_hvector\fP(count, blocklength, stride) T} T{ Create a vector (strided) datatype T} _ T{ \fI\%Create_indexed\fP(blocklengths, displacements) T} T{ Create an indexed datatype T} _ T{ \fI\%Create_indexed_block\fP(blocklength, displacements) T} T{ Create an indexed datatype with constant\-sized blocks T} _ T{ \fI\%Create_keyval\fP([copy_fn, delete_fn, nopython]) T} T{ Create a new attribute key for datatypes T} _ T{ \fI\%Create_resized\fP(lb, extent) T} T{ Create a datatype with a new lower bound and extent T} _ T{ \fI\%Create_struct\fP(blocklengths, displacements, ...) T} T{ Create an datatype from a general set of block sizes, displacements and datatypes T} _ T{ \fI\%Create_subarray\fP(sizes, subsizes, starts[, order]) T} T{ Create a datatype for a subarray of a regular, multidimensional array T} _ T{ \fI\%Create_vector\fP(count, blocklength, stride) T} T{ Create a vector (strided) datatype T} _ T{ \fI\%Delete_attr\fP(keyval) T} T{ Delete attribute value associated with a key T} _ T{ \fI\%Dup\fP() T} T{ Duplicate a datatype T} _ T{ \fI\%Free\fP() T} T{ Free the datatype T} _ T{ \fI\%Free_keyval\fP(keyval) T} T{ Free an attribute key for datatypes T} _ T{ \fI\%Get_attr\fP(keyval) T} T{ Retrieve attribute value by key T} _ T{ \fI\%Get_contents\fP() T} T{ Retrieve the actual arguments used in the call that created a datatype T} _ T{ \fI\%Get_envelope\fP() T} T{ Return information on the number and type of input arguments used in the call that created a datatype T} _ T{ \fI\%Get_extent\fP() T} T{ Return lower bound and extent of datatype T} _ T{ \fI\%Get_name\fP() T} T{ Get the print name for this datatype T} _ T{ \fI\%Get_size\fP() T} T{ Return the number of bytes occupied by entries in the datatype T} _ T{ \fI\%Get_true_extent\fP() T} T{ Return the true lower bound and extent of a datatype T} _ T{ \fI\%Match_size\fP(typeclass, size) T} T{ Find a datatype matching a specified size in bytes T} _ T{ \fI\%Pack\fP(inbuf, outbuf, position, comm) T} T{ Pack into contiguous memory according to datatype. T} _ T{ \fI\%Pack_external\fP(datarep, inbuf, outbuf, position) T} T{ Pack into contiguous memory according to datatype, using a portable data representation (\fBexternal32\fP). T} _ T{ \fI\%Pack_external_size\fP(datarep, count) T} T{ Return the upper bound on the amount of space (in bytes) needed to pack a message according to datatype, using a portable data representation (\fBexternal32\fP). T} _ T{ \fI\%Pack_size\fP(count, comm) T} T{ Return the upper bound on the amount of space (in bytes) needed to pack a message according to datatype. T} _ T{ \fI\%Set_attr\fP(keyval, attrval) T} T{ Store attribute value associated with a key T} _ T{ \fI\%Set_name\fP(name) T} T{ Set the print name for this datatype T} _ T{ \fI\%Unpack\fP(inbuf, position, outbuf, comm) T} T{ Unpack from contiguous memory according to datatype. T} _ T{ \fI\%Unpack_external\fP(datarep, inbuf, position, outbuf) T} T{ Unpack from contiguous memory according to datatype, using a portable data representation (\fBexternal32\fP). T} _ T{ \fI\%decode\fP() T} T{ Convenience method for decoding a datatype T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%combiner\fP T} T{ datatype combiner T} _ T{ \fI\%contents\fP T} T{ datatype contents T} _ T{ \fI\%envelope\fP T} T{ datatype envelope T} _ T{ \fI\%extent\fP T} T{ T} _ T{ \fI\%is_named\fP T} T{ is a named datatype T} _ T{ \fI\%is_predefined\fP T} T{ is a predefined datatype T} _ T{ \fI\%lb\fP T} T{ lower bound T} _ T{ \fI\%name\fP T} T{ datatype name T} _ T{ \fI\%size\fP T} T{ T} _ T{ \fI\%true_extent\fP T} T{ true extent T} _ T{ \fI\%true_lb\fP T} T{ true lower bound T} _ T{ \fI\%true_ub\fP T} T{ true upper bound T} _ T{ \fI\%ub\fP T} T{ upper bound T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Commit() Commit the datatype .INDENT 7.0 .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_contiguous(count) Create a contiguous datatype .INDENT 7.0 .TP .B Parameters \fBcount\fP (\fIint\fP) – .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_darray(size, rank, gsizes, distribs, dargs, psizes, order=ORDER_C) Create a datatype representing an HPF\-like distributed array on Cartesian process grids .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsize\fP (\fIint\fP) – .IP \(bu 2 \fBrank\fP (\fIint\fP) – .IP \(bu 2 \fBgsizes\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdistribs\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdargs\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBpsizes\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBorder\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_f90_complex(p, r) Return a bounded complex datatype .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBp\fP (\fIint\fP) – .IP \(bu 2 \fBr\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_f90_integer(r) Return a bounded integer datatype .INDENT 7.0 .TP .B Parameters \fBr\fP (\fIint\fP) – .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_f90_real(p, r) Return a bounded real datatype .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBp\fP (\fIint\fP) – .IP \(bu 2 \fBr\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_hindexed(blocklengths, displacements) Create an indexed datatype with displacements in bytes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBblocklengths\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdisplacements\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_hindexed_block(blocklength, displacements) Create an indexed datatype with constant\-sized blocks and displacements in bytes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBblocklength\fP (\fIint\fP) – .IP \(bu 2 \fBdisplacements\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_hvector(count, blocklength, stride) Create a vector (strided) datatype .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcount\fP (\fIint\fP) – .IP \(bu 2 \fBblocklength\fP (\fIint\fP) – .IP \(bu 2 \fBstride\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_indexed(blocklengths, displacements) Create an indexed datatype .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBblocklengths\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdisplacements\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_indexed_block(blocklength, displacements) Create an indexed datatype with constant\-sized blocks .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBblocklength\fP (\fIint\fP) – .IP \(bu 2 \fBdisplacements\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_keyval(copy_fn=None, delete_fn=None, nopython=False) Create a new attribute key for datatypes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcopy_fn\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIDatatype\fP\fI, \fP\fIint\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBdelete_fn\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIDatatype\fP\fI, \fP\fIint\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fINone\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBnopython\fP (\fIbool\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_resized(lb, extent) Create a datatype with a new lower bound and extent .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBlb\fP (\fIint\fP) – .IP \(bu 2 \fBextent\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_struct(blocklengths, displacements, datatypes) Create an datatype from a general set of block sizes, displacements and datatypes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBblocklengths\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdisplacements\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdatatypes\fP (\fISequence\fP\fI[\fP\fIDatatype\fP\fI]\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_subarray(sizes, subsizes, starts, order=ORDER_C) Create a datatype for a subarray of a regular, multidimensional array .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsizes\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBsubsizes\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBstarts\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBorder\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_vector(count, blocklength, stride) Create a vector (strided) datatype .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcount\fP (\fIint\fP) – .IP \(bu 2 \fBblocklength\fP (\fIint\fP) – .IP \(bu 2 \fBstride\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Delete_attr(keyval) Delete attribute value associated with a key .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Dup() Duplicate a datatype .INDENT 7.0 .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free the datatype .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Free_keyval(keyval) Free an attribute key for datatypes .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_attr(keyval) Retrieve attribute value by key .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type Optional[Union[int, Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_contents() Retrieve the actual arguments used in the call that created a datatype .INDENT 7.0 .TP .B Return type Tuple[List[int], List[int], List[Datatype]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_envelope() Return information on the number and type of input arguments used in the call that created a datatype .INDENT 7.0 .TP .B Return type Tuple[int, int, int, int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_extent() Return lower bound and extent of datatype .INDENT 7.0 .TP .B Return type Tuple[int, int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_name() Get the print name for this datatype .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_size() Return the number of bytes occupied by entries in the datatype .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_true_extent() Return the true lower bound and extent of a datatype .INDENT 7.0 .TP .B Return type Tuple[int, int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Match_size(typeclass, size) Find a datatype matching a specified size in bytes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBtypeclass\fP (\fIint\fP) – .IP \(bu 2 \fBsize\fP (\fIint\fP) – .UNINDENT .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Pack(inbuf, outbuf, position, comm) Pack into contiguous memory according to datatype. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBinbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBoutbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBposition\fP (\fIint\fP) – .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Pack_external(datarep, inbuf, outbuf, position) Pack into contiguous memory according to datatype, using a portable data representation (\fBexternal32\fP). .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdatarep\fP (\fIstr\fP) – .IP \(bu 2 \fBinbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBoutbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBposition\fP (\fIint\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Pack_external_size(datarep, count) Return the upper bound on the amount of space (in bytes) needed to pack a message according to datatype, using a portable data representation (\fBexternal32\fP). .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdatarep\fP (\fIstr\fP) – .IP \(bu 2 \fBcount\fP (\fIint\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Pack_size(count, comm) Return the upper bound on the amount of space (in bytes) needed to pack a message according to datatype. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcount\fP (\fIint\fP) – .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_attr(keyval, attrval) Store attribute value associated with a key .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkeyval\fP (\fIint\fP) – .IP \(bu 2 \fBattrval\fP (\fIAny\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_name(name) Set the print name for this datatype .INDENT 7.0 .TP .B Parameters \fBname\fP (\fIstr\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Unpack(inbuf, position, outbuf, comm) Unpack from contiguous memory according to datatype. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBinbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBposition\fP (\fIint\fP) – .IP \(bu 2 \fBoutbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Unpack_external(datarep, inbuf, position, outbuf) Unpack from contiguous memory according to datatype, using a portable data representation (\fBexternal32\fP). .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdatarep\fP (\fIstr\fP) – .IP \(bu 2 \fBinbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBposition\fP (\fIint\fP) – .IP \(bu 2 \fBoutbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B decode() Convenience method for decoding a datatype .INDENT 7.0 .TP .B Return type Tuple[Datatype, str, Dict[str, Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Datatype .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B combiner datatype combiner .UNINDENT .INDENT 7.0 .TP .B contents datatype contents .UNINDENT .INDENT 7.0 .TP .B envelope datatype envelope .UNINDENT .INDENT 7.0 .TP .B extent .UNINDENT .INDENT 7.0 .TP .B is_named is a named datatype .UNINDENT .INDENT 7.0 .TP .B is_predefined is a predefined datatype .UNINDENT .INDENT 7.0 .TP .B lb lower bound .UNINDENT .INDENT 7.0 .TP .B name datatype name .UNINDENT .INDENT 7.0 .TP .B size .UNINDENT .INDENT 7.0 .TP .B true_extent true extent .UNINDENT .INDENT 7.0 .TP .B true_lb true lower bound .UNINDENT .INDENT 7.0 .TP .B true_ub true upper bound .UNINDENT .INDENT 7.0 .TP .B ub upper bound .UNINDENT .UNINDENT .SS mpi4py.MPI.Distgraphcomm .INDENT 0.0 .TP .B class mpi4py.MPI.Distgraphcomm(comm=None) Bases: \fBmpi4py.MPI.Topocomm\fP .sp Distributed graph topology intracommunicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIDistgraphcomm\fP\fI]\fP) – .TP .B Return type Distgraphcomm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIDistgraphcomm\fP\fI]\fP) – .TP .B Return type Distgraphcomm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Get_dist_neighbors\fP() T} T{ Return adjacency information for a distributed graph topology T} _ T{ \fI\%Get_dist_neighbors_count\fP() T} T{ Return adjacency information for a distributed graph topology T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Get_dist_neighbors() Return adjacency information for a distributed graph topology .INDENT 7.0 .TP .B Return type Tuple[List[int], List[int], Optional[Tuple[List[int], List[int]]]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_dist_neighbors_count() Return adjacency information for a distributed graph topology .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Errhandler .INDENT 0.0 .TP .B class mpi4py.MPI.Errhandler(errhandler=None) Bases: \fBobject\fP .sp Error handler .INDENT 7.0 .TP .B Parameters \fBerrhandler\fP (\fIOptional\fP\fI[\fP\fIErrhandler\fP\fI]\fP) – .TP .B Return type Errhandler .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, errhandler=None) .INDENT 7.0 .TP .B Parameters \fBerrhandler\fP (\fIOptional\fP\fI[\fP\fIErrhandler\fP\fI]\fP) – .TP .B Return type Errhandler .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Free\fP() T} T{ Free an error handler T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Free() Free an error handler .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Errhandler .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.File .INDENT 0.0 .TP .B class mpi4py.MPI.File(file=None) Bases: \fBobject\fP .sp File handle .INDENT 7.0 .TP .B Parameters \fBfile\fP (\fIOptional\fP\fI[\fP\fIFile\fP\fI]\fP) – .TP .B Return type File .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, file=None) .INDENT 7.0 .TP .B Parameters \fBfile\fP (\fIOptional\fP\fI[\fP\fIFile\fP\fI]\fP) – .TP .B Return type File .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Call_errhandler\fP(errorcode) T} T{ Call the error handler installed on a file T} _ T{ \fI\%Close\fP() T} T{ Close a file T} _ T{ \fI\%Delete\fP(filename[, info]) T} T{ Delete a file T} _ T{ \fI\%Get_amode\fP() T} T{ Return the file access mode T} _ T{ \fI\%Get_atomicity\fP() T} T{ Return the atomicity mode T} _ T{ \fI\%Get_byte_offset\fP(offset) T} T{ Return the absolute byte position in the file corresponding to \(aqoffset\(aq etypes relative to the current view T} _ T{ \fI\%Get_errhandler\fP() T} T{ Get the error handler for a file T} _ T{ \fI\%Get_group\fP() T} T{ Return the group of processes that opened the file T} _ T{ \fI\%Get_info\fP() T} T{ Return the hints for a file that that are currently in use T} _ T{ \fI\%Get_position\fP() T} T{ Return the current position of the individual file pointer in etype units relative to the current view T} _ T{ \fI\%Get_position_shared\fP() T} T{ Return the current position of the shared file pointer in etype units relative to the current view T} _ T{ \fI\%Get_size\fP() T} T{ Return the file size T} _ T{ \fI\%Get_type_extent\fP(datatype) T} T{ Return the extent of datatype in the file T} _ T{ \fI\%Get_view\fP() T} T{ Return the file view T} _ T{ \fI\%Iread\fP(buf) T} T{ Nonblocking read using individual file pointer T} _ T{ \fI\%Iread_all\fP(buf) T} T{ Nonblocking collective read using individual file pointer T} _ T{ \fI\%Iread_at\fP(offset, buf) T} T{ Nonblocking read using explicit offset T} _ T{ \fI\%Iread_at_all\fP(offset, buf) T} T{ Nonblocking collective read using explicit offset T} _ T{ \fI\%Iread_shared\fP(buf) T} T{ Nonblocking read using shared file pointer T} _ T{ \fI\%Iwrite\fP(buf) T} T{ Nonblocking write using individual file pointer T} _ T{ \fI\%Iwrite_all\fP(buf) T} T{ Nonblocking collective write using individual file pointer T} _ T{ \fI\%Iwrite_at\fP(offset, buf) T} T{ Nonblocking write using explicit offset T} _ T{ \fI\%Iwrite_at_all\fP(offset, buf) T} T{ Nonblocking collective write using explicit offset T} _ T{ \fI\%Iwrite_shared\fP(buf) T} T{ Nonblocking write using shared file pointer T} _ T{ \fI\%Open\fP(comm, filename[, amode, info]) T} T{ Open a file T} _ T{ \fI\%Preallocate\fP(size) T} T{ Preallocate storage space for a file T} _ T{ \fI\%Read\fP(buf[, status]) T} T{ Read using individual file pointer T} _ T{ \fI\%Read_all\fP(buf[, status]) T} T{ Collective read using individual file pointer T} _ T{ \fI\%Read_all_begin\fP(buf) T} T{ Start a split collective read using individual file pointer T} _ T{ \fI\%Read_all_end\fP(buf[, status]) T} T{ Complete a split collective read using individual file pointer T} _ T{ \fI\%Read_at\fP(offset, buf[, status]) T} T{ Read using explicit offset T} _ T{ \fI\%Read_at_all\fP(offset, buf[, status]) T} T{ Collective read using explicit offset T} _ T{ \fI\%Read_at_all_begin\fP(offset, buf) T} T{ Start a split collective read using explict offset T} _ T{ \fI\%Read_at_all_end\fP(buf[, status]) T} T{ Complete a split collective read using explict offset T} _ T{ \fI\%Read_ordered\fP(buf[, status]) T} T{ Collective read using shared file pointer T} _ T{ \fI\%Read_ordered_begin\fP(buf) T} T{ Start a split collective read using shared file pointer T} _ T{ \fI\%Read_ordered_end\fP(buf[, status]) T} T{ Complete a split collective read using shared file pointer T} _ T{ \fI\%Read_shared\fP(buf[, status]) T} T{ Read using shared file pointer T} _ T{ \fI\%Seek\fP(offset[, whence]) T} T{ Update the individual file pointer T} _ T{ \fI\%Seek_shared\fP(offset[, whence]) T} T{ Update the shared file pointer T} _ T{ \fI\%Set_atomicity\fP(flag) T} T{ Set the atomicity mode T} _ T{ \fI\%Set_errhandler\fP(errhandler) T} T{ Set the error handler for a file T} _ T{ \fI\%Set_info\fP(info) T} T{ Set new values for the hints associated with a file T} _ T{ \fI\%Set_size\fP(size) T} T{ Sets the file size T} _ T{ \fI\%Set_view\fP([disp, etype, filetype, datarep, info]) T} T{ Set the file view T} _ T{ \fI\%Sync\fP() T} T{ Causes all previous writes to be transferred to the storage device T} _ T{ \fI\%Write\fP(buf[, status]) T} T{ Write using individual file pointer T} _ T{ \fI\%Write_all\fP(buf[, status]) T} T{ Collective write using individual file pointer T} _ T{ \fI\%Write_all_begin\fP(buf) T} T{ Start a split collective write using individual file pointer T} _ T{ \fI\%Write_all_end\fP(buf[, status]) T} T{ Complete a split collective write using individual file pointer T} _ T{ \fI\%Write_at\fP(offset, buf[, status]) T} T{ Write using explicit offset T} _ T{ \fI\%Write_at_all\fP(offset, buf[, status]) T} T{ Collective write using explicit offset T} _ T{ \fI\%Write_at_all_begin\fP(offset, buf) T} T{ Start a split collective write using explict offset T} _ T{ \fI\%Write_at_all_end\fP(buf[, status]) T} T{ Complete a split collective write using explict offset T} _ T{ \fI\%Write_ordered\fP(buf[, status]) T} T{ Collective write using shared file pointer T} _ T{ \fI\%Write_ordered_begin\fP(buf) T} T{ Start a split collective write using shared file pointer T} _ T{ \fI\%Write_ordered_end\fP(buf[, status]) T} T{ Complete a split collective write using shared file pointer T} _ T{ \fI\%Write_shared\fP(buf[, status]) T} T{ Write using shared file pointer T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%amode\fP T} T{ file access mode T} _ T{ \fI\%atomicity\fP T} T{ T} _ T{ \fI\%group\fP T} T{ file group T} _ T{ \fI\%info\fP T} T{ file info T} _ T{ \fI\%size\fP T} T{ file size T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Call_errhandler(errorcode) Call the error handler installed on a file .INDENT 7.0 .TP .B Parameters \fBerrorcode\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Close() Close a file .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Delete(filename, info=INFO_NULL) Delete a file .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBfilename\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_amode() Return the file access mode .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_atomicity() Return the atomicity mode .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_byte_offset(offset) Return the absolute byte position in the file corresponding to ‘offset’ etypes relative to the current view .INDENT 7.0 .TP .B Parameters \fBoffset\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_errhandler() Get the error handler for a file .INDENT 7.0 .TP .B Return type Errhandler .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_group() Return the group of processes that opened the file .INDENT 7.0 .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_info() Return the hints for a file that that are currently in use .INDENT 7.0 .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_position() Return the current position of the individual file pointer in etype units relative to the current view .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_position_shared() Return the current position of the shared file pointer in etype units relative to the current view .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_size() Return the file size .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_type_extent(datatype) Return the extent of datatype in the file .INDENT 7.0 .TP .B Parameters \fBdatatype\fP (\fIDatatype\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_view() Return the file view .INDENT 7.0 .TP .B Return type Tuple[int, Datatype, Datatype, str] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iread(buf) Nonblocking read using individual file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iread_all(buf) Nonblocking collective read using individual file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iread_at(offset, buf) Nonblocking read using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iread_at_all(offset, buf) Nonblocking collective read using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iread_shared(buf) Nonblocking read using shared file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iwrite(buf) Nonblocking write using individual file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iwrite_all(buf) Nonblocking collective write using individual file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iwrite_at(offset, buf) Nonblocking write using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iwrite_at_all(offset, buf) Nonblocking collective write using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iwrite_shared(buf) Nonblocking write using shared file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Open(comm, filename, amode=MODE_RDONLY, info=INFO_NULL) Open a file .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcomm\fP (\fIIntracomm\fP) – .IP \(bu 2 \fBfilename\fP (\fIstr\fP) – .IP \(bu 2 \fBamode\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type File .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Preallocate(size) Preallocate storage space for a file .INDENT 7.0 .TP .B Parameters \fBsize\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read(buf, status=None) Read using individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_all(buf, status=None) Collective read using individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_all_begin(buf) Start a split collective read using individual file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_all_end(buf, status=None) Complete a split collective read using individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_at(offset, buf, status=None) Read using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_at_all(offset, buf, status=None) Collective read using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_at_all_begin(offset, buf) Start a split collective read using explict offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_at_all_end(buf, status=None) Complete a split collective read using explict offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_ordered(buf, status=None) Collective read using shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_ordered_begin(buf) Start a split collective read using shared file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_ordered_end(buf, status=None) Complete a split collective read using shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Read_shared(buf, status=None) Read using shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Seek(offset, whence=SEEK_SET) Update the individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBwhence\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Seek_shared(offset, whence=SEEK_SET) Update the shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBwhence\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_atomicity(flag) Set the atomicity mode .INDENT 7.0 .TP .B Parameters \fBflag\fP (\fIbool\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_errhandler(errhandler) Set the error handler for a file .INDENT 7.0 .TP .B Parameters \fBerrhandler\fP (\fIErrhandler\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_info(info) Set new values for the hints associated with a file .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIInfo\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_size(size) Sets the file size .INDENT 7.0 .TP .B Parameters \fBsize\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_view(disp=0, etype=BYTE, filetype=None, datarep=\(aqnative\(aq, info=INFO_NULL) Set the file view .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdisp\fP (\fIint\fP) – .IP \(bu 2 \fBetype\fP (\fIDatatype\fP) – .IP \(bu 2 \fBfiletype\fP (\fIOptional\fP\fI[\fP\fIDatatype\fP\fI]\fP) – .IP \(bu 2 \fBdatarep\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Sync() Causes all previous writes to be transferred to the storage device .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write(buf, status=None) Write using individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_all(buf, status=None) Collective write using individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_all_begin(buf) Start a split collective write using individual file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_all_end(buf, status=None) Complete a split collective write using individual file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_at(offset, buf, status=None) Write using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_at_all(offset, buf, status=None) Collective write using explicit offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_at_all_begin(offset, buf) Start a split collective write using explict offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBoffset\fP (\fIint\fP) – .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_at_all_end(buf, status=None) Complete a split collective write using explict offset .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_ordered(buf, status=None) Collective write using shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_ordered_begin(buf) Start a split collective write using shared file pointer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_ordered_end(buf, status=None) Complete a split collective write using shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Write_shared(buf, status=None) Write using shared file pointer .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type File .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B amode file access mode .UNINDENT .INDENT 7.0 .TP .B atomicity .UNINDENT .INDENT 7.0 .TP .B group file group .UNINDENT .INDENT 7.0 .TP .B info file info .UNINDENT .INDENT 7.0 .TP .B size file size .UNINDENT .UNINDENT .SS mpi4py.MPI.Graphcomm .INDENT 0.0 .TP .B class mpi4py.MPI.Graphcomm(comm=None) Bases: \fBmpi4py.MPI.Topocomm\fP .sp General graph topology intracommunicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIGraphcomm\fP\fI]\fP) – .TP .B Return type Graphcomm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIGraphcomm\fP\fI]\fP) – .TP .B Return type Graphcomm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Get_dims\fP() T} T{ Return the number of nodes and edges T} _ T{ \fI\%Get_neighbors\fP(rank) T} T{ Return list of neighbors of a process T} _ T{ \fI\%Get_neighbors_count\fP(rank) T} T{ Return number of neighbors of a process T} _ T{ \fI\%Get_topo\fP() T} T{ Return index and edges T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%dims\fP T} T{ number of nodes and edges T} _ T{ \fI\%edges\fP T} T{ T} _ T{ \fI\%index\fP T} T{ T} _ T{ \fI\%nedges\fP T} T{ number of edges T} _ T{ \fI\%neighbors\fP T} T{ T} _ T{ \fI\%nneighbors\fP T} T{ number of neighbors T} _ T{ \fI\%nnodes\fP T} T{ number of nodes T} _ T{ \fI\%topo\fP T} T{ topology information T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Get_dims() Return the number of nodes and edges .INDENT 7.0 .TP .B Return type Tuple[int, int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_neighbors(rank) Return list of neighbors of a process .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type List[int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_neighbors_count(rank) Return number of neighbors of a process .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_topo() Return index and edges .INDENT 7.0 .TP .B Return type Tuple[List[int], List[int]] .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B dims number of nodes and edges .UNINDENT .INDENT 7.0 .TP .B edges .UNINDENT .INDENT 7.0 .TP .B index .UNINDENT .INDENT 7.0 .TP .B nedges number of edges .UNINDENT .INDENT 7.0 .TP .B neighbors .UNINDENT .INDENT 7.0 .TP .B nneighbors number of neighbors .UNINDENT .INDENT 7.0 .TP .B nnodes number of nodes .UNINDENT .INDENT 7.0 .TP .B topo topology information .UNINDENT .UNINDENT .SS mpi4py.MPI.Grequest .INDENT 0.0 .TP .B class mpi4py.MPI.Grequest(request=None) Bases: \fBmpi4py.MPI.Request\fP .sp Generalized request handle .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIOptional\fP\fI[\fP\fIGrequest\fP\fI]\fP) – .TP .B Return type Grequest .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, request=None) .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIOptional\fP\fI[\fP\fIGrequest\fP\fI]\fP) – .TP .B Return type Grequest .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Complete\fP() T} T{ Notify that a user\-defined request is complete T} _ T{ \fI\%Start\fP(query_fn, free_fn, cancel_fn[, args, ...]) T} T{ Create and return a user\-defined request T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Complete() Notify that a user\-defined request is complete .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Start(query_fn, free_fn, cancel_fn, args=None, kargs=None) Create and return a user\-defined request .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBquery_fn\fP (\fICallable\fP\fI[\fP\fI\&...\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBfree_fn\fP (\fICallable\fP\fI[\fP\fI\&...\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBcancel_fn\fP (\fICallable\fP\fI[\fP\fI\&...\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBargs\fP (\fIOptional\fP\fI[\fP\fITuple\fP\fI[\fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBkargs\fP (\fIOptional\fP\fI[\fP\fIDict\fP\fI[\fP\fIstr\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Grequest .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Group .INDENT 0.0 .TP .B class mpi4py.MPI.Group(group=None) Bases: \fBobject\fP .sp Group of processes .INDENT 7.0 .TP .B Parameters \fBgroup\fP (\fIOptional\fP\fI[\fP\fIGroup\fP\fI]\fP) – .TP .B Return type Group .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, group=None) .INDENT 7.0 .TP .B Parameters \fBgroup\fP (\fIOptional\fP\fI[\fP\fIGroup\fP\fI]\fP) – .TP .B Return type Group .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Compare\fP(group1, group2) T} T{ Compare two groups T} _ T{ \fI\%Difference\fP(group1, group2) T} T{ Produce a group from the difference of two existing groups T} _ T{ \fI\%Dup\fP() T} T{ Duplicate a group T} _ T{ \fI\%Excl\fP(ranks) T} T{ Produce a group by reordering an existing group and taking only unlisted members T} _ T{ \fI\%Free\fP() T} T{ Free a group T} _ T{ \fI\%Get_rank\fP() T} T{ Return the rank of this process in a group T} _ T{ \fI\%Get_size\fP() T} T{ Return the size of a group T} _ T{ \fI\%Incl\fP(ranks) T} T{ Produce a group by reordering an existing group and taking only listed members T} _ T{ \fI\%Intersection\fP(group1, group2) T} T{ Produce a group as the intersection of two existing groups T} _ T{ \fI\%Range_excl\fP(ranks) T} T{ Create a new group by excluding ranges of processes from an existing group T} _ T{ \fI\%Range_incl\fP(ranks) T} T{ Create a new group from ranges of of ranks in an existing group T} _ T{ \fI\%Translate_ranks\fP(group1, ranks1[, group2]) T} T{ Translate the ranks of processes in one group to those in another group T} _ T{ \fI\%Union\fP(group1, group2) T} T{ Produce a group by combining two existing groups T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%rank\fP T} T{ rank of this process in group T} _ T{ \fI\%size\fP T} T{ number of processes in group T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B classmethod Compare(group1, group2) Compare two groups .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup1\fP (\fIGroup\fP) – .IP \(bu 2 \fBgroup2\fP (\fIGroup\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Difference(group1, group2) Produce a group from the difference of two existing groups .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup1\fP (\fIGroup\fP) – .IP \(bu 2 \fBgroup2\fP (\fIGroup\fP) – .UNINDENT .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Dup() Duplicate a group .INDENT 7.0 .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Excl(ranks) Produce a group by reordering an existing group and taking only unlisted members .INDENT 7.0 .TP .B Parameters \fBranks\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free a group .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_rank() Return the rank of this process in a group .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_size() Return the size of a group .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Incl(ranks) Produce a group by reordering an existing group and taking only listed members .INDENT 7.0 .TP .B Parameters \fBranks\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Intersection(group1, group2) Produce a group as the intersection of two existing groups .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup1\fP (\fIGroup\fP) – .IP \(bu 2 \fBgroup2\fP (\fIGroup\fP) – .UNINDENT .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Range_excl(ranks) Create a new group by excluding ranges of processes from an existing group .INDENT 7.0 .TP .B Parameters \fBranks\fP (\fISequence\fP\fI[\fP\fITuple\fP\fI[\fP\fIint\fP\fI, \fP\fIint\fP\fI, \fP\fIint\fP\fI]\fP\fI]\fP) – .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Range_incl(ranks) Create a new group from ranges of of ranks in an existing group .INDENT 7.0 .TP .B Parameters \fBranks\fP (\fISequence\fP\fI[\fP\fITuple\fP\fI[\fP\fIint\fP\fI, \fP\fIint\fP\fI, \fP\fIint\fP\fI]\fP\fI]\fP) – .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Translate_ranks(group1, ranks1, group2=None) Translate the ranks of processes in one group to those in another group .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup1\fP (\fIGroup\fP) – .IP \(bu 2 \fBranks1\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBgroup2\fP (\fIOptional\fP\fI[\fP\fIGroup\fP\fI]\fP) – .UNINDENT .TP .B Return type List[int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Union(group1, group2) Produce a group by combining two existing groups .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup1\fP (\fIGroup\fP) – .IP \(bu 2 \fBgroup2\fP (\fIGroup\fP) – .UNINDENT .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B rank rank of this process in group .UNINDENT .INDENT 7.0 .TP .B size number of processes in group .UNINDENT .UNINDENT .SS mpi4py.MPI.Info .INDENT 0.0 .TP .B class mpi4py.MPI.Info(info=None) Bases: \fBobject\fP .sp Info object .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIOptional\fP\fI[\fP\fIInfo\fP\fI]\fP) – .TP .B Return type Info .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, info=None) .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIOptional\fP\fI[\fP\fIInfo\fP\fI]\fP) – .TP .B Return type Info .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Create\fP() T} T{ Create a new, empty info object T} _ T{ \fI\%Delete\fP(key) T} T{ Remove a (key, value) pair from info T} _ T{ \fI\%Dup\fP() T} T{ Duplicate an existing info object, creating a new object, with the same (key, value) pairs and the same ordering of keys T} _ T{ \fI\%Free\fP() T} T{ Free a info object T} _ T{ \fI\%Get\fP(key[, maxlen]) T} T{ Retrieve the value associated with a key T} _ T{ \fI\%Get_nkeys\fP() T} T{ Return the number of currently defined keys in info T} _ T{ \fI\%Get_nthkey\fP(n) T} T{ Return the nth defined key in info. T} _ T{ \fI\%Set\fP(key, value) T} T{ Add the (key, value) pair to info, and overrides the value if a value for the same key was previously set T} _ T{ \fI\%clear\fP() T} T{ info clear T} _ T{ \fI\%copy\fP() T} T{ info copy T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%get\fP(key[, default]) T} T{ info get T} _ T{ \fI\%items\fP() T} T{ info items T} _ T{ \fI\%keys\fP() T} T{ info keys T} _ T{ \fI\%pop\fP(key, *default) T} T{ info pop T} _ T{ \fI\%popitem\fP() T} T{ info popitem T} _ T{ \fI\%py2f\fP() T} T{ T} _ T{ \fI\%update\fP([other]) T} T{ info update T} _ T{ \fI\%values\fP() T} T{ info values T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B classmethod Create() Create a new, empty info object .INDENT 7.0 .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Delete(key) Remove a (key, value) pair from info .INDENT 7.0 .TP .B Parameters \fBkey\fP (\fIstr\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Dup() Duplicate an existing info object, creating a new object, with the same (key, value) pairs and the same ordering of keys .INDENT 7.0 .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free a info object .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get(key, maxlen=\- 1) Retrieve the value associated with a key .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkey\fP (\fIstr\fP) – .IP \(bu 2 \fBmaxlen\fP (\fIint\fP) – .UNINDENT .TP .B Return type Optional[str] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_nkeys() Return the number of currently defined keys in info .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_nthkey(n) Return the nth defined key in info. Keys are numbered in the range [0, N) where N is the value returned by \fI\%Info.Get_nkeys()\fP .INDENT 7.0 .TP .B Parameters \fBn\fP (\fIint\fP) – .TP .B Return type str .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set(key, value) Add the (key, value) pair to info, and overrides the value if a value for the same key was previously set .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkey\fP (\fIstr\fP) – .IP \(bu 2 \fBvalue\fP (\fIstr\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B clear() info clear .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B copy() info copy .INDENT 7.0 .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B get(key, default=None) info get .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkey\fP (\fIstr\fP) – .IP \(bu 2 \fBdefault\fP (\fIOptional\fP\fI[\fP\fIstr\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[str] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B items() info items .INDENT 7.0 .TP .B Return type List[Tuple[str, str]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B keys() info keys .INDENT 7.0 .TP .B Return type List[str] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B pop(key, *default) info pop .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkey\fP (\fIstr\fP) – .IP \(bu 2 \fBdefault\fP (\fIstr\fP) – .UNINDENT .TP .B Return type str .UNINDENT .UNINDENT .INDENT 7.0 .TP .B popitem() info popitem .INDENT 7.0 .TP .B Return type Tuple[str, str] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B update(other=(), **kwds) info update .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBother\fP (\fIUnion\fP\fI[\fP\fIInfo\fP\fI, \fP\fIMapping\fP\fI[\fP\fIstr\fP\fI, \fP\fIstr\fP\fI]\fP\fI, \fP\fIIterable\fP\fI[\fP\fITuple\fP\fI[\fP\fIstr\fP\fI, \fP\fIstr\fP\fI]\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBkwds\fP (\fIstr\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B values() info values .INDENT 7.0 .TP .B Return type List[str] .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Intercomm .INDENT 0.0 .TP .B class mpi4py.MPI.Intercomm(comm=None) Bases: \fBmpi4py.MPI.Comm\fP .sp Intercommunicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIIntercomm\fP\fI]\fP) – .TP .B Return type Intercomm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIIntercomm\fP\fI]\fP) – .TP .B Return type Intercomm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Get_remote_group\fP() T} T{ Access the remote group associated with the inter\-communicator T} _ T{ \fI\%Get_remote_size\fP() T} T{ Intercommunicator remote size T} _ T{ \fI\%Merge\fP([high]) T} T{ Merge intercommunicator T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%remote_group\fP T} T{ remote group T} _ T{ \fI\%remote_size\fP T} T{ number of remote processes T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Get_remote_group() Access the remote group associated with the inter\-communicator .INDENT 7.0 .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_remote_size() Intercommunicator remote size .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Merge(high=False) Merge intercommunicator .INDENT 7.0 .TP .B Parameters \fBhigh\fP (\fIbool\fP) – .TP .B Return type Intracomm .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B remote_group remote group .UNINDENT .INDENT 7.0 .TP .B remote_size number of remote processes .UNINDENT .UNINDENT .SS mpi4py.MPI.Intracomm .INDENT 0.0 .TP .B class mpi4py.MPI.Intracomm(comm=None) Bases: \fBmpi4py.MPI.Comm\fP .sp Intracommunicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIIntracomm\fP\fI]\fP) – .TP .B Return type Intracomm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fIIntracomm\fP\fI]\fP) – .TP .B Return type Intracomm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Accept\fP(port_name[, info, root]) T} T{ Accept a request to form a new intercommunicator T} _ T{ \fI\%Cart_map\fP(dims[, periods]) T} T{ Return an optimal placement for the calling process on the physical machine T} _ T{ \fI\%Connect\fP(port_name[, info, root]) T} T{ Make a request to form a new intercommunicator T} _ T{ \fI\%Create_cart\fP(dims[, periods, reorder]) T} T{ Create cartesian communicator T} _ T{ \fI\%Create_dist_graph\fP(sources, degrees, destinations) T} T{ Create distributed graph communicator T} _ T{ \fI\%Create_dist_graph_adjacent\fP(sources, destinations) T} T{ Create distributed graph communicator T} _ T{ \fI\%Create_graph\fP(index, edges[, reorder]) T} T{ Create graph communicator T} _ T{ \fI\%Create_intercomm\fP(local_leader, peer_comm, ...) T} T{ Create intercommunicator T} _ T{ \fI\%Exscan\fP(sendbuf, recvbuf[, op]) T} T{ Exclusive Scan T} _ T{ \fI\%Graph_map\fP(index, edges) T} T{ Return an optimal placement for the calling process on the physical machine T} _ T{ \fI\%Iexscan\fP(sendbuf, recvbuf[, op]) T} T{ Inclusive Scan T} _ T{ \fI\%Iscan\fP(sendbuf, recvbuf[, op]) T} T{ Inclusive Scan T} _ T{ \fI\%Scan\fP(sendbuf, recvbuf[, op]) T} T{ Inclusive Scan T} _ T{ \fI\%Spawn\fP(command[, args, maxprocs, info, root, ...]) T} T{ Spawn instances of a single MPI application T} _ T{ \fI\%Spawn_multiple\fP(command[, args, maxprocs, ...]) T} T{ Spawn instances of multiple MPI applications T} _ T{ \fI\%exscan\fP(sendobj[, op]) T} T{ Exclusive Scan T} _ T{ \fI\%scan\fP(sendobj[, op]) T} T{ Inclusive Scan T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Accept(port_name, info=INFO_NULL, root=0) Accept a request to form a new intercommunicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBport_name\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Cart_map(dims, periods=None) Return an optimal placement for the calling process on the physical machine .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdims\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBperiods\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIbool\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Connect(port_name, info=INFO_NULL, root=0) Make a request to form a new intercommunicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBport_name\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .UNINDENT .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_cart(dims, periods=None, reorder=False) Create cartesian communicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdims\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBperiods\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIbool\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBreorder\fP (\fIbool\fP) – .UNINDENT .TP .B Return type Cartcomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_dist_graph(sources, degrees, destinations, weights=None, info=INFO_NULL, reorder=False) Create distributed graph communicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsources\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdegrees\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdestinations\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBweights\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBreorder\fP (\fIbool\fP) – .UNINDENT .TP .B Return type Distgraphcomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_dist_graph_adjacent(sources, destinations, sourceweights=None, destweights=None, info=INFO_NULL, reorder=False) Create distributed graph communicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsources\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBdestinations\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBsourceweights\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBdestweights\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBreorder\fP (\fIbool\fP) – .UNINDENT .TP .B Return type Distgraphcomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_graph(index, edges, reorder=False) Create graph communicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBindex\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBedges\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBreorder\fP (\fIbool\fP) – .UNINDENT .TP .B Return type Graphcomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Create_intercomm(local_leader, peer_comm, remote_leader, tag=0) Create intercommunicator .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBlocal_leader\fP (\fIint\fP) – .IP \(bu 2 \fBpeer_comm\fP (\fIIntracomm\fP) – .IP \(bu 2 \fBremote_leader\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .UNINDENT .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Exscan(sendbuf, recvbuf, op=SUM) Exclusive Scan .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Graph_map(index, edges) Return an optimal placement for the calling process on the physical machine .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBindex\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .IP \(bu 2 \fBedges\fP (\fISequence\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iexscan(sendbuf, recvbuf, op=SUM) Inclusive Scan .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Iscan(sendbuf, recvbuf, op=SUM) Inclusive Scan .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Scan(sendbuf, recvbuf, op=SUM) Inclusive Scan .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIUnion\fP\fI[\fP\fIBufSpec\fP\fI, \fP\fIInPlace\fP\fI]\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Spawn(command, args=None, maxprocs=1, info=INFO_NULL, root=0, errcodes=None) Spawn instances of a single MPI application .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcommand\fP (\fIstr\fP) – .IP \(bu 2 \fBargs\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIstr\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBmaxprocs\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .IP \(bu 2 \fBerrcodes\fP (\fIOptional\fP\fI[\fP\fIlist\fP\fI]\fP) – .UNINDENT .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Spawn_multiple(command, args=None, maxprocs=None, info=INFO_NULL, root=0, errcodes=None) Spawn instances of multiple MPI applications .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcommand\fP (\fISequence\fP\fI[\fP\fIstr\fP\fI]\fP) – .IP \(bu 2 \fBargs\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fISequence\fP\fI[\fP\fIstr\fP\fI]\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBmaxprocs\fP (\fIOptional\fP\fI[\fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBinfo\fP (\fIUnion\fP\fI[\fP\fIInfo\fP\fI, \fP\fISequence\fP\fI[\fP\fIInfo\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBroot\fP (\fIint\fP) – .IP \(bu 2 \fBerrcodes\fP (\fIOptional\fP\fI[\fP\fIlist\fP\fI]\fP) – .UNINDENT .TP .B Return type Intercomm .UNINDENT .UNINDENT .INDENT 7.0 .TP .B exscan(sendobj, op=SUM) Exclusive Scan .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBop\fP (\fIUnion\fP\fI[\fP\fIOp\fP\fI, \fP\fICallable\fP\fI[\fP\fI[\fP\fIAny\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B scan(sendobj, op=SUM) Inclusive Scan .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendobj\fP (\fIAny\fP) – .IP \(bu 2 \fBop\fP (\fIUnion\fP\fI[\fP\fIOp\fP\fI, \fP\fICallable\fP\fI[\fP\fI[\fP\fIAny\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Message .INDENT 0.0 .TP .B class mpi4py.MPI.Message(message=None) Bases: \fBobject\fP .sp Matched message handle .INDENT 7.0 .TP .B Parameters \fBmessage\fP (\fIOptional\fP\fI[\fP\fIMessage\fP\fI]\fP) – .TP .B Return type Message .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, message=None) .INDENT 7.0 .TP .B Parameters \fBmessage\fP (\fIOptional\fP\fI[\fP\fIMessage\fP\fI]\fP) – .TP .B Return type Message .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Iprobe\fP(comm[, source, tag, status]) T} T{ Nonblocking test for a matched message T} _ T{ \fI\%Irecv\fP(buf) T} T{ Nonblocking receive of matched message T} _ T{ \fI\%Probe\fP(comm[, source, tag, status]) T} T{ Blocking test for a matched message T} _ T{ \fI\%Recv\fP(buf[, status]) T} T{ Blocking receive of matched message T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%iprobe\fP(comm[, source, tag, status]) T} T{ Nonblocking test for a matched message T} _ T{ \fI\%irecv\fP() T} T{ Nonblocking receive of matched message T} _ T{ \fI\%probe\fP(comm[, source, tag, status]) T} T{ Blocking test for a matched message T} _ T{ \fI\%py2f\fP() T} T{ T} _ T{ \fI\%recv\fP([status]) T} T{ Blocking receive of matched message T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B classmethod Iprobe(comm, source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[Message] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Irecv(buf) Nonblocking receive of matched message .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBufSpec\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Probe(comm, source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Message .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Recv(buf, status=None) Blocking receive of matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Message .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod iprobe(comm, source=ANY_SOURCE, tag=ANY_TAG, status=None) Nonblocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[Message] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B irecv() Nonblocking receive of matched message .INDENT 7.0 .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod probe(comm, source=ANY_SOURCE, tag=ANY_TAG, status=None) Blocking test for a matched message .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcomm\fP (\fIComm\fP) – .IP \(bu 2 \fBsource\fP (\fIint\fP) – .IP \(bu 2 \fBtag\fP (\fIint\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Message .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B recv(status=None) Blocking receive of matched message .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Any .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Op .INDENT 0.0 .TP .B class mpi4py.MPI.Op(op=None) Bases: \fBobject\fP .sp Operation object .INDENT 7.0 .TP .B Parameters \fBop\fP (\fIOptional\fP\fI[\fP\fIOp\fP\fI]\fP) – .TP .B Return type Op .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, op=None) .INDENT 7.0 .TP .B Parameters \fBop\fP (\fIOptional\fP\fI[\fP\fIOp\fP\fI]\fP) – .TP .B Return type Op .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Create\fP(function[, commute]) T} T{ Create a user\-defined operation T} _ T{ \fI\%Free\fP() T} T{ Free the operation T} _ T{ \fI\%Is_commutative\fP() T} T{ Query reduction operations for their commutativity T} _ T{ \fI\%Reduce_local\fP(inbuf, inoutbuf) T} T{ Apply a reduction operator to local data T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%is_commutative\fP T} T{ is commutative T} _ T{ \fI\%is_predefined\fP T} T{ is a predefined operation T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B classmethod Create(function, commute=False) Create a user\-defined operation .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBfunction\fP (\fICallable\fP\fI[\fP\fI[\fP\fIBuffer\fP\fI, \fP\fIBuffer\fP\fI, \fP\fIDatatype\fP\fI]\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBcommute\fP (\fIbool\fP) – .UNINDENT .TP .B Return type Op .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free the operation .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Is_commutative() Query reduction operations for their commutativity .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Reduce_local(inbuf, inoutbuf) Apply a reduction operator to local data .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBinbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBinoutbuf\fP (\fIBufSpec\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Op .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B is_commutative is commutative .UNINDENT .INDENT 7.0 .TP .B is_predefined is a predefined operation .UNINDENT .UNINDENT .SS mpi4py.MPI.Pickle .INDENT 0.0 .TP .B class mpi4py.MPI.Pickle(dumps=None, loads=None, protocol=None) Bases: \fBobject\fP .sp Pickle/unpickle Python objects .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdumps\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIAny\fP\fI, \fP\fIint\fP\fI]\fP\fI, \fP\fIbytes\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBloads\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIBuffer\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBprotocol\fP (\fIOptional\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .INDENT 7.0 .TP .B __init__(dumps=None, loads=None, protocol=None) .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdumps\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIAny\fP\fI, \fP\fIint\fP\fI]\fP\fI, \fP\fIbytes\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBloads\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIBuffer\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBprotocol\fP (\fIOptional\fP\fI[\fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%dumps\fP(obj[, buffer_callback]) T} T{ Serialize object to pickle data stream. T} _ T{ \fI\%loads\fP(data[, buffers]) T} T{ Deserialize object from pickle data stream. T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%PROTOCOL\fP T} T{ pickle protocol T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B dumps(obj, buffer_callback=None) Serialize object to pickle data stream. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIAny\fP) – .IP \(bu 2 \fBbuffer_callback\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIBuffer\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type bytes .UNINDENT .UNINDENT .INDENT 7.0 .TP .B loads(data, buffers=None) Deserialize object from pickle data stream. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdata\fP (\fIBuffer\fP) – .IP \(bu 2 \fBbuffers\fP (\fIOptional\fP\fI[\fP\fIIterable\fP\fI[\fP\fIBuffer\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B PROTOCOL pickle protocol .UNINDENT .UNINDENT .SS mpi4py.MPI.Prequest .INDENT 0.0 .TP .B class mpi4py.MPI.Prequest(request=None) Bases: \fBmpi4py.MPI.Request\fP .sp Persistent request handle .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIOptional\fP\fI[\fP\fIPrequest\fP\fI]\fP) – .TP .B Return type Prequest .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, request=None) .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIOptional\fP\fI[\fP\fIPrequest\fP\fI]\fP) – .TP .B Return type Prequest .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Start\fP() T} T{ Initiate a communication with a persistent request T} _ T{ \fI\%Startall\fP(requests) T} T{ Start a collection of persistent requests T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Start() Initiate a communication with a persistent request .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Startall(requests) Start a collection of persistent requests .INDENT 7.0 .TP .B Parameters \fBrequests\fP (\fIList\fP\fI[\fP\fIPrequest\fP\fI]\fP) – .TP .B Return type None .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Request .INDENT 0.0 .TP .B class mpi4py.MPI.Request(request=None) Bases: \fBobject\fP .sp Request handle .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIOptional\fP\fI[\fP\fIRequest\fP\fI]\fP) – .TP .B Return type Request .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, request=None) .INDENT 7.0 .TP .B Parameters \fBrequest\fP (\fIOptional\fP\fI[\fP\fIRequest\fP\fI]\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Cancel\fP() T} T{ Cancel a communication request T} _ T{ \fI\%Free\fP() T} T{ Free a communication request T} _ T{ \fI\%Get_status\fP([status]) T} T{ Non\-destructive test for the completion of a request T} _ T{ \fI\%Test\fP([status]) T} T{ Test for the completion of a send or receive T} _ T{ \fI\%Testall\fP(requests[, statuses]) T} T{ Test for completion of all previously initiated requests T} _ T{ \fI\%Testany\fP(requests[, status]) T} T{ Test for completion of any previously initiated request T} _ T{ \fI\%Testsome\fP(requests[, statuses]) T} T{ Test for completion of some previously initiated requests T} _ T{ \fI\%Wait\fP([status]) T} T{ Wait for a send or receive to complete T} _ T{ \fI\%Waitall\fP(requests[, statuses]) T} T{ Wait for all previously initiated requests to complete T} _ T{ \fI\%Waitany\fP(requests[, status]) T} T{ Wait for any previously initiated request to complete T} _ T{ \fI\%Waitsome\fP(requests[, statuses]) T} T{ Wait for some previously initiated requests to complete T} _ T{ \fI\%cancel\fP() T} T{ Cancel a communication request T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%get_status\fP([status]) T} T{ Non\-destructive test for the completion of a request T} _ T{ \fI\%py2f\fP() T} T{ T} _ T{ \fI\%test\fP([status]) T} T{ Test for the completion of a send or receive T} _ T{ \fI\%testall\fP(requests[, statuses]) T} T{ Test for completion of all previously initiated requests T} _ T{ \fI\%testany\fP(requests[, status]) T} T{ Test for completion of any previously initiated request T} _ T{ \fI\%testsome\fP(requests[, statuses]) T} T{ Test for completion of some previously initiated requests T} _ T{ \fI\%wait\fP([status]) T} T{ Wait for a send or receive to complete T} _ T{ \fI\%waitall\fP(requests[, statuses]) T} T{ Wait for all previously initiated requests to complete T} _ T{ \fI\%waitany\fP(requests[, status]) T} T{ Wait for any previously initiated request to complete T} _ T{ \fI\%waitsome\fP(requests[, statuses]) T} T{ Wait for some previously initiated requests to complete T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Cancel() Cancel a communication request .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free a communication request .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_status(status=None) Non\-destructive test for the completion of a request .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Test(status=None) Test for the completion of a send or receive .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Testall(requests, statuses=None) Test for completion of all previously initiated requests .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Testany(requests, status=None) Test for completion of any previously initiated request .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Tuple[int, bool] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Testsome(requests, statuses=None) Test for completion of some previously initiated requests .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[List[int]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Wait(status=None) Wait for a send or receive to complete .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Literal[True] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Waitall(requests, statuses=None) Wait for all previously initiated requests to complete .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Literal[True] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Waitany(requests, status=None) Wait for any previously initiated request to complete .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Waitsome(requests, statuses=None) Wait for some previously initiated requests to complete .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Optional[List[int]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B cancel() Cancel a communication request .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B get_status(status=None) Non\-destructive test for the completion of a request .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B test(status=None) Test for the completion of a send or receive .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Tuple[bool, Optional[Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod testall(requests, statuses=None) Test for completion of all previously initiated requests .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Tuple[bool, Optional[List[Any]]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod testany(requests, status=None) Test for completion of any previously initiated request .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Tuple[int, bool, Optional[Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod testsome(requests, statuses=None) Test for completion of some previously initiated requests .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Tuple[Optional[List[int]], Optional[List[Any]]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B wait(status=None) Wait for a send or receive to complete .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Any .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod waitall(requests, statuses=None) Wait for all previously initiated requests to complete .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type List[Any] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod waitany(requests, status=None) Wait for any previously initiated request to complete .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .UNINDENT .TP .B Return type Tuple[int, Any] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod waitsome(requests, statuses=None) Wait for some previously initiated requests to complete .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrequests\fP (\fISequence\fP\fI[\fP\fIRequest\fP\fI]\fP) – .IP \(bu 2 \fBstatuses\fP (\fIOptional\fP\fI[\fP\fIList\fP\fI[\fP\fIStatus\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type Tuple[Optional[List[int]], Optional[List[Any]]] .UNINDENT .UNINDENT .UNINDENT .SS mpi4py.MPI.Status .INDENT 0.0 .TP .B class mpi4py.MPI.Status(status=None) Bases: \fBobject\fP .sp Status object .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Status .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, status=None) .INDENT 7.0 .TP .B Parameters \fBstatus\fP (\fIOptional\fP\fI[\fP\fIStatus\fP\fI]\fP) – .TP .B Return type Status .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Get_count\fP([datatype]) T} T{ Get the number of \fItop level\fP elements T} _ T{ \fI\%Get_elements\fP(datatype) T} T{ Get the number of basic elements in a datatype T} _ T{ \fI\%Get_error\fP() T} T{ Get message error T} _ T{ \fI\%Get_source\fP() T} T{ Get message source T} _ T{ \fI\%Get_tag\fP() T} T{ Get message tag T} _ T{ \fI\%Is_cancelled\fP() T} T{ Test to see if a request was cancelled T} _ T{ \fI\%Set_cancelled\fP(flag) T} T{ Set the cancelled state associated with a status T} _ T{ \fI\%Set_elements\fP(datatype, count) T} T{ Set the number of elements in a status T} _ T{ \fI\%Set_error\fP(error) T} T{ Set message error T} _ T{ \fI\%Set_source\fP(source) T} T{ Set message source T} _ T{ \fI\%Set_tag\fP(tag) T} T{ Set message tag T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%cancelled\fP T} T{ cancelled state T} _ T{ \fI\%count\fP T} T{ byte count T} _ T{ \fI\%error\fP T} T{ T} _ T{ \fI\%source\fP T} T{ T} _ T{ \fI\%tag\fP T} T{ T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Get_count(datatype=BYTE) Get the number of \fItop level\fP elements .INDENT 7.0 .TP .B Parameters \fBdatatype\fP (\fIDatatype\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_elements(datatype) Get the number of basic elements in a datatype .INDENT 7.0 .TP .B Parameters \fBdatatype\fP (\fIDatatype\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_error() Get message error .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_source() Get message source .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_tag() Get message tag .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Is_cancelled() Test to see if a request was cancelled .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_cancelled(flag) Set the cancelled state associated with a status .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This should be only used when implementing query callback functions for generalized requests .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters \fBflag\fP (\fIbool\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_elements(datatype, count) Set the number of elements in a status .sp \fBNOTE:\fP .INDENT 7.0 .INDENT 3.5 This should be only used when implementing query callback functions for generalized requests .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdatatype\fP (\fIDatatype\fP) – .IP \(bu 2 \fBcount\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_error(error) Set message error .INDENT 7.0 .TP .B Parameters \fBerror\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_source(source) Set message source .INDENT 7.0 .TP .B Parameters \fBsource\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_tag(tag) Set message tag .INDENT 7.0 .TP .B Parameters \fBtag\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIList\fP\fI[\fP\fIint\fP\fI]\fP) – .TP .B Return type Status .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type List[int] .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B cancelled cancelled state .UNINDENT .INDENT 7.0 .TP .B count byte count .UNINDENT .INDENT 7.0 .TP .B error .UNINDENT .INDENT 7.0 .TP .B source .UNINDENT .INDENT 7.0 .TP .B tag .UNINDENT .UNINDENT .SS mpi4py.MPI.Topocomm .INDENT 0.0 .TP .B class mpi4py.MPI.Topocomm(comm=None) Bases: \fBmpi4py.MPI.Intracomm\fP .sp Topology intracommunicator .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fITopocomm\fP\fI]\fP) – .TP .B Return type Topocomm .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, comm=None) .INDENT 7.0 .TP .B Parameters \fBcomm\fP (\fIOptional\fP\fI[\fP\fITopocomm\fP\fI]\fP) – .TP .B Return type Topocomm .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Ineighbor_allgather\fP(sendbuf, recvbuf) T} T{ Nonblocking Neighbor Gather to All T} _ T{ \fI\%Ineighbor_allgatherv\fP(sendbuf, recvbuf) T} T{ Nonblocking Neighbor Gather to All Vector T} _ T{ \fI\%Ineighbor_alltoall\fP(sendbuf, recvbuf) T} T{ Nonblocking Neighbor All\-to\-All T} _ T{ \fI\%Ineighbor_alltoallv\fP(sendbuf, recvbuf) T} T{ Nonblocking Neighbor All\-to\-All Vector T} _ T{ \fI\%Ineighbor_alltoallw\fP(sendbuf, recvbuf) T} T{ Nonblocking Neighbor All\-to\-All Generalized T} _ T{ \fI\%Neighbor_allgather\fP(sendbuf, recvbuf) T} T{ Neighbor Gather to All T} _ T{ \fI\%Neighbor_allgatherv\fP(sendbuf, recvbuf) T} T{ Neighbor Gather to All Vector T} _ T{ \fI\%Neighbor_alltoall\fP(sendbuf, recvbuf) T} T{ Neighbor All\-to\-All T} _ T{ \fI\%Neighbor_alltoallv\fP(sendbuf, recvbuf) T} T{ Neighbor All\-to\-All Vector T} _ T{ \fI\%Neighbor_alltoallw\fP(sendbuf, recvbuf) T} T{ Neighbor All\-to\-All Generalized T} _ T{ \fI\%neighbor_allgather\fP(sendobj) T} T{ Neighbor Gather to All T} _ T{ \fI\%neighbor_alltoall\fP(sendobj) T} T{ Neighbor All to All Scatter/Gather T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%degrees\fP T} T{ number of incoming and outgoing neighbors T} _ T{ \fI\%indegree\fP T} T{ number of incoming neighbors T} _ T{ \fI\%inedges\fP T} T{ incoming neighbors T} _ T{ \fI\%inoutedges\fP T} T{ incoming and outgoing neighbors T} _ T{ \fI\%outdegree\fP T} T{ number of outgoing neighbors T} _ T{ \fI\%outedges\fP T} T{ outgoing neighbors T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Ineighbor_allgather(sendbuf, recvbuf) Nonblocking Neighbor Gather to All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ineighbor_allgatherv(sendbuf, recvbuf) Nonblocking Neighbor Gather to All Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ineighbor_alltoall(sendbuf, recvbuf) Nonblocking Neighbor All\-to\-All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpecB\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ineighbor_alltoallv(sendbuf, recvbuf) Nonblocking Neighbor All\-to\-All Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpecV\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Ineighbor_alltoallw(sendbuf, recvbuf) Nonblocking Neighbor All\-to\-All Generalized .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpecW\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecW\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Neighbor_allgather(sendbuf, recvbuf) Neighbor Gather to All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Neighbor_allgatherv(sendbuf, recvbuf) Neighbor Gather to All Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Neighbor_alltoall(sendbuf, recvbuf) Neighbor All\-to\-All .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpecB\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecB\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Neighbor_alltoallv(sendbuf, recvbuf) Neighbor All\-to\-All Vector .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpecV\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecV\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Neighbor_alltoallw(sendbuf, recvbuf) Neighbor All\-to\-All Generalized .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsendbuf\fP (\fIBufSpecW\fP) – .IP \(bu 2 \fBrecvbuf\fP (\fIBufSpecW\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B neighbor_allgather(sendobj) Neighbor Gather to All .INDENT 7.0 .TP .B Parameters \fBsendobj\fP (\fIAny\fP) – .TP .B Return type List[Any] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B neighbor_alltoall(sendobj) Neighbor All to All Scatter/Gather .INDENT 7.0 .TP .B Parameters \fBsendobj\fP (\fIList\fP\fI[\fP\fIAny\fP\fI]\fP) – .TP .B Return type List[Any] .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B degrees number of incoming and outgoing neighbors .UNINDENT .INDENT 7.0 .TP .B indegree number of incoming neighbors .UNINDENT .INDENT 7.0 .TP .B inedges incoming neighbors .UNINDENT .INDENT 7.0 .TP .B inoutedges incoming and outgoing neighbors .UNINDENT .INDENT 7.0 .TP .B outdegree number of outgoing neighbors .UNINDENT .INDENT 7.0 .TP .B outedges outgoing neighbors .UNINDENT .UNINDENT .SS mpi4py.MPI.Win .INDENT 0.0 .TP .B class mpi4py.MPI.Win(win=None) Bases: \fBobject\fP .sp Window handle .INDENT 7.0 .TP .B Parameters \fBwin\fP (\fIOptional\fP\fI[\fP\fIWin\fP\fI]\fP) – .TP .B Return type Win .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, win=None) .INDENT 7.0 .TP .B Parameters \fBwin\fP (\fIOptional\fP\fI[\fP\fIWin\fP\fI]\fP) – .TP .B Return type Win .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Accumulate\fP(origin, target_rank[, target, op]) T} T{ Accumulate data into the target process T} _ T{ \fI\%Allocate\fP(size[, disp_unit, info, comm]) T} T{ Create an window object for one\-sided communication T} _ T{ \fI\%Allocate_shared\fP(size[, disp_unit, info, comm]) T} T{ Create an window object for one\-sided communication T} _ T{ \fI\%Attach\fP(memory) T} T{ Attach a local memory region T} _ T{ \fI\%Call_errhandler\fP(errorcode) T} T{ Call the error handler installed on a window T} _ T{ \fI\%Compare_and_swap\fP(origin, compare, result, ...) T} T{ Perform one\-sided atomic compare\-and\-swap T} _ T{ \fI\%Complete\fP() T} T{ Completes an RMA operations begun after an \fI\%Win.Start()\fP T} _ T{ \fI\%Create\fP(memory[, disp_unit, info, comm]) T} T{ Create an window object for one\-sided communication T} _ T{ \fI\%Create_dynamic\fP([info, comm]) T} T{ Create an window object for one\-sided communication T} _ T{ \fI\%Create_keyval\fP([copy_fn, delete_fn, nopython]) T} T{ Create a new attribute key for windows T} _ T{ \fI\%Delete_attr\fP(keyval) T} T{ Delete attribute value associated with a key T} _ T{ \fI\%Detach\fP(memory) T} T{ Detach a local memory region T} _ T{ \fI\%Fence\fP([assertion]) T} T{ Perform an MPI fence synchronization on a window T} _ T{ \fI\%Fetch_and_op\fP(origin, result, target_rank[, ...]) T} T{ Perform one\-sided read\-modify\-write T} _ T{ \fI\%Flush\fP(rank) T} T{ Complete all outstanding RMA operations at the given target T} _ T{ \fI\%Flush_all\fP() T} T{ Complete all outstanding RMA operations at all targets T} _ T{ \fI\%Flush_local\fP(rank) T} T{ Complete locally all outstanding RMA operations at the given target T} _ T{ \fI\%Flush_local_all\fP() T} T{ Complete locally all outstanding RMA opera\- tions at all targets T} _ T{ \fI\%Free\fP() T} T{ Free a window T} _ T{ \fI\%Free_keyval\fP(keyval) T} T{ Free an attribute key for windows T} _ T{ \fI\%Get\fP(origin, target_rank[, target]) T} T{ Get data from a memory window on a remote process. T} _ T{ \fI\%Get_accumulate\fP(origin, result, target_rank) T} T{ Fetch\-and\-accumulate data into the target process T} _ T{ \fI\%Get_attr\fP(keyval) T} T{ Retrieve attribute value by key T} _ T{ \fI\%Get_errhandler\fP() T} T{ Get the error handler for a window T} _ T{ \fI\%Get_group\fP() T} T{ Return a duplicate of the group of the communicator used to create the window T} _ T{ \fI\%Get_info\fP() T} T{ Return the hints for a windows that are currently in use T} _ T{ \fI\%Get_name\fP() T} T{ Get the print name associated with the window T} _ T{ \fI\%Lock\fP(rank[, lock_type, assertion]) T} T{ Begin an RMA access epoch at the target process T} _ T{ \fI\%Lock_all\fP([assertion]) T} T{ Begin an RMA access epoch at all processes T} _ T{ \fI\%Post\fP(group[, assertion]) T} T{ Start an RMA exposure epoch T} _ T{ \fI\%Put\fP(origin, target_rank[, target]) T} T{ Put data into a memory window on a remote process. T} _ T{ \fI\%Raccumulate\fP(origin, target_rank[, target, op]) T} T{ Fetch\-and\-accumulate data into the target process T} _ T{ \fI\%Rget\fP(origin, target_rank[, target]) T} T{ Get data from a memory window on a remote process. T} _ T{ \fI\%Rget_accumulate\fP(origin, result, target_rank) T} T{ Accumulate data into the target process using remote memory access. T} _ T{ \fI\%Rput\fP(origin, target_rank[, target]) T} T{ Put data into a memory window on a remote process. T} _ T{ \fI\%Set_attr\fP(keyval, attrval) T} T{ Store attribute value associated with a key T} _ T{ \fI\%Set_errhandler\fP(errhandler) T} T{ Set the error handler for a window T} _ T{ \fI\%Set_info\fP(info) T} T{ Set new values for the hints associated with a window T} _ T{ \fI\%Set_name\fP(name) T} T{ Set the print name associated with the window T} _ T{ \fI\%Shared_query\fP(rank) T} T{ Query the process\-local address for remote memory segments created with \fI\%Win.Allocate_shared()\fP T} _ T{ \fI\%Start\fP(group[, assertion]) T} T{ Start an RMA access epoch for MPI T} _ T{ \fI\%Sync\fP() T} T{ Synchronize public and private copies of the given window T} _ T{ \fI\%Test\fP() T} T{ Test whether an RMA exposure epoch has completed T} _ T{ \fI\%Unlock\fP(rank) T} T{ Complete an RMA access epoch at the target process T} _ T{ \fI\%Unlock_all\fP() T} T{ Complete an RMA access epoch at all processes T} _ T{ \fI\%Wait\fP() T} T{ Complete an RMA exposure epoch begun with \fI\%Win.Post()\fP T} _ T{ \fI\%f2py\fP(arg) T} T{ T} _ T{ \fI\%py2f\fP() T} T{ T} _ T{ \fI\%tomemory\fP() T} T{ Return window memory buffer T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%attrs\fP T} T{ window attributes T} _ T{ \fI\%flavor\fP T} T{ window create flavor T} _ T{ \fI\%group\fP T} T{ window group T} _ T{ \fI\%info\fP T} T{ window info T} _ T{ \fI\%model\fP T} T{ window memory model T} _ T{ \fI\%name\fP T} T{ window name T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Accumulate(origin, target_rank, target=None, op=SUM) Accumulate data into the target process .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Allocate(size, disp_unit=1, info=INFO_NULL, comm=COMM_SELF) Create an window object for one\-sided communication .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsize\fP (\fIint\fP) – .IP \(bu 2 \fBdisp_unit\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBcomm\fP (\fIIntracomm\fP) – .UNINDENT .TP .B Return type Win .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Allocate_shared(size, disp_unit=1, info=INFO_NULL, comm=COMM_SELF) Create an window object for one\-sided communication .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsize\fP (\fIint\fP) – .IP \(bu 2 \fBdisp_unit\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBcomm\fP (\fIIntracomm\fP) – .UNINDENT .TP .B Return type Win .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Attach(memory) Attach a local memory region .INDENT 7.0 .TP .B Parameters \fBmemory\fP (\fIBuffer\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Call_errhandler(errorcode) Call the error handler installed on a window .INDENT 7.0 .TP .B Parameters \fBerrorcode\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Compare_and_swap(origin, compare, result, target_rank, target_disp=0) Perform one\-sided atomic compare\-and\-swap .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBcompare\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBresult\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget_disp\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Complete() Completes an RMA operations begun after an \fI\%Win.Start()\fP .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create(memory, disp_unit=1, info=INFO_NULL, comm=COMM_SELF) Create an window object for one\-sided communication .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBmemory\fP (\fIUnion\fP\fI[\fP\fIBuffer\fP\fI, \fP\fIBottom\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBdisp_unit\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBcomm\fP (\fIIntracomm\fP) – .UNINDENT .TP .B Return type Win .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_dynamic(info=INFO_NULL, comm=COMM_SELF) Create an window object for one\-sided communication .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .IP \(bu 2 \fBcomm\fP (\fIIntracomm\fP) – .UNINDENT .TP .B Return type Win .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Create_keyval(copy_fn=None, delete_fn=None, nopython=False) Create a new attribute key for windows .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBcopy_fn\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIWin\fP\fI, \fP\fIint\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fIAny\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBdelete_fn\fP (\fIOptional\fP\fI[\fP\fICallable\fP\fI[\fP\fI[\fP\fIWin\fP\fI, \fP\fIint\fP\fI, \fP\fIAny\fP\fI]\fP\fI, \fP\fINone\fP\fI]\fP\fI]\fP) – .IP \(bu 2 \fBnopython\fP (\fIbool\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Delete_attr(keyval) Delete attribute value associated with a key .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Detach(memory) Detach a local memory region .INDENT 7.0 .TP .B Parameters \fBmemory\fP (\fIBuffer\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Fence(assertion=0) Perform an MPI fence synchronization on a window .INDENT 7.0 .TP .B Parameters \fBassertion\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Fetch_and_op(origin, result, target_rank, target_disp=0, op=SUM) Perform one\-sided read\-modify\-write .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBresult\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget_disp\fP (\fIint\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Flush(rank) Complete all outstanding RMA operations at the given target .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Flush_all() Complete all outstanding RMA operations at all targets .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Flush_local(rank) Complete locally all outstanding RMA operations at the given target .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Flush_local_all() Complete locally all outstanding RMA opera\- tions at all targets .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Free() Free a window .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod Free_keyval(keyval) Free an attribute key for windows .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get(origin, target_rank, target=None) Get data from a memory window on a remote process. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_accumulate(origin, result, target_rank, target=None, op=SUM) Fetch\-and\-accumulate data into the target process .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBresult\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_attr(keyval) Retrieve attribute value by key .INDENT 7.0 .TP .B Parameters \fBkeyval\fP (\fIint\fP) – .TP .B Return type Optional[Union[int, Any]] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_errhandler() Get the error handler for a window .INDENT 7.0 .TP .B Return type Errhandler .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_group() Return a duplicate of the group of the communicator used to create the window .INDENT 7.0 .TP .B Return type Group .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_info() Return the hints for a windows that are currently in use .INDENT 7.0 .TP .B Return type Info .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_name() Get the print name associated with the window .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Lock(rank, lock_type=LOCK_EXCLUSIVE, assertion=0) Begin an RMA access epoch at the target process .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBrank\fP (\fIint\fP) – .IP \(bu 2 \fBlock_type\fP (\fIint\fP) – .IP \(bu 2 \fBassertion\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Lock_all(assertion=0) Begin an RMA access epoch at all processes .INDENT 7.0 .TP .B Parameters \fBassertion\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Post(group, assertion=0) Start an RMA exposure epoch .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup\fP (\fIGroup\fP) – .IP \(bu 2 \fBassertion\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Put(origin, target_rank, target=None) Put data into a memory window on a remote process. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Raccumulate(origin, target_rank, target=None, op=SUM) Fetch\-and\-accumulate data into the target process .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Rget(origin, target_rank, target=None) Get data from a memory window on a remote process. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Rget_accumulate(origin, result, target_rank, target=None, op=SUM) Accumulate data into the target process using remote memory access. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBresult\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .IP \(bu 2 \fBop\fP (\fIOp\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Rput(origin, target_rank, target=None) Put data into a memory window on a remote process. .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBorigin\fP (\fIBufSpec\fP) – .IP \(bu 2 \fBtarget_rank\fP (\fIint\fP) – .IP \(bu 2 \fBtarget\fP (\fIOptional\fP\fI[\fP\fITargetSpec\fP\fI]\fP) – .UNINDENT .TP .B Return type Request .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_attr(keyval, attrval) Store attribute value associated with a key .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBkeyval\fP (\fIint\fP) – .IP \(bu 2 \fBattrval\fP (\fIAny\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_errhandler(errhandler) Set the error handler for a window .INDENT 7.0 .TP .B Parameters \fBerrhandler\fP (\fIErrhandler\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_info(info) Set new values for the hints associated with a window .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIInfo\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Set_name(name) Set the print name associated with the window .INDENT 7.0 .TP .B Parameters \fBname\fP (\fIstr\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Shared_query(rank) Query the process\-local address for remote memory segments created with \fI\%Win.Allocate_shared()\fP .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type Tuple[memory, int] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Start(group, assertion=0) Start an RMA access epoch for MPI .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBgroup\fP (\fIGroup\fP) – .IP \(bu 2 \fBassertion\fP (\fIint\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Sync() Synchronize public and private copies of the given window .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Test() Test whether an RMA exposure epoch has completed .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Unlock(rank) Complete an RMA access epoch at the target process .INDENT 7.0 .TP .B Parameters \fBrank\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Unlock_all() Complete an RMA access epoch at all processes .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Wait() Complete an RMA exposure epoch begun with \fI\%Win.Post()\fP .INDENT 7.0 .TP .B Return type Literal[True] .UNINDENT .UNINDENT .INDENT 7.0 .TP .B classmethod f2py(arg) .INDENT 7.0 .TP .B Parameters \fBarg\fP (\fIint\fP) – .TP .B Return type Win .UNINDENT .UNINDENT .INDENT 7.0 .TP .B py2f() .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B tomemory() Return window memory buffer .INDENT 7.0 .TP .B Return type memory .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B attrs window attributes .UNINDENT .INDENT 7.0 .TP .B flavor window create flavor .UNINDENT .INDENT 7.0 .TP .B group window group .UNINDENT .INDENT 7.0 .TP .B info window info .UNINDENT .INDENT 7.0 .TP .B model window memory model .UNINDENT .INDENT 7.0 .TP .B name window name .UNINDENT .UNINDENT .SS mpi4py.MPI.memory .INDENT 0.0 .TP .B class mpi4py.MPI.memory(buf) Bases: \fBobject\fP .sp Memory buffer .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBuffer\fP) – .TP .B Return type memory .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, buf) .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBuffer\fP) – .TP .B Return type memory .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%allocate\fP(nbytes[, clear]) T} T{ Memory allocation T} _ T{ \fI\%fromaddress\fP(address, nbytes[, readonly]) T} T{ Memory from address and size in bytes T} _ T{ \fI\%frombuffer\fP(obj[, readonly]) T} T{ Memory from buffer\-like object T} _ T{ \fI\%release\fP() T} T{ Release the underlying buffer exposed by the memory object T} _ T{ \fI\%tobytes\fP([order]) T} T{ Return the data in the buffer as a byte string T} _ T{ \fI\%toreadonly\fP() T} T{ Return a readonly version of the memory object T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%address\fP T} T{ Memory address T} _ T{ \fI\%format\fP T} T{ A string with the format of each element T} _ T{ \fI\%itemsize\fP T} T{ The size in bytes of each element T} _ T{ \fI\%nbytes\fP T} T{ Memory size (in bytes) T} _ T{ \fI\%obj\fP T} T{ The underlying object of the memory T} _ T{ \fI\%readonly\fP T} T{ Boolean indicating whether the memory is read\-only T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B static allocate(nbytes, clear=False) Memory allocation .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBnbytes\fP (\fIint\fP) – .IP \(bu 2 \fBclear\fP (\fIbool\fP) – .UNINDENT .TP .B Return type memory .UNINDENT .UNINDENT .INDENT 7.0 .TP .B static fromaddress(address, nbytes, readonly=False) Memory from address and size in bytes .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBaddress\fP (\fIint\fP) – .IP \(bu 2 \fBnbytes\fP (\fIint\fP) – .IP \(bu 2 \fBreadonly\fP (\fIbool\fP) – .UNINDENT .TP .B Return type memory .UNINDENT .UNINDENT .INDENT 7.0 .TP .B static frombuffer(obj, readonly=False) Memory from buffer\-like object .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBobj\fP (\fIBuffer\fP) – .IP \(bu 2 \fBreadonly\fP (\fIbool\fP) – .UNINDENT .TP .B Return type memory .UNINDENT .UNINDENT .INDENT 7.0 .TP .B release() Release the underlying buffer exposed by the memory object .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .INDENT 7.0 .TP .B tobytes(order=None) Return the data in the buffer as a byte string .INDENT 7.0 .TP .B Parameters \fBorder\fP (\fIOptional\fP\fI[\fP\fIstr\fP\fI]\fP) – .TP .B Return type bytes .UNINDENT .UNINDENT .INDENT 7.0 .TP .B toreadonly() Return a readonly version of the memory object .INDENT 7.0 .TP .B Return type memory .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B address Memory address .UNINDENT .INDENT 7.0 .TP .B format A string with the format of each element .UNINDENT .INDENT 7.0 .TP .B itemsize The size in bytes of each element .UNINDENT .INDENT 7.0 .TP .B nbytes Memory size (in bytes) .UNINDENT .INDENT 7.0 .TP .B obj The underlying object of the memory .UNINDENT .INDENT 7.0 .TP .B readonly Boolean indicating whether the memory is read\-only .UNINDENT .UNINDENT .sp Exceptions .TS center; |l|l|. _ T{ \fBException\fP([ierr]) T} T{ Exception class T} _ .TE .SS mpi4py.MPI.Exception .INDENT 0.0 .TP .B exception mpi4py.MPI.Exception(ierr=SUCCESS) Bases: \fBRuntimeError\fP .sp Exception class .INDENT 7.0 .TP .B Parameters \fBierr\fP (\fIint\fP) – .TP .B Return type Exception .UNINDENT .INDENT 7.0 .TP .B static __new__(cls, ierr=SUCCESS) .INDENT 7.0 .TP .B Parameters \fBierr\fP (\fIint\fP) – .TP .B Return type Exception .UNINDENT .UNINDENT .sp Methods Summary .TS center; |l|l|. _ T{ \fI\%Get_error_class\fP() T} T{ Error class T} _ T{ \fI\%Get_error_code\fP() T} T{ Error code T} _ T{ \fI\%Get_error_string\fP() T} T{ Error string T} _ .TE .sp Attributes Summary .TS center; |l|l|. _ T{ \fI\%error_class\fP T} T{ error class T} _ T{ \fI\%error_code\fP T} T{ error code T} _ T{ \fI\%error_string\fP T} T{ error string T} _ .TE .sp Methods Documentation .INDENT 7.0 .TP .B Get_error_class() Error class .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_error_code() Error code .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .INDENT 7.0 .TP .B Get_error_string() Error string .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .sp Attributes Documentation .INDENT 7.0 .TP .B error_class error class .UNINDENT .INDENT 7.0 .TP .B error_code error code .UNINDENT .INDENT 7.0 .TP .B error_string error string .UNINDENT .UNINDENT .sp Functions .TS center; |l|l|. _ T{ \fBAdd_error_class\fP() T} T{ Add an \fIerror class\fP to the known error classes T} _ T{ \fBAdd_error_code\fP(errorclass) T} T{ Add an \fIerror code\fP to an \fIerror class\fP T} _ T{ \fBAdd_error_string\fP(errorcode, string) T} T{ Associate an \fIerror string\fP with an \fIerror class\fP or \fIerrorcode\fP T} _ T{ \fBAint_add\fP(base, disp) T} T{ Return the sum of base address and displacement T} _ T{ \fBAint_diff\fP(addr1, addr2) T} T{ Return the difference between absolute addresses T} _ T{ \fBAlloc_mem\fP(size[, info]) T} T{ Allocate memory for message passing and RMA T} _ T{ \fBAttach_buffer\fP(buf) T} T{ Attach a user\-provided buffer for sending in buffered mode T} _ T{ \fBClose_port\fP(port_name) T} T{ Close a port T} _ T{ \fBCompute_dims\fP(nnodes, dims) T} T{ Return a balanced distribution of processes per coordinate direction T} _ T{ \fBDetach_buffer\fP() T} T{ Remove an existing attached buffer T} _ T{ \fBFinalize\fP() T} T{ Terminate the MPI execution environment T} _ T{ \fBFree_mem\fP(mem) T} T{ Free memory allocated with \fBAlloc_mem()\fP T} _ T{ \fBGet_address\fP(location) T} T{ Get the address of a location in memory T} _ T{ \fBGet_error_class\fP(errorcode) T} T{ Convert an \fIerror code\fP into an \fIerror class\fP T} _ T{ \fBGet_error_string\fP(errorcode) T} T{ Return the \fIerror string\fP for a given \fIerror class\fP or \fIerror code\fP T} _ T{ \fBGet_library_version\fP() T} T{ Obtain the version string of the MPI library T} _ T{ \fBGet_processor_name\fP() T} T{ Obtain the name of the calling processor T} _ T{ \fBGet_version\fP() T} T{ Obtain the version number of the MPI standard supported by the implementation as a tuple \fB(version, subversion)\fP T} _ T{ \fBInit\fP() T} T{ Initialize the MPI execution environment T} _ T{ \fBInit_thread\fP([required]) T} T{ Initialize the MPI execution environment T} _ T{ \fBIs_finalized\fP() T} T{ Indicates whether \fBFinalize\fP has completed T} _ T{ \fBIs_initialized\fP() T} T{ Indicates whether \fBInit\fP has been called T} _ T{ \fBIs_thread_main\fP() T} T{ Indicate whether this thread called \fBInit\fP or \fBInit_thread\fP T} _ T{ \fBLookup_name\fP(service_name[, info]) T} T{ Lookup a port name given a service name T} _ T{ \fBOpen_port\fP([info]) T} T{ Return an address that can be used to establish connections between groups of MPI processes T} _ T{ \fBPcontrol\fP(level) T} T{ Control profiling T} _ T{ \fBPublish_name\fP(service_name, port_name[, info]) T} T{ Publish a service name T} _ T{ \fBQuery_thread\fP() T} T{ Return the level of thread support provided by the MPI library T} _ T{ \fBRegister_datarep\fP(datarep, read_fn, write_fn, ...) T} T{ Register user\-defined data representations T} _ T{ \fBUnpublish_name\fP(service_name, port_name[, info]) T} T{ Unpublish a service name T} _ T{ \fBWtick\fP() T} T{ Return the resolution of \fBWtime\fP T} _ T{ \fBWtime\fP() T} T{ Return an elapsed time on the calling processor T} _ T{ \fBget_vendor\fP() T} T{ Infomation about the underlying MPI implementation T} _ .TE .SS mpi4py.MPI.Add_error_class .INDENT 0.0 .TP .B mpi4py.MPI.Add_error_class() Add an \fIerror class\fP to the known error classes .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Add_error_code .INDENT 0.0 .TP .B mpi4py.MPI.Add_error_code(errorclass) Add an \fIerror code\fP to an \fIerror class\fP .INDENT 7.0 .TP .B Parameters \fBerrorclass\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Add_error_string .INDENT 0.0 .TP .B mpi4py.MPI.Add_error_string(errorcode, string) Associate an \fIerror string\fP with an \fIerror class\fP or \fIerrorcode\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBerrorcode\fP (\fIint\fP) – .IP \(bu 2 \fBstring\fP (\fIstr\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Aint_add .INDENT 0.0 .TP .B mpi4py.MPI.Aint_add(base, disp) Return the sum of base address and displacement .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBbase\fP (\fIint\fP) – .IP \(bu 2 \fBdisp\fP (\fIint\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Aint_diff .INDENT 0.0 .TP .B mpi4py.MPI.Aint_diff(addr1, addr2) Return the difference between absolute addresses .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBaddr1\fP (\fIint\fP) – .IP \(bu 2 \fBaddr2\fP (\fIint\fP) – .UNINDENT .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Alloc_mem .INDENT 0.0 .TP .B mpi4py.MPI.Alloc_mem(size, info=INFO_NULL) Allocate memory for message passing and RMA .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBsize\fP (\fIint\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type memory .UNINDENT .UNINDENT .SS mpi4py.MPI.Attach_buffer .INDENT 0.0 .TP .B mpi4py.MPI.Attach_buffer(buf) Attach a user\-provided buffer for sending in buffered mode .INDENT 7.0 .TP .B Parameters \fBbuf\fP (\fIBuffer\fP) – .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Close_port .INDENT 0.0 .TP .B mpi4py.MPI.Close_port(port_name) Close a port .INDENT 7.0 .TP .B Parameters \fBport_name\fP (\fIstr\fP) – .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Compute_dims .INDENT 0.0 .TP .B mpi4py.MPI.Compute_dims(nnodes, dims) Return a balanced distribution of processes per coordinate direction .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBnnodes\fP (\fIint\fP) – .IP \(bu 2 \fBdims\fP (\fIUnion\fP\fI[\fP\fIint\fP\fI, \fP\fISequence\fP\fI[\fP\fIint\fP\fI]\fP\fI]\fP) – .UNINDENT .TP .B Return type List[int] .UNINDENT .UNINDENT .SS mpi4py.MPI.Detach_buffer .INDENT 0.0 .TP .B mpi4py.MPI.Detach_buffer() Remove an existing attached buffer .INDENT 7.0 .TP .B Return type Buffer .UNINDENT .UNINDENT .SS mpi4py.MPI.Finalize .INDENT 0.0 .TP .B mpi4py.MPI.Finalize() Terminate the MPI execution environment .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Free_mem .INDENT 0.0 .TP .B mpi4py.MPI.Free_mem(mem) Free memory allocated with \fBAlloc_mem()\fP .INDENT 7.0 .TP .B Parameters \fBmem\fP (\fImemory\fP) – .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Get_address .INDENT 0.0 .TP .B mpi4py.MPI.Get_address(location) Get the address of a location in memory .INDENT 7.0 .TP .B Parameters \fBlocation\fP (\fIUnion\fP\fI[\fP\fIBuffer\fP\fI, \fP\fIBottom\fP\fI]\fP) – .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Get_error_class .INDENT 0.0 .TP .B mpi4py.MPI.Get_error_class(errorcode) Convert an \fIerror code\fP into an \fIerror class\fP .INDENT 7.0 .TP .B Parameters \fBerrorcode\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Get_error_string .INDENT 0.0 .TP .B mpi4py.MPI.Get_error_string(errorcode) Return the \fIerror string\fP for a given \fIerror class\fP or \fIerror code\fP .INDENT 7.0 .TP .B Parameters \fBerrorcode\fP (\fIint\fP) – .TP .B Return type str .UNINDENT .UNINDENT .SS mpi4py.MPI.Get_library_version .INDENT 0.0 .TP .B mpi4py.MPI.Get_library_version() Obtain the version string of the MPI library .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .SS mpi4py.MPI.Get_processor_name .INDENT 0.0 .TP .B mpi4py.MPI.Get_processor_name() Obtain the name of the calling processor .INDENT 7.0 .TP .B Return type str .UNINDENT .UNINDENT .SS mpi4py.MPI.Get_version .INDENT 0.0 .TP .B mpi4py.MPI.Get_version() Obtain the version number of the MPI standard supported by the implementation as a tuple \fB(version, subversion)\fP .INDENT 7.0 .TP .B Return type Tuple[int, int] .UNINDENT .UNINDENT .SS mpi4py.MPI.Init .INDENT 0.0 .TP .B mpi4py.MPI.Init() Initialize the MPI execution environment .INDENT 7.0 .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Init_thread .INDENT 0.0 .TP .B mpi4py.MPI.Init_thread(required=THREAD_MULTIPLE) Initialize the MPI execution environment .INDENT 7.0 .TP .B Parameters \fBrequired\fP (\fIint\fP) – .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Is_finalized .INDENT 0.0 .TP .B mpi4py.MPI.Is_finalized() Indicates whether \fBFinalize\fP has completed .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .SS mpi4py.MPI.Is_initialized .INDENT 0.0 .TP .B mpi4py.MPI.Is_initialized() Indicates whether \fBInit\fP has been called .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .SS mpi4py.MPI.Is_thread_main .INDENT 0.0 .TP .B mpi4py.MPI.Is_thread_main() Indicate whether this thread called \fBInit\fP or \fBInit_thread\fP .INDENT 7.0 .TP .B Return type bool .UNINDENT .UNINDENT .SS mpi4py.MPI.Lookup_name .INDENT 0.0 .TP .B mpi4py.MPI.Lookup_name(service_name, info=INFO_NULL) Lookup a port name given a service name .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBservice_name\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type str .UNINDENT .UNINDENT .SS mpi4py.MPI.Open_port .INDENT 0.0 .TP .B mpi4py.MPI.Open_port(info=INFO_NULL) Return an address that can be used to establish connections between groups of MPI processes .INDENT 7.0 .TP .B Parameters \fBinfo\fP (\fIInfo\fP) – .TP .B Return type str .UNINDENT .UNINDENT .SS mpi4py.MPI.Pcontrol .INDENT 0.0 .TP .B mpi4py.MPI.Pcontrol(level) Control profiling .INDENT 7.0 .TP .B Parameters \fBlevel\fP (\fIint\fP) – .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Publish_name .INDENT 0.0 .TP .B mpi4py.MPI.Publish_name(service_name, port_name, info=INFO_NULL) Publish a service name .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBservice_name\fP (\fIstr\fP) – .IP \(bu 2 \fBport_name\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Query_thread .INDENT 0.0 .TP .B mpi4py.MPI.Query_thread() Return the level of thread support provided by the MPI library .INDENT 7.0 .TP .B Return type int .UNINDENT .UNINDENT .SS mpi4py.MPI.Register_datarep .INDENT 0.0 .TP .B mpi4py.MPI.Register_datarep(datarep, read_fn, write_fn, extent_fn) Register user\-defined data representations .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBdatarep\fP (\fIstr\fP) – .IP \(bu 2 \fBread_fn\fP (\fICallable\fP\fI[\fP\fI[\fP\fIBuffer\fP\fI, \fP\fIDatatype\fP\fI, \fP\fIint\fP\fI, \fP\fIBuffer\fP\fI, \fP\fIint\fP\fI]\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBwrite_fn\fP (\fICallable\fP\fI[\fP\fI[\fP\fIBuffer\fP\fI, \fP\fIDatatype\fP\fI, \fP\fIint\fP\fI, \fP\fIBuffer\fP\fI, \fP\fIint\fP\fI]\fP\fI, \fP\fINone\fP\fI]\fP) – .IP \(bu 2 \fBextent_fn\fP (\fICallable\fP\fI[\fP\fI[\fP\fIDatatype\fP\fI]\fP\fI, \fP\fIint\fP\fI]\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Unpublish_name .INDENT 0.0 .TP .B mpi4py.MPI.Unpublish_name(service_name, port_name, info=INFO_NULL) Unpublish a service name .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBservice_name\fP (\fIstr\fP) – .IP \(bu 2 \fBport_name\fP (\fIstr\fP) – .IP \(bu 2 \fBinfo\fP (\fIInfo\fP) – .UNINDENT .TP .B Return type None .UNINDENT .UNINDENT .SS mpi4py.MPI.Wtick .INDENT 0.0 .TP .B mpi4py.MPI.Wtick() Return the resolution of \fBWtime\fP .INDENT 7.0 .TP .B Return type float .UNINDENT .UNINDENT .SS mpi4py.MPI.Wtime .INDENT 0.0 .TP .B mpi4py.MPI.Wtime() Return an elapsed time on the calling processor .INDENT 7.0 .TP .B Return type float .UNINDENT .UNINDENT .SS mpi4py.MPI.get_vendor .INDENT 0.0 .TP .B mpi4py.MPI.get_vendor() Infomation about the underlying MPI implementation .INDENT 7.0 .TP .B Returns .INDENT 7.0 .IP \(bu 2 a string with the name of the MPI implementation .IP \(bu 2 an integer 3\-tuple version \fB(major, minor, micro)\fP .UNINDENT .TP .B Return type Tuple[str, Tuple[int, int, int]] .UNINDENT .UNINDENT .sp Attributes .TS center; |l|l|. _ T{ \fBUNDEFINED\fP T} T{ \fBint\fP \fBUNDEFINED\fP T} _ T{ \fBANY_SOURCE\fP T} T{ \fBint\fP \fBANY_SOURCE\fP T} _ T{ \fBANY_TAG\fP T} T{ \fBint\fP \fBANY_TAG\fP T} _ T{ \fBPROC_NULL\fP T} T{ \fBint\fP \fBPROC_NULL\fP T} _ T{ \fBROOT\fP T} T{ \fBint\fP \fBROOT\fP T} _ T{ \fBBOTTOM\fP T} T{ \fBBottom\fP \fBBOTTOM\fP T} _ T{ \fBIN_PLACE\fP T} T{ \fBInPlace\fP \fBIN_PLACE\fP T} _ T{ \fBKEYVAL_INVALID\fP T} T{ \fBint\fP \fBKEYVAL_INVALID\fP T} _ T{ \fBTAG_UB\fP T} T{ \fBint\fP \fBTAG_UB\fP T} _ T{ \fBHOST\fP T} T{ \fBint\fP \fBHOST\fP T} _ T{ \fBIO\fP T} T{ \fBint\fP \fBIO\fP T} _ T{ \fBWTIME_IS_GLOBAL\fP T} T{ \fBint\fP \fBWTIME_IS_GLOBAL\fP T} _ T{ \fBUNIVERSE_SIZE\fP T} T{ \fBint\fP \fBUNIVERSE_SIZE\fP T} _ T{ \fBAPPNUM\fP T} T{ \fBint\fP \fBAPPNUM\fP T} _ T{ \fBLASTUSEDCODE\fP T} T{ \fBint\fP \fBLASTUSEDCODE\fP T} _ T{ \fBWIN_BASE\fP T} T{ \fBint\fP \fBWIN_BASE\fP T} _ T{ \fBWIN_SIZE\fP T} T{ \fBint\fP \fBWIN_SIZE\fP T} _ T{ \fBWIN_DISP_UNIT\fP T} T{ \fBint\fP \fBWIN_DISP_UNIT\fP T} _ T{ \fBWIN_CREATE_FLAVOR\fP T} T{ \fBint\fP \fBWIN_CREATE_FLAVOR\fP T} _ T{ \fBWIN_FLAVOR\fP T} T{ \fBint\fP \fBWIN_FLAVOR\fP T} _ T{ \fBWIN_MODEL\fP T} T{ \fBint\fP \fBWIN_MODEL\fP T} _ T{ \fBSUCCESS\fP T} T{ \fBint\fP \fBSUCCESS\fP T} _ T{ \fBERR_LASTCODE\fP T} T{ \fBint\fP \fBERR_LASTCODE\fP T} _ T{ \fBERR_COMM\fP T} T{ \fBint\fP \fBERR_COMM\fP T} _ T{ \fBERR_GROUP\fP T} T{ \fBint\fP \fBERR_GROUP\fP T} _ T{ \fBERR_TYPE\fP T} T{ \fBint\fP \fBERR_TYPE\fP T} _ T{ \fBERR_REQUEST\fP T} T{ \fBint\fP \fBERR_REQUEST\fP T} _ T{ \fBERR_OP\fP T} T{ \fBint\fP \fBERR_OP\fP T} _ T{ \fBERR_BUFFER\fP T} T{ \fBint\fP \fBERR_BUFFER\fP T} _ T{ \fBERR_COUNT\fP T} T{ \fBint\fP \fBERR_COUNT\fP T} _ T{ \fBERR_TAG\fP T} T{ \fBint\fP \fBERR_TAG\fP T} _ T{ \fBERR_RANK\fP T} T{ \fBint\fP \fBERR_RANK\fP T} _ T{ \fBERR_ROOT\fP T} T{ \fBint\fP \fBERR_ROOT\fP T} _ T{ \fBERR_TRUNCATE\fP T} T{ \fBint\fP \fBERR_TRUNCATE\fP T} _ T{ \fBERR_IN_STATUS\fP T} T{ \fBint\fP \fBERR_IN_STATUS\fP T} _ T{ \fBERR_PENDING\fP T} T{ \fBint\fP \fBERR_PENDING\fP T} _ T{ \fBERR_TOPOLOGY\fP T} T{ \fBint\fP \fBERR_TOPOLOGY\fP T} _ T{ \fBERR_DIMS\fP T} T{ \fBint\fP \fBERR_DIMS\fP T} _ T{ \fBERR_ARG\fP T} T{ \fBint\fP \fBERR_ARG\fP T} _ T{ \fBERR_OTHER\fP T} T{ \fBint\fP \fBERR_OTHER\fP T} _ T{ \fBERR_UNKNOWN\fP T} T{ \fBint\fP \fBERR_UNKNOWN\fP T} _ T{ \fBERR_INTERN\fP T} T{ \fBint\fP \fBERR_INTERN\fP T} _ T{ \fBERR_INFO\fP T} T{ \fBint\fP \fBERR_INFO\fP T} _ T{ \fBERR_FILE\fP T} T{ \fBint\fP \fBERR_FILE\fP T} _ T{ \fBERR_WIN\fP T} T{ \fBint\fP \fBERR_WIN\fP T} _ T{ \fBERR_KEYVAL\fP T} T{ \fBint\fP \fBERR_KEYVAL\fP T} _ T{ \fBERR_INFO_KEY\fP T} T{ \fBint\fP \fBERR_INFO_KEY\fP T} _ T{ \fBERR_INFO_VALUE\fP T} T{ \fBint\fP \fBERR_INFO_VALUE\fP T} _ T{ \fBERR_INFO_NOKEY\fP T} T{ \fBint\fP \fBERR_INFO_NOKEY\fP T} _ T{ \fBERR_ACCESS\fP T} T{ \fBint\fP \fBERR_ACCESS\fP T} _ T{ \fBERR_AMODE\fP T} T{ \fBint\fP \fBERR_AMODE\fP T} _ T{ \fBERR_BAD_FILE\fP T} T{ \fBint\fP \fBERR_BAD_FILE\fP T} _ T{ \fBERR_FILE_EXISTS\fP T} T{ \fBint\fP \fBERR_FILE_EXISTS\fP T} _ T{ \fBERR_FILE_IN_USE\fP T} T{ \fBint\fP \fBERR_FILE_IN_USE\fP T} _ T{ \fBERR_NO_SPACE\fP T} T{ \fBint\fP \fBERR_NO_SPACE\fP T} _ T{ \fBERR_NO_SUCH_FILE\fP T} T{ \fBint\fP \fBERR_NO_SUCH_FILE\fP T} _ T{ \fBERR_IO\fP T} T{ \fBint\fP \fBERR_IO\fP T} _ T{ \fBERR_READ_ONLY\fP T} T{ \fBint\fP \fBERR_READ_ONLY\fP T} _ T{ \fBERR_CONVERSION\fP T} T{ \fBint\fP \fBERR_CONVERSION\fP T} _ T{ \fBERR_DUP_DATAREP\fP T} T{ \fBint\fP \fBERR_DUP_DATAREP\fP T} _ T{ \fBERR_UNSUPPORTED_DATAREP\fP T} T{ \fBint\fP \fBERR_UNSUPPORTED_DATAREP\fP T} _ T{ \fBERR_UNSUPPORTED_OPERATION\fP T} T{ \fBint\fP \fBERR_UNSUPPORTED_OPERATION\fP T} _ T{ \fBERR_NAME\fP T} T{ \fBint\fP \fBERR_NAME\fP T} _ T{ \fBERR_NO_MEM\fP T} T{ \fBint\fP \fBERR_NO_MEM\fP T} _ T{ \fBERR_NOT_SAME\fP T} T{ \fBint\fP \fBERR_NOT_SAME\fP T} _ T{ \fBERR_PORT\fP T} T{ \fBint\fP \fBERR_PORT\fP T} _ T{ \fBERR_QUOTA\fP T} T{ \fBint\fP \fBERR_QUOTA\fP T} _ T{ \fBERR_SERVICE\fP T} T{ \fBint\fP \fBERR_SERVICE\fP T} _ T{ \fBERR_SPAWN\fP T} T{ \fBint\fP \fBERR_SPAWN\fP T} _ T{ \fBERR_BASE\fP T} T{ \fBint\fP \fBERR_BASE\fP T} _ T{ \fBERR_SIZE\fP T} T{ \fBint\fP \fBERR_SIZE\fP T} _ T{ \fBERR_DISP\fP T} T{ \fBint\fP \fBERR_DISP\fP T} _ T{ \fBERR_ASSERT\fP T} T{ \fBint\fP \fBERR_ASSERT\fP T} _ T{ \fBERR_LOCKTYPE\fP T} T{ \fBint\fP \fBERR_LOCKTYPE\fP T} _ T{ \fBERR_RMA_CONFLICT\fP T} T{ \fBint\fP \fBERR_RMA_CONFLICT\fP T} _ T{ \fBERR_RMA_SYNC\fP T} T{ \fBint\fP \fBERR_RMA_SYNC\fP T} _ T{ \fBERR_RMA_RANGE\fP T} T{ \fBint\fP \fBERR_RMA_RANGE\fP T} _ T{ \fBERR_RMA_ATTACH\fP T} T{ \fBint\fP \fBERR_RMA_ATTACH\fP T} _ T{ \fBERR_RMA_SHARED\fP T} T{ \fBint\fP \fBERR_RMA_SHARED\fP T} _ T{ \fBERR_RMA_FLAVOR\fP T} T{ \fBint\fP \fBERR_RMA_FLAVOR\fP T} _ T{ \fBORDER_C\fP T} T{ \fBint\fP \fBORDER_C\fP T} _ T{ \fBORDER_FORTRAN\fP T} T{ \fBint\fP \fBORDER_FORTRAN\fP T} _ T{ \fBORDER_F\fP T} T{ \fBint\fP \fBORDER_F\fP T} _ T{ \fBTYPECLASS_INTEGER\fP T} T{ \fBint\fP \fBTYPECLASS_INTEGER\fP T} _ T{ \fBTYPECLASS_REAL\fP T} T{ \fBint\fP \fBTYPECLASS_REAL\fP T} _ T{ \fBTYPECLASS_COMPLEX\fP T} T{ \fBint\fP \fBTYPECLASS_COMPLEX\fP T} _ T{ \fBDISTRIBUTE_NONE\fP T} T{ \fBint\fP \fBDISTRIBUTE_NONE\fP T} _ T{ \fBDISTRIBUTE_BLOCK\fP T} T{ \fBint\fP \fBDISTRIBUTE_BLOCK\fP T} _ T{ \fBDISTRIBUTE_CYCLIC\fP T} T{ \fBint\fP \fBDISTRIBUTE_CYCLIC\fP T} _ T{ \fBDISTRIBUTE_DFLT_DARG\fP T} T{ \fBint\fP \fBDISTRIBUTE_DFLT_DARG\fP T} _ T{ \fBCOMBINER_NAMED\fP T} T{ \fBint\fP \fBCOMBINER_NAMED\fP T} _ T{ \fBCOMBINER_DUP\fP T} T{ \fBint\fP \fBCOMBINER_DUP\fP T} _ T{ \fBCOMBINER_CONTIGUOUS\fP T} T{ \fBint\fP \fBCOMBINER_CONTIGUOUS\fP T} _ T{ \fBCOMBINER_VECTOR\fP T} T{ \fBint\fP \fBCOMBINER_VECTOR\fP T} _ T{ \fBCOMBINER_HVECTOR\fP T} T{ \fBint\fP \fBCOMBINER_HVECTOR\fP T} _ T{ \fBCOMBINER_INDEXED\fP T} T{ \fBint\fP \fBCOMBINER_INDEXED\fP T} _ T{ \fBCOMBINER_HINDEXED\fP T} T{ \fBint\fP \fBCOMBINER_HINDEXED\fP T} _ T{ \fBCOMBINER_INDEXED_BLOCK\fP T} T{ \fBint\fP \fBCOMBINER_INDEXED_BLOCK\fP T} _ T{ \fBCOMBINER_HINDEXED_BLOCK\fP T} T{ \fBint\fP \fBCOMBINER_HINDEXED_BLOCK\fP T} _ T{ \fBCOMBINER_STRUCT\fP T} T{ \fBint\fP \fBCOMBINER_STRUCT\fP T} _ T{ \fBCOMBINER_SUBARRAY\fP T} T{ \fBint\fP \fBCOMBINER_SUBARRAY\fP T} _ T{ \fBCOMBINER_DARRAY\fP T} T{ \fBint\fP \fBCOMBINER_DARRAY\fP T} _ T{ \fBCOMBINER_RESIZED\fP T} T{ \fBint\fP \fBCOMBINER_RESIZED\fP T} _ T{ \fBCOMBINER_F90_REAL\fP T} T{ \fBint\fP \fBCOMBINER_F90_REAL\fP T} _ T{ \fBCOMBINER_F90_COMPLEX\fP T} T{ \fBint\fP \fBCOMBINER_F90_COMPLEX\fP T} _ T{ \fBCOMBINER_F90_INTEGER\fP T} T{ \fBint\fP \fBCOMBINER_F90_INTEGER\fP T} _ T{ \fBIDENT\fP T} T{ \fBint\fP \fBIDENT\fP T} _ T{ \fBCONGRUENT\fP T} T{ \fBint\fP \fBCONGRUENT\fP T} _ T{ \fBSIMILAR\fP T} T{ \fBint\fP \fBSIMILAR\fP T} _ T{ \fBUNEQUAL\fP T} T{ \fBint\fP \fBUNEQUAL\fP T} _ T{ \fBCART\fP T} T{ \fBint\fP \fBCART\fP T} _ T{ \fBGRAPH\fP T} T{ \fBint\fP \fBGRAPH\fP T} _ T{ \fBDIST_GRAPH\fP T} T{ \fBint\fP \fBDIST_GRAPH\fP T} _ T{ \fBUNWEIGHTED\fP T} T{ \fBint\fP \fBUNWEIGHTED\fP T} _ T{ \fBWEIGHTS_EMPTY\fP T} T{ \fBint\fP \fBWEIGHTS_EMPTY\fP T} _ T{ \fBCOMM_TYPE_SHARED\fP T} T{ \fBint\fP \fBCOMM_TYPE_SHARED\fP T} _ T{ \fBBSEND_OVERHEAD\fP T} T{ \fBint\fP \fBBSEND_OVERHEAD\fP T} _ T{ \fBWIN_FLAVOR_CREATE\fP T} T{ \fBint\fP \fBWIN_FLAVOR_CREATE\fP T} _ T{ \fBWIN_FLAVOR_ALLOCATE\fP T} T{ \fBint\fP \fBWIN_FLAVOR_ALLOCATE\fP T} _ T{ \fBWIN_FLAVOR_DYNAMIC\fP T} T{ \fBint\fP \fBWIN_FLAVOR_DYNAMIC\fP T} _ T{ \fBWIN_FLAVOR_SHARED\fP T} T{ \fBint\fP \fBWIN_FLAVOR_SHARED\fP T} _ T{ \fBWIN_SEPARATE\fP T} T{ \fBint\fP \fBWIN_SEPARATE\fP T} _ T{ \fBWIN_UNIFIED\fP T} T{ \fBint\fP \fBWIN_UNIFIED\fP T} _ T{ \fBMODE_NOCHECK\fP T} T{ \fBint\fP \fBMODE_NOCHECK\fP T} _ T{ \fBMODE_NOSTORE\fP T} T{ \fBint\fP \fBMODE_NOSTORE\fP T} _ T{ \fBMODE_NOPUT\fP T} T{ \fBint\fP \fBMODE_NOPUT\fP T} _ T{ \fBMODE_NOPRECEDE\fP T} T{ \fBint\fP \fBMODE_NOPRECEDE\fP T} _ T{ \fBMODE_NOSUCCEED\fP T} T{ \fBint\fP \fBMODE_NOSUCCEED\fP T} _ T{ \fBLOCK_EXCLUSIVE\fP T} T{ \fBint\fP \fBLOCK_EXCLUSIVE\fP T} _ T{ \fBLOCK_SHARED\fP T} T{ \fBint\fP \fBLOCK_SHARED\fP T} _ T{ \fBMODE_RDONLY\fP T} T{ \fBint\fP \fBMODE_RDONLY\fP T} _ T{ \fBMODE_WRONLY\fP T} T{ \fBint\fP \fBMODE_WRONLY\fP T} _ T{ \fBMODE_RDWR\fP T} T{ \fBint\fP \fBMODE_RDWR\fP T} _ T{ \fBMODE_CREATE\fP T} T{ \fBint\fP \fBMODE_CREATE\fP T} _ T{ \fBMODE_EXCL\fP T} T{ \fBint\fP \fBMODE_EXCL\fP T} _ T{ \fBMODE_DELETE_ON_CLOSE\fP T} T{ \fBint\fP \fBMODE_DELETE_ON_CLOSE\fP T} _ T{ \fBMODE_UNIQUE_OPEN\fP T} T{ \fBint\fP \fBMODE_UNIQUE_OPEN\fP T} _ T{ \fBMODE_SEQUENTIAL\fP T} T{ \fBint\fP \fBMODE_SEQUENTIAL\fP T} _ T{ \fBMODE_APPEND\fP T} T{ \fBint\fP \fBMODE_APPEND\fP T} _ T{ \fBSEEK_SET\fP T} T{ \fBint\fP \fBSEEK_SET\fP T} _ T{ \fBSEEK_CUR\fP T} T{ \fBint\fP \fBSEEK_CUR\fP T} _ T{ \fBSEEK_END\fP T} T{ \fBint\fP \fBSEEK_END\fP T} _ T{ \fBDISPLACEMENT_CURRENT\fP T} T{ \fBint\fP \fBDISPLACEMENT_CURRENT\fP T} _ T{ \fBDISP_CUR\fP T} T{ \fBint\fP \fBDISP_CUR\fP T} _ T{ \fBTHREAD_SINGLE\fP T} T{ \fBint\fP \fBTHREAD_SINGLE\fP T} _ T{ \fBTHREAD_FUNNELED\fP T} T{ \fBint\fP \fBTHREAD_FUNNELED\fP T} _ T{ \fBTHREAD_SERIALIZED\fP T} T{ \fBint\fP \fBTHREAD_SERIALIZED\fP T} _ T{ \fBTHREAD_MULTIPLE\fP T} T{ \fBint\fP \fBTHREAD_MULTIPLE\fP T} _ T{ \fBVERSION\fP T} T{ \fBint\fP \fBVERSION\fP T} _ T{ \fBSUBVERSION\fP T} T{ \fBint\fP \fBSUBVERSION\fP T} _ T{ \fBMAX_PROCESSOR_NAME\fP T} T{ \fBint\fP \fBMAX_PROCESSOR_NAME\fP T} _ T{ \fBMAX_ERROR_STRING\fP T} T{ \fBint\fP \fBMAX_ERROR_STRING\fP T} _ T{ \fBMAX_PORT_NAME\fP T} T{ \fBint\fP \fBMAX_PORT_NAME\fP T} _ T{ \fBMAX_INFO_KEY\fP T} T{ \fBint\fP \fBMAX_INFO_KEY\fP T} _ T{ \fBMAX_INFO_VAL\fP T} T{ \fBint\fP \fBMAX_INFO_VAL\fP T} _ T{ \fBMAX_OBJECT_NAME\fP T} T{ \fBint\fP \fBMAX_OBJECT_NAME\fP T} _ T{ \fBMAX_DATAREP_STRING\fP T} T{ \fBint\fP \fBMAX_DATAREP_STRING\fP T} _ T{ \fBMAX_LIBRARY_VERSION_STRING\fP T} T{ \fBint\fP \fBMAX_LIBRARY_VERSION_STRING\fP T} _ T{ \fBDATATYPE_NULL\fP T} T{ \fBDatatype\fP \fBDATATYPE_NULL\fP T} _ T{ \fBUB\fP T} T{ \fBDatatype\fP \fBUB\fP T} _ T{ \fBLB\fP T} T{ \fBDatatype\fP \fBLB\fP T} _ T{ \fBPACKED\fP T} T{ \fBDatatype\fP \fBPACKED\fP T} _ T{ \fBBYTE\fP T} T{ \fBDatatype\fP \fBBYTE\fP T} _ T{ \fBAINT\fP T} T{ \fBDatatype\fP \fBAINT\fP T} _ T{ \fBOFFSET\fP T} T{ \fBDatatype\fP \fBOFFSET\fP T} _ T{ \fBCOUNT\fP T} T{ \fBDatatype\fP \fBCOUNT\fP T} _ T{ \fBCHAR\fP T} T{ \fBDatatype\fP \fBCHAR\fP T} _ T{ \fBWCHAR\fP T} T{ \fBDatatype\fP \fBWCHAR\fP T} _ T{ \fBSIGNED_CHAR\fP T} T{ \fBDatatype\fP \fBSIGNED_CHAR\fP T} _ T{ \fBSHORT\fP T} T{ \fBDatatype\fP \fBSHORT\fP T} _ T{ \fBINT\fP T} T{ \fBDatatype\fP \fBINT\fP T} _ T{ \fBLONG\fP T} T{ \fBDatatype\fP \fBLONG\fP T} _ T{ \fBLONG_LONG\fP T} T{ \fBDatatype\fP \fBLONG_LONG\fP T} _ T{ \fBUNSIGNED_CHAR\fP T} T{ \fBDatatype\fP \fBUNSIGNED_CHAR\fP T} _ T{ \fBUNSIGNED_SHORT\fP T} T{ \fBDatatype\fP \fBUNSIGNED_SHORT\fP T} _ T{ \fBUNSIGNED\fP T} T{ \fBDatatype\fP \fBUNSIGNED\fP T} _ T{ \fBUNSIGNED_LONG\fP T} T{ \fBDatatype\fP \fBUNSIGNED_LONG\fP T} _ T{ \fBUNSIGNED_LONG_LONG\fP T} T{ \fBDatatype\fP \fBUNSIGNED_LONG_LONG\fP T} _ T{ \fBFLOAT\fP T} T{ \fBDatatype\fP \fBFLOAT\fP T} _ T{ \fBDOUBLE\fP T} T{ \fBDatatype\fP \fBDOUBLE\fP T} _ T{ \fBLONG_DOUBLE\fP T} T{ \fBDatatype\fP \fBLONG_DOUBLE\fP T} _ T{ \fBC_BOOL\fP T} T{ \fBDatatype\fP \fBC_BOOL\fP T} _ T{ \fBINT8_T\fP T} T{ \fBDatatype\fP \fBINT8_T\fP T} _ T{ \fBINT16_T\fP T} T{ \fBDatatype\fP \fBINT16_T\fP T} _ T{ \fBINT32_T\fP T} T{ \fBDatatype\fP \fBINT32_T\fP T} _ T{ \fBINT64_T\fP T} T{ \fBDatatype\fP \fBINT64_T\fP T} _ T{ \fBUINT8_T\fP T} T{ \fBDatatype\fP \fBUINT8_T\fP T} _ T{ \fBUINT16_T\fP T} T{ \fBDatatype\fP \fBUINT16_T\fP T} _ T{ \fBUINT32_T\fP T} T{ \fBDatatype\fP \fBUINT32_T\fP T} _ T{ \fBUINT64_T\fP T} T{ \fBDatatype\fP \fBUINT64_T\fP T} _ T{ \fBC_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_COMPLEX\fP T} _ T{ \fBC_FLOAT_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_FLOAT_COMPLEX\fP T} _ T{ \fBC_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_DOUBLE_COMPLEX\fP T} _ T{ \fBC_LONG_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBC_LONG_DOUBLE_COMPLEX\fP T} _ T{ \fBCXX_BOOL\fP T} T{ \fBDatatype\fP \fBCXX_BOOL\fP T} _ T{ \fBCXX_FLOAT_COMPLEX\fP T} T{ \fBDatatype\fP \fBCXX_FLOAT_COMPLEX\fP T} _ T{ \fBCXX_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBCXX_DOUBLE_COMPLEX\fP T} _ T{ \fBCXX_LONG_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBCXX_LONG_DOUBLE_COMPLEX\fP T} _ T{ \fBSHORT_INT\fP T} T{ \fBDatatype\fP \fBSHORT_INT\fP T} _ T{ \fBINT_INT\fP T} T{ \fBDatatype\fP \fBINT_INT\fP T} _ T{ \fBTWOINT\fP T} T{ \fBDatatype\fP \fBTWOINT\fP T} _ T{ \fBLONG_INT\fP T} T{ \fBDatatype\fP \fBLONG_INT\fP T} _ T{ \fBFLOAT_INT\fP T} T{ \fBDatatype\fP \fBFLOAT_INT\fP T} _ T{ \fBDOUBLE_INT\fP T} T{ \fBDatatype\fP \fBDOUBLE_INT\fP T} _ T{ \fBLONG_DOUBLE_INT\fP T} T{ \fBDatatype\fP \fBLONG_DOUBLE_INT\fP T} _ T{ \fBCHARACTER\fP T} T{ \fBDatatype\fP \fBCHARACTER\fP T} _ T{ \fBLOGICAL\fP T} T{ \fBDatatype\fP \fBLOGICAL\fP T} _ T{ \fBINTEGER\fP T} T{ \fBDatatype\fP \fBINTEGER\fP T} _ T{ \fBREAL\fP T} T{ \fBDatatype\fP \fBREAL\fP T} _ T{ \fBDOUBLE_PRECISION\fP T} T{ \fBDatatype\fP \fBDOUBLE_PRECISION\fP T} _ T{ \fBCOMPLEX\fP T} T{ \fBDatatype\fP \fBCOMPLEX\fP T} _ T{ \fBDOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBDOUBLE_COMPLEX\fP T} _ T{ \fBLOGICAL1\fP T} T{ \fBDatatype\fP \fBLOGICAL1\fP T} _ T{ \fBLOGICAL2\fP T} T{ \fBDatatype\fP \fBLOGICAL2\fP T} _ T{ \fBLOGICAL4\fP T} T{ \fBDatatype\fP \fBLOGICAL4\fP T} _ T{ \fBLOGICAL8\fP T} T{ \fBDatatype\fP \fBLOGICAL8\fP T} _ T{ \fBINTEGER1\fP T} T{ \fBDatatype\fP \fBINTEGER1\fP T} _ T{ \fBINTEGER2\fP T} T{ \fBDatatype\fP \fBINTEGER2\fP T} _ T{ \fBINTEGER4\fP T} T{ \fBDatatype\fP \fBINTEGER4\fP T} _ T{ \fBINTEGER8\fP T} T{ \fBDatatype\fP \fBINTEGER8\fP T} _ T{ \fBINTEGER16\fP T} T{ \fBDatatype\fP \fBINTEGER16\fP T} _ T{ \fBREAL2\fP T} T{ \fBDatatype\fP \fBREAL2\fP T} _ T{ \fBREAL4\fP T} T{ \fBDatatype\fP \fBREAL4\fP T} _ T{ \fBREAL8\fP T} T{ \fBDatatype\fP \fBREAL8\fP T} _ T{ \fBREAL16\fP T} T{ \fBDatatype\fP \fBREAL16\fP T} _ T{ \fBCOMPLEX4\fP T} T{ \fBDatatype\fP \fBCOMPLEX4\fP T} _ T{ \fBCOMPLEX8\fP T} T{ \fBDatatype\fP \fBCOMPLEX8\fP T} _ T{ \fBCOMPLEX16\fP T} T{ \fBDatatype\fP \fBCOMPLEX16\fP T} _ T{ \fBCOMPLEX32\fP T} T{ \fBDatatype\fP \fBCOMPLEX32\fP T} _ T{ \fBUNSIGNED_INT\fP T} T{ \fBDatatype\fP \fBUNSIGNED_INT\fP T} _ T{ \fBSIGNED_SHORT\fP T} T{ \fBDatatype\fP \fBSIGNED_SHORT\fP T} _ T{ \fBSIGNED_INT\fP T} T{ \fBDatatype\fP \fBSIGNED_INT\fP T} _ T{ \fBSIGNED_LONG\fP T} T{ \fBDatatype\fP \fBSIGNED_LONG\fP T} _ T{ \fBSIGNED_LONG_LONG\fP T} T{ \fBDatatype\fP \fBSIGNED_LONG_LONG\fP T} _ T{ \fBBOOL\fP T} T{ \fBDatatype\fP \fBBOOL\fP T} _ T{ \fBSINT8_T\fP T} T{ \fBDatatype\fP \fBSINT8_T\fP T} _ T{ \fBSINT16_T\fP T} T{ \fBDatatype\fP \fBSINT16_T\fP T} _ T{ \fBSINT32_T\fP T} T{ \fBDatatype\fP \fBSINT32_T\fP T} _ T{ \fBSINT64_T\fP T} T{ \fBDatatype\fP \fBSINT64_T\fP T} _ T{ \fBF_BOOL\fP T} T{ \fBDatatype\fP \fBF_BOOL\fP T} _ T{ \fBF_INT\fP T} T{ \fBDatatype\fP \fBF_INT\fP T} _ T{ \fBF_FLOAT\fP T} T{ \fBDatatype\fP \fBF_FLOAT\fP T} _ T{ \fBF_DOUBLE\fP T} T{ \fBDatatype\fP \fBF_DOUBLE\fP T} _ T{ \fBF_COMPLEX\fP T} T{ \fBDatatype\fP \fBF_COMPLEX\fP T} _ T{ \fBF_FLOAT_COMPLEX\fP T} T{ \fBDatatype\fP \fBF_FLOAT_COMPLEX\fP T} _ T{ \fBF_DOUBLE_COMPLEX\fP T} T{ \fBDatatype\fP \fBF_DOUBLE_COMPLEX\fP T} _ T{ \fBREQUEST_NULL\fP T} T{ \fBRequest\fP \fBREQUEST_NULL\fP T} _ T{ \fBMESSAGE_NULL\fP T} T{ \fBMessage\fP \fBMESSAGE_NULL\fP T} _ T{ \fBMESSAGE_NO_PROC\fP T} T{ \fBMessage\fP \fBMESSAGE_NO_PROC\fP T} _ T{ \fBOP_NULL\fP T} T{ \fBOp\fP \fBOP_NULL\fP T} _ T{ \fBMAX\fP T} T{ \fBOp\fP \fBMAX\fP T} _ T{ \fBMIN\fP T} T{ \fBOp\fP \fBMIN\fP T} _ T{ \fBSUM\fP T} T{ \fBOp\fP \fBSUM\fP T} _ T{ \fBPROD\fP T} T{ \fBOp\fP \fBPROD\fP T} _ T{ \fBLAND\fP T} T{ \fBOp\fP \fBLAND\fP T} _ T{ \fBBAND\fP T} T{ \fBOp\fP \fBBAND\fP T} _ T{ \fBLOR\fP T} T{ \fBOp\fP \fBLOR\fP T} _ T{ \fBBOR\fP T} T{ \fBOp\fP \fBBOR\fP T} _ T{ \fBLXOR\fP T} T{ \fBOp\fP \fBLXOR\fP T} _ T{ \fBBXOR\fP T} T{ \fBOp\fP \fBBXOR\fP T} _ T{ \fBMAXLOC\fP T} T{ \fBOp\fP \fBMAXLOC\fP T} _ T{ \fBMINLOC\fP T} T{ \fBOp\fP \fBMINLOC\fP T} _ T{ \fBREPLACE\fP T} T{ \fBOp\fP \fBREPLACE\fP T} _ T{ \fBNO_OP\fP T} T{ \fBOp\fP \fBNO_OP\fP T} _ T{ \fBGROUP_NULL\fP T} T{ \fBGroup\fP \fBGROUP_NULL\fP T} _ T{ \fBGROUP_EMPTY\fP T} T{ \fBGroup\fP \fBGROUP_EMPTY\fP T} _ T{ \fBINFO_NULL\fP T} T{ \fBInfo\fP \fBINFO_NULL\fP T} _ T{ \fBINFO_ENV\fP T} T{ \fBInfo\fP \fBINFO_ENV\fP T} _ T{ \fBERRHANDLER_NULL\fP T} T{ \fBErrhandler\fP \fBERRHANDLER_NULL\fP T} _ T{ \fBERRORS_RETURN\fP T} T{ \fBErrhandler\fP \fBERRORS_RETURN\fP T} _ T{ \fBERRORS_ARE_FATAL\fP T} T{ \fBErrhandler\fP \fBERRORS_ARE_FATAL\fP T} _ T{ \fBCOMM_NULL\fP T} T{ \fBComm\fP \fBCOMM_NULL\fP T} _ T{ \fBCOMM_SELF\fP T} T{ \fBIntracomm\fP \fBCOMM_SELF\fP T} _ T{ \fBCOMM_WORLD\fP T} T{ \fBIntracomm\fP \fBCOMM_WORLD\fP T} _ T{ \fBWIN_NULL\fP T} T{ \fBWin\fP \fBWIN_NULL\fP T} _ T{ \fBFILE_NULL\fP T} T{ \fBFile\fP \fBFILE_NULL\fP T} _ T{ \fBpickle\fP T} T{ \fBPickle\fP \fBpickle\fP T} _ .TE .SS mpi4py.MPI.UNDEFINED .INDENT 0.0 .TP .B mpi4py.MPI.UNDEFINED: int = UNDEFINED \fBint\fP \fBUNDEFINED\fP .UNINDENT .SS mpi4py.MPI.ANY_SOURCE .INDENT 0.0 .TP .B mpi4py.MPI.ANY_SOURCE: int = ANY_SOURCE \fBint\fP \fBANY_SOURCE\fP .UNINDENT .SS mpi4py.MPI.ANY_TAG .INDENT 0.0 .TP .B mpi4py.MPI.ANY_TAG: int = ANY_TAG \fBint\fP \fBANY_TAG\fP .UNINDENT .SS mpi4py.MPI.PROC_NULL .INDENT 0.0 .TP .B mpi4py.MPI.PROC_NULL: int = PROC_NULL \fBint\fP \fBPROC_NULL\fP .UNINDENT .SS mpi4py.MPI.ROOT .INDENT 0.0 .TP .B mpi4py.MPI.ROOT: int = ROOT \fBint\fP \fBROOT\fP .UNINDENT .SS mpi4py.MPI.BOTTOM .INDENT 0.0 .TP .B mpi4py.MPI.BOTTOM: Bottom = BOTTOM \fBBottom\fP \fBBOTTOM\fP .UNINDENT .SS mpi4py.MPI.IN_PLACE .INDENT 0.0 .TP .B mpi4py.MPI.IN_PLACE: InPlace = IN_PLACE \fBInPlace\fP \fBIN_PLACE\fP .UNINDENT .SS mpi4py.MPI.KEYVAL_INVALID .INDENT 0.0 .TP .B mpi4py.MPI.KEYVAL_INVALID: int = KEYVAL_INVALID \fBint\fP \fBKEYVAL_INVALID\fP .UNINDENT .SS mpi4py.MPI.TAG_UB .INDENT 0.0 .TP .B mpi4py.MPI.TAG_UB: int = TAG_UB \fBint\fP \fBTAG_UB\fP .UNINDENT .SS mpi4py.MPI.HOST .INDENT 0.0 .TP .B mpi4py.MPI.HOST: int = HOST \fBint\fP \fBHOST\fP .UNINDENT .SS mpi4py.MPI.IO .INDENT 0.0 .TP .B mpi4py.MPI.IO: int = IO \fBint\fP \fBIO\fP .UNINDENT .SS mpi4py.MPI.WTIME_IS_GLOBAL .INDENT 0.0 .TP .B mpi4py.MPI.WTIME_IS_GLOBAL: int = WTIME_IS_GLOBAL \fBint\fP \fBWTIME_IS_GLOBAL\fP .UNINDENT .SS mpi4py.MPI.UNIVERSE_SIZE .INDENT 0.0 .TP .B mpi4py.MPI.UNIVERSE_SIZE: int = UNIVERSE_SIZE \fBint\fP \fBUNIVERSE_SIZE\fP .UNINDENT .SS mpi4py.MPI.APPNUM .INDENT 0.0 .TP .B mpi4py.MPI.APPNUM: int = APPNUM \fBint\fP \fBAPPNUM\fP .UNINDENT .SS mpi4py.MPI.LASTUSEDCODE .INDENT 0.0 .TP .B mpi4py.MPI.LASTUSEDCODE: int = LASTUSEDCODE \fBint\fP \fBLASTUSEDCODE\fP .UNINDENT .SS mpi4py.MPI.WIN_BASE .INDENT 0.0 .TP .B mpi4py.MPI.WIN_BASE: int = WIN_BASE \fBint\fP \fBWIN_BASE\fP .UNINDENT .SS mpi4py.MPI.WIN_SIZE .INDENT 0.0 .TP .B mpi4py.MPI.WIN_SIZE: int = WIN_SIZE \fBint\fP \fBWIN_SIZE\fP .UNINDENT .SS mpi4py.MPI.WIN_DISP_UNIT .INDENT 0.0 .TP .B mpi4py.MPI.WIN_DISP_UNIT: int = WIN_DISP_UNIT \fBint\fP \fBWIN_DISP_UNIT\fP .UNINDENT .SS mpi4py.MPI.WIN_CREATE_FLAVOR .INDENT 0.0 .TP .B mpi4py.MPI.WIN_CREATE_FLAVOR: int = WIN_CREATE_FLAVOR \fBint\fP \fBWIN_CREATE_FLAVOR\fP .UNINDENT .SS mpi4py.MPI.WIN_FLAVOR .INDENT 0.0 .TP .B mpi4py.MPI.WIN_FLAVOR: int = WIN_FLAVOR \fBint\fP \fBWIN_FLAVOR\fP .UNINDENT .SS mpi4py.MPI.WIN_MODEL .INDENT 0.0 .TP .B mpi4py.MPI.WIN_MODEL: int = WIN_MODEL \fBint\fP \fBWIN_MODEL\fP .UNINDENT .SS mpi4py.MPI.SUCCESS .INDENT 0.0 .TP .B mpi4py.MPI.SUCCESS: int = SUCCESS \fBint\fP \fBSUCCESS\fP .UNINDENT .SS mpi4py.MPI.ERR_LASTCODE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_LASTCODE: int = ERR_LASTCODE \fBint\fP \fBERR_LASTCODE\fP .UNINDENT .SS mpi4py.MPI.ERR_COMM .INDENT 0.0 .TP .B mpi4py.MPI.ERR_COMM: int = ERR_COMM \fBint\fP \fBERR_COMM\fP .UNINDENT .SS mpi4py.MPI.ERR_GROUP .INDENT 0.0 .TP .B mpi4py.MPI.ERR_GROUP: int = ERR_GROUP \fBint\fP \fBERR_GROUP\fP .UNINDENT .SS mpi4py.MPI.ERR_TYPE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_TYPE: int = ERR_TYPE \fBint\fP \fBERR_TYPE\fP .UNINDENT .SS mpi4py.MPI.ERR_REQUEST .INDENT 0.0 .TP .B mpi4py.MPI.ERR_REQUEST: int = ERR_REQUEST \fBint\fP \fBERR_REQUEST\fP .UNINDENT .SS mpi4py.MPI.ERR_OP .INDENT 0.0 .TP .B mpi4py.MPI.ERR_OP: int = ERR_OP \fBint\fP \fBERR_OP\fP .UNINDENT .SS mpi4py.MPI.ERR_BUFFER .INDENT 0.0 .TP .B mpi4py.MPI.ERR_BUFFER: int = ERR_BUFFER \fBint\fP \fBERR_BUFFER\fP .UNINDENT .SS mpi4py.MPI.ERR_COUNT .INDENT 0.0 .TP .B mpi4py.MPI.ERR_COUNT: int = ERR_COUNT \fBint\fP \fBERR_COUNT\fP .UNINDENT .SS mpi4py.MPI.ERR_TAG .INDENT 0.0 .TP .B mpi4py.MPI.ERR_TAG: int = ERR_TAG \fBint\fP \fBERR_TAG\fP .UNINDENT .SS mpi4py.MPI.ERR_RANK .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RANK: int = ERR_RANK \fBint\fP \fBERR_RANK\fP .UNINDENT .SS mpi4py.MPI.ERR_ROOT .INDENT 0.0 .TP .B mpi4py.MPI.ERR_ROOT: int = ERR_ROOT \fBint\fP \fBERR_ROOT\fP .UNINDENT .SS mpi4py.MPI.ERR_TRUNCATE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_TRUNCATE: int = ERR_TRUNCATE \fBint\fP \fBERR_TRUNCATE\fP .UNINDENT .SS mpi4py.MPI.ERR_IN_STATUS .INDENT 0.0 .TP .B mpi4py.MPI.ERR_IN_STATUS: int = ERR_IN_STATUS \fBint\fP \fBERR_IN_STATUS\fP .UNINDENT .SS mpi4py.MPI.ERR_PENDING .INDENT 0.0 .TP .B mpi4py.MPI.ERR_PENDING: int = ERR_PENDING \fBint\fP \fBERR_PENDING\fP .UNINDENT .SS mpi4py.MPI.ERR_TOPOLOGY .INDENT 0.0 .TP .B mpi4py.MPI.ERR_TOPOLOGY: int = ERR_TOPOLOGY \fBint\fP \fBERR_TOPOLOGY\fP .UNINDENT .SS mpi4py.MPI.ERR_DIMS .INDENT 0.0 .TP .B mpi4py.MPI.ERR_DIMS: int = ERR_DIMS \fBint\fP \fBERR_DIMS\fP .UNINDENT .SS mpi4py.MPI.ERR_ARG .INDENT 0.0 .TP .B mpi4py.MPI.ERR_ARG: int = ERR_ARG \fBint\fP \fBERR_ARG\fP .UNINDENT .SS mpi4py.MPI.ERR_OTHER .INDENT 0.0 .TP .B mpi4py.MPI.ERR_OTHER: int = ERR_OTHER \fBint\fP \fBERR_OTHER\fP .UNINDENT .SS mpi4py.MPI.ERR_UNKNOWN .INDENT 0.0 .TP .B mpi4py.MPI.ERR_UNKNOWN: int = ERR_UNKNOWN \fBint\fP \fBERR_UNKNOWN\fP .UNINDENT .SS mpi4py.MPI.ERR_INTERN .INDENT 0.0 .TP .B mpi4py.MPI.ERR_INTERN: int = ERR_INTERN \fBint\fP \fBERR_INTERN\fP .UNINDENT .SS mpi4py.MPI.ERR_INFO .INDENT 0.0 .TP .B mpi4py.MPI.ERR_INFO: int = ERR_INFO \fBint\fP \fBERR_INFO\fP .UNINDENT .SS mpi4py.MPI.ERR_FILE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_FILE: int = ERR_FILE \fBint\fP \fBERR_FILE\fP .UNINDENT .SS mpi4py.MPI.ERR_WIN .INDENT 0.0 .TP .B mpi4py.MPI.ERR_WIN: int = ERR_WIN \fBint\fP \fBERR_WIN\fP .UNINDENT .SS mpi4py.MPI.ERR_KEYVAL .INDENT 0.0 .TP .B mpi4py.MPI.ERR_KEYVAL: int = ERR_KEYVAL \fBint\fP \fBERR_KEYVAL\fP .UNINDENT .SS mpi4py.MPI.ERR_INFO_KEY .INDENT 0.0 .TP .B mpi4py.MPI.ERR_INFO_KEY: int = ERR_INFO_KEY \fBint\fP \fBERR_INFO_KEY\fP .UNINDENT .SS mpi4py.MPI.ERR_INFO_VALUE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_INFO_VALUE: int = ERR_INFO_VALUE \fBint\fP \fBERR_INFO_VALUE\fP .UNINDENT .SS mpi4py.MPI.ERR_INFO_NOKEY .INDENT 0.0 .TP .B mpi4py.MPI.ERR_INFO_NOKEY: int = ERR_INFO_NOKEY \fBint\fP \fBERR_INFO_NOKEY\fP .UNINDENT .SS mpi4py.MPI.ERR_ACCESS .INDENT 0.0 .TP .B mpi4py.MPI.ERR_ACCESS: int = ERR_ACCESS \fBint\fP \fBERR_ACCESS\fP .UNINDENT .SS mpi4py.MPI.ERR_AMODE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_AMODE: int = ERR_AMODE \fBint\fP \fBERR_AMODE\fP .UNINDENT .SS mpi4py.MPI.ERR_BAD_FILE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_BAD_FILE: int = ERR_BAD_FILE \fBint\fP \fBERR_BAD_FILE\fP .UNINDENT .SS mpi4py.MPI.ERR_FILE_EXISTS .INDENT 0.0 .TP .B mpi4py.MPI.ERR_FILE_EXISTS: int = ERR_FILE_EXISTS \fBint\fP \fBERR_FILE_EXISTS\fP .UNINDENT .SS mpi4py.MPI.ERR_FILE_IN_USE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_FILE_IN_USE: int = ERR_FILE_IN_USE \fBint\fP \fBERR_FILE_IN_USE\fP .UNINDENT .SS mpi4py.MPI.ERR_NO_SPACE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_NO_SPACE: int = ERR_NO_SPACE \fBint\fP \fBERR_NO_SPACE\fP .UNINDENT .SS mpi4py.MPI.ERR_NO_SUCH_FILE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_NO_SUCH_FILE: int = ERR_NO_SUCH_FILE \fBint\fP \fBERR_NO_SUCH_FILE\fP .UNINDENT .SS mpi4py.MPI.ERR_IO .INDENT 0.0 .TP .B mpi4py.MPI.ERR_IO: int = ERR_IO \fBint\fP \fBERR_IO\fP .UNINDENT .SS mpi4py.MPI.ERR_READ_ONLY .INDENT 0.0 .TP .B mpi4py.MPI.ERR_READ_ONLY: int = ERR_READ_ONLY \fBint\fP \fBERR_READ_ONLY\fP .UNINDENT .SS mpi4py.MPI.ERR_CONVERSION .INDENT 0.0 .TP .B mpi4py.MPI.ERR_CONVERSION: int = ERR_CONVERSION \fBint\fP \fBERR_CONVERSION\fP .UNINDENT .SS mpi4py.MPI.ERR_DUP_DATAREP .INDENT 0.0 .TP .B mpi4py.MPI.ERR_DUP_DATAREP: int = ERR_DUP_DATAREP \fBint\fP \fBERR_DUP_DATAREP\fP .UNINDENT .SS mpi4py.MPI.ERR_UNSUPPORTED_DATAREP .INDENT 0.0 .TP .B mpi4py.MPI.ERR_UNSUPPORTED_DATAREP: int = ERR_UNSUPPORTED_DATAREP \fBint\fP \fBERR_UNSUPPORTED_DATAREP\fP .UNINDENT .SS mpi4py.MPI.ERR_UNSUPPORTED_OPERATION .INDENT 0.0 .TP .B mpi4py.MPI.ERR_UNSUPPORTED_OPERATION: int = ERR_UNSUPPORTED_OPERATION \fBint\fP \fBERR_UNSUPPORTED_OPERATION\fP .UNINDENT .SS mpi4py.MPI.ERR_NAME .INDENT 0.0 .TP .B mpi4py.MPI.ERR_NAME: int = ERR_NAME \fBint\fP \fBERR_NAME\fP .UNINDENT .SS mpi4py.MPI.ERR_NO_MEM .INDENT 0.0 .TP .B mpi4py.MPI.ERR_NO_MEM: int = ERR_NO_MEM \fBint\fP \fBERR_NO_MEM\fP .UNINDENT .SS mpi4py.MPI.ERR_NOT_SAME .INDENT 0.0 .TP .B mpi4py.MPI.ERR_NOT_SAME: int = ERR_NOT_SAME \fBint\fP \fBERR_NOT_SAME\fP .UNINDENT .SS mpi4py.MPI.ERR_PORT .INDENT 0.0 .TP .B mpi4py.MPI.ERR_PORT: int = ERR_PORT \fBint\fP \fBERR_PORT\fP .UNINDENT .SS mpi4py.MPI.ERR_QUOTA .INDENT 0.0 .TP .B mpi4py.MPI.ERR_QUOTA: int = ERR_QUOTA \fBint\fP \fBERR_QUOTA\fP .UNINDENT .SS mpi4py.MPI.ERR_SERVICE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_SERVICE: int = ERR_SERVICE \fBint\fP \fBERR_SERVICE\fP .UNINDENT .SS mpi4py.MPI.ERR_SPAWN .INDENT 0.0 .TP .B mpi4py.MPI.ERR_SPAWN: int = ERR_SPAWN \fBint\fP \fBERR_SPAWN\fP .UNINDENT .SS mpi4py.MPI.ERR_BASE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_BASE: int = ERR_BASE \fBint\fP \fBERR_BASE\fP .UNINDENT .SS mpi4py.MPI.ERR_SIZE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_SIZE: int = ERR_SIZE \fBint\fP \fBERR_SIZE\fP .UNINDENT .SS mpi4py.MPI.ERR_DISP .INDENT 0.0 .TP .B mpi4py.MPI.ERR_DISP: int = ERR_DISP \fBint\fP \fBERR_DISP\fP .UNINDENT .SS mpi4py.MPI.ERR_ASSERT .INDENT 0.0 .TP .B mpi4py.MPI.ERR_ASSERT: int = ERR_ASSERT \fBint\fP \fBERR_ASSERT\fP .UNINDENT .SS mpi4py.MPI.ERR_LOCKTYPE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_LOCKTYPE: int = ERR_LOCKTYPE \fBint\fP \fBERR_LOCKTYPE\fP .UNINDENT .SS mpi4py.MPI.ERR_RMA_CONFLICT .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RMA_CONFLICT: int = ERR_RMA_CONFLICT \fBint\fP \fBERR_RMA_CONFLICT\fP .UNINDENT .SS mpi4py.MPI.ERR_RMA_SYNC .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RMA_SYNC: int = ERR_RMA_SYNC \fBint\fP \fBERR_RMA_SYNC\fP .UNINDENT .SS mpi4py.MPI.ERR_RMA_RANGE .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RMA_RANGE: int = ERR_RMA_RANGE \fBint\fP \fBERR_RMA_RANGE\fP .UNINDENT .SS mpi4py.MPI.ERR_RMA_ATTACH .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RMA_ATTACH: int = ERR_RMA_ATTACH \fBint\fP \fBERR_RMA_ATTACH\fP .UNINDENT .SS mpi4py.MPI.ERR_RMA_SHARED .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RMA_SHARED: int = ERR_RMA_SHARED \fBint\fP \fBERR_RMA_SHARED\fP .UNINDENT .SS mpi4py.MPI.ERR_RMA_FLAVOR .INDENT 0.0 .TP .B mpi4py.MPI.ERR_RMA_FLAVOR: int = ERR_RMA_FLAVOR \fBint\fP \fBERR_RMA_FLAVOR\fP .UNINDENT .SS mpi4py.MPI.ORDER_C .INDENT 0.0 .TP .B mpi4py.MPI.ORDER_C: int = ORDER_C \fBint\fP \fBORDER_C\fP .UNINDENT .SS mpi4py.MPI.ORDER_FORTRAN .INDENT 0.0 .TP .B mpi4py.MPI.ORDER_FORTRAN: int = ORDER_FORTRAN \fBint\fP \fBORDER_FORTRAN\fP .UNINDENT .SS mpi4py.MPI.ORDER_F .INDENT 0.0 .TP .B mpi4py.MPI.ORDER_F: int = ORDER_F \fBint\fP \fBORDER_F\fP .UNINDENT .SS mpi4py.MPI.TYPECLASS_INTEGER .INDENT 0.0 .TP .B mpi4py.MPI.TYPECLASS_INTEGER: int = TYPECLASS_INTEGER \fBint\fP \fBTYPECLASS_INTEGER\fP .UNINDENT .SS mpi4py.MPI.TYPECLASS_REAL .INDENT 0.0 .TP .B mpi4py.MPI.TYPECLASS_REAL: int = TYPECLASS_REAL \fBint\fP \fBTYPECLASS_REAL\fP .UNINDENT .SS mpi4py.MPI.TYPECLASS_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.TYPECLASS_COMPLEX: int = TYPECLASS_COMPLEX \fBint\fP \fBTYPECLASS_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.DISTRIBUTE_NONE .INDENT 0.0 .TP .B mpi4py.MPI.DISTRIBUTE_NONE: int = DISTRIBUTE_NONE \fBint\fP \fBDISTRIBUTE_NONE\fP .UNINDENT .SS mpi4py.MPI.DISTRIBUTE_BLOCK .INDENT 0.0 .TP .B mpi4py.MPI.DISTRIBUTE_BLOCK: int = DISTRIBUTE_BLOCK \fBint\fP \fBDISTRIBUTE_BLOCK\fP .UNINDENT .SS mpi4py.MPI.DISTRIBUTE_CYCLIC .INDENT 0.0 .TP .B mpi4py.MPI.DISTRIBUTE_CYCLIC: int = DISTRIBUTE_CYCLIC \fBint\fP \fBDISTRIBUTE_CYCLIC\fP .UNINDENT .SS mpi4py.MPI.DISTRIBUTE_DFLT_DARG .INDENT 0.0 .TP .B mpi4py.MPI.DISTRIBUTE_DFLT_DARG: int = DISTRIBUTE_DFLT_DARG \fBint\fP \fBDISTRIBUTE_DFLT_DARG\fP .UNINDENT .SS mpi4py.MPI.COMBINER_NAMED .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_NAMED: int = COMBINER_NAMED \fBint\fP \fBCOMBINER_NAMED\fP .UNINDENT .SS mpi4py.MPI.COMBINER_DUP .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_DUP: int = COMBINER_DUP \fBint\fP \fBCOMBINER_DUP\fP .UNINDENT .SS mpi4py.MPI.COMBINER_CONTIGUOUS .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_CONTIGUOUS: int = COMBINER_CONTIGUOUS \fBint\fP \fBCOMBINER_CONTIGUOUS\fP .UNINDENT .SS mpi4py.MPI.COMBINER_VECTOR .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_VECTOR: int = COMBINER_VECTOR \fBint\fP \fBCOMBINER_VECTOR\fP .UNINDENT .SS mpi4py.MPI.COMBINER_HVECTOR .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_HVECTOR: int = COMBINER_HVECTOR \fBint\fP \fBCOMBINER_HVECTOR\fP .UNINDENT .SS mpi4py.MPI.COMBINER_INDEXED .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_INDEXED: int = COMBINER_INDEXED \fBint\fP \fBCOMBINER_INDEXED\fP .UNINDENT .SS mpi4py.MPI.COMBINER_HINDEXED .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_HINDEXED: int = COMBINER_HINDEXED \fBint\fP \fBCOMBINER_HINDEXED\fP .UNINDENT .SS mpi4py.MPI.COMBINER_INDEXED_BLOCK .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_INDEXED_BLOCK: int = COMBINER_INDEXED_BLOCK \fBint\fP \fBCOMBINER_INDEXED_BLOCK\fP .UNINDENT .SS mpi4py.MPI.COMBINER_HINDEXED_BLOCK .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_HINDEXED_BLOCK: int = COMBINER_HINDEXED_BLOCK \fBint\fP \fBCOMBINER_HINDEXED_BLOCK\fP .UNINDENT .SS mpi4py.MPI.COMBINER_STRUCT .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_STRUCT: int = COMBINER_STRUCT \fBint\fP \fBCOMBINER_STRUCT\fP .UNINDENT .SS mpi4py.MPI.COMBINER_SUBARRAY .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_SUBARRAY: int = COMBINER_SUBARRAY \fBint\fP \fBCOMBINER_SUBARRAY\fP .UNINDENT .SS mpi4py.MPI.COMBINER_DARRAY .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_DARRAY: int = COMBINER_DARRAY \fBint\fP \fBCOMBINER_DARRAY\fP .UNINDENT .SS mpi4py.MPI.COMBINER_RESIZED .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_RESIZED: int = COMBINER_RESIZED \fBint\fP \fBCOMBINER_RESIZED\fP .UNINDENT .SS mpi4py.MPI.COMBINER_F90_REAL .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_F90_REAL: int = COMBINER_F90_REAL \fBint\fP \fBCOMBINER_F90_REAL\fP .UNINDENT .SS mpi4py.MPI.COMBINER_F90_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_F90_COMPLEX: int = COMBINER_F90_COMPLEX \fBint\fP \fBCOMBINER_F90_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.COMBINER_F90_INTEGER .INDENT 0.0 .TP .B mpi4py.MPI.COMBINER_F90_INTEGER: int = COMBINER_F90_INTEGER \fBint\fP \fBCOMBINER_F90_INTEGER\fP .UNINDENT .SS mpi4py.MPI.IDENT .INDENT 0.0 .TP .B mpi4py.MPI.IDENT: int = IDENT \fBint\fP \fBIDENT\fP .UNINDENT .SS mpi4py.MPI.CONGRUENT .INDENT 0.0 .TP .B mpi4py.MPI.CONGRUENT: int = CONGRUENT \fBint\fP \fBCONGRUENT\fP .UNINDENT .SS mpi4py.MPI.SIMILAR .INDENT 0.0 .TP .B mpi4py.MPI.SIMILAR: int = SIMILAR \fBint\fP \fBSIMILAR\fP .UNINDENT .SS mpi4py.MPI.UNEQUAL .INDENT 0.0 .TP .B mpi4py.MPI.UNEQUAL: int = UNEQUAL \fBint\fP \fBUNEQUAL\fP .UNINDENT .SS mpi4py.MPI.CART .INDENT 0.0 .TP .B mpi4py.MPI.CART: int = CART \fBint\fP \fBCART\fP .UNINDENT .SS mpi4py.MPI.GRAPH .INDENT 0.0 .TP .B mpi4py.MPI.GRAPH: int = GRAPH \fBint\fP \fBGRAPH\fP .UNINDENT .SS mpi4py.MPI.DIST_GRAPH .INDENT 0.0 .TP .B mpi4py.MPI.DIST_GRAPH: int = DIST_GRAPH \fBint\fP \fBDIST_GRAPH\fP .UNINDENT .SS mpi4py.MPI.UNWEIGHTED .INDENT 0.0 .TP .B mpi4py.MPI.UNWEIGHTED: int = UNWEIGHTED \fBint\fP \fBUNWEIGHTED\fP .UNINDENT .SS mpi4py.MPI.WEIGHTS_EMPTY .INDENT 0.0 .TP .B mpi4py.MPI.WEIGHTS_EMPTY: int = WEIGHTS_EMPTY \fBint\fP \fBWEIGHTS_EMPTY\fP .UNINDENT .SS mpi4py.MPI.COMM_TYPE_SHARED .INDENT 0.0 .TP .B mpi4py.MPI.COMM_TYPE_SHARED: int = COMM_TYPE_SHARED \fBint\fP \fBCOMM_TYPE_SHARED\fP .UNINDENT .SS mpi4py.MPI.BSEND_OVERHEAD .INDENT 0.0 .TP .B mpi4py.MPI.BSEND_OVERHEAD: int = BSEND_OVERHEAD \fBint\fP \fBBSEND_OVERHEAD\fP .UNINDENT .SS mpi4py.MPI.WIN_FLAVOR_CREATE .INDENT 0.0 .TP .B mpi4py.MPI.WIN_FLAVOR_CREATE: int = WIN_FLAVOR_CREATE \fBint\fP \fBWIN_FLAVOR_CREATE\fP .UNINDENT .SS mpi4py.MPI.WIN_FLAVOR_ALLOCATE .INDENT 0.0 .TP .B mpi4py.MPI.WIN_FLAVOR_ALLOCATE: int = WIN_FLAVOR_ALLOCATE \fBint\fP \fBWIN_FLAVOR_ALLOCATE\fP .UNINDENT .SS mpi4py.MPI.WIN_FLAVOR_DYNAMIC .INDENT 0.0 .TP .B mpi4py.MPI.WIN_FLAVOR_DYNAMIC: int = WIN_FLAVOR_DYNAMIC \fBint\fP \fBWIN_FLAVOR_DYNAMIC\fP .UNINDENT .SS mpi4py.MPI.WIN_FLAVOR_SHARED .INDENT 0.0 .TP .B mpi4py.MPI.WIN_FLAVOR_SHARED: int = WIN_FLAVOR_SHARED \fBint\fP \fBWIN_FLAVOR_SHARED\fP .UNINDENT .SS mpi4py.MPI.WIN_SEPARATE .INDENT 0.0 .TP .B mpi4py.MPI.WIN_SEPARATE: int = WIN_SEPARATE \fBint\fP \fBWIN_SEPARATE\fP .UNINDENT .SS mpi4py.MPI.WIN_UNIFIED .INDENT 0.0 .TP .B mpi4py.MPI.WIN_UNIFIED: int = WIN_UNIFIED \fBint\fP \fBWIN_UNIFIED\fP .UNINDENT .SS mpi4py.MPI.MODE_NOCHECK .INDENT 0.0 .TP .B mpi4py.MPI.MODE_NOCHECK: int = MODE_NOCHECK \fBint\fP \fBMODE_NOCHECK\fP .UNINDENT .SS mpi4py.MPI.MODE_NOSTORE .INDENT 0.0 .TP .B mpi4py.MPI.MODE_NOSTORE: int = MODE_NOSTORE \fBint\fP \fBMODE_NOSTORE\fP .UNINDENT .SS mpi4py.MPI.MODE_NOPUT .INDENT 0.0 .TP .B mpi4py.MPI.MODE_NOPUT: int = MODE_NOPUT \fBint\fP \fBMODE_NOPUT\fP .UNINDENT .SS mpi4py.MPI.MODE_NOPRECEDE .INDENT 0.0 .TP .B mpi4py.MPI.MODE_NOPRECEDE: int = MODE_NOPRECEDE \fBint\fP \fBMODE_NOPRECEDE\fP .UNINDENT .SS mpi4py.MPI.MODE_NOSUCCEED .INDENT 0.0 .TP .B mpi4py.MPI.MODE_NOSUCCEED: int = MODE_NOSUCCEED \fBint\fP \fBMODE_NOSUCCEED\fP .UNINDENT .SS mpi4py.MPI.LOCK_EXCLUSIVE .INDENT 0.0 .TP .B mpi4py.MPI.LOCK_EXCLUSIVE: int = LOCK_EXCLUSIVE \fBint\fP \fBLOCK_EXCLUSIVE\fP .UNINDENT .SS mpi4py.MPI.LOCK_SHARED .INDENT 0.0 .TP .B mpi4py.MPI.LOCK_SHARED: int = LOCK_SHARED \fBint\fP \fBLOCK_SHARED\fP .UNINDENT .SS mpi4py.MPI.MODE_RDONLY .INDENT 0.0 .TP .B mpi4py.MPI.MODE_RDONLY: int = MODE_RDONLY \fBint\fP \fBMODE_RDONLY\fP .UNINDENT .SS mpi4py.MPI.MODE_WRONLY .INDENT 0.0 .TP .B mpi4py.MPI.MODE_WRONLY: int = MODE_WRONLY \fBint\fP \fBMODE_WRONLY\fP .UNINDENT .SS mpi4py.MPI.MODE_RDWR .INDENT 0.0 .TP .B mpi4py.MPI.MODE_RDWR: int = MODE_RDWR \fBint\fP \fBMODE_RDWR\fP .UNINDENT .SS mpi4py.MPI.MODE_CREATE .INDENT 0.0 .TP .B mpi4py.MPI.MODE_CREATE: int = MODE_CREATE \fBint\fP \fBMODE_CREATE\fP .UNINDENT .SS mpi4py.MPI.MODE_EXCL .INDENT 0.0 .TP .B mpi4py.MPI.MODE_EXCL: int = MODE_EXCL \fBint\fP \fBMODE_EXCL\fP .UNINDENT .SS mpi4py.MPI.MODE_DELETE_ON_CLOSE .INDENT 0.0 .TP .B mpi4py.MPI.MODE_DELETE_ON_CLOSE: int = MODE_DELETE_ON_CLOSE \fBint\fP \fBMODE_DELETE_ON_CLOSE\fP .UNINDENT .SS mpi4py.MPI.MODE_UNIQUE_OPEN .INDENT 0.0 .TP .B mpi4py.MPI.MODE_UNIQUE_OPEN: int = MODE_UNIQUE_OPEN \fBint\fP \fBMODE_UNIQUE_OPEN\fP .UNINDENT .SS mpi4py.MPI.MODE_SEQUENTIAL .INDENT 0.0 .TP .B mpi4py.MPI.MODE_SEQUENTIAL: int = MODE_SEQUENTIAL \fBint\fP \fBMODE_SEQUENTIAL\fP .UNINDENT .SS mpi4py.MPI.MODE_APPEND .INDENT 0.0 .TP .B mpi4py.MPI.MODE_APPEND: int = MODE_APPEND \fBint\fP \fBMODE_APPEND\fP .UNINDENT .SS mpi4py.MPI.SEEK_SET .INDENT 0.0 .TP .B mpi4py.MPI.SEEK_SET: int = SEEK_SET \fBint\fP \fBSEEK_SET\fP .UNINDENT .SS mpi4py.MPI.SEEK_CUR .INDENT 0.0 .TP .B mpi4py.MPI.SEEK_CUR: int = SEEK_CUR \fBint\fP \fBSEEK_CUR\fP .UNINDENT .SS mpi4py.MPI.SEEK_END .INDENT 0.0 .TP .B mpi4py.MPI.SEEK_END: int = SEEK_END \fBint\fP \fBSEEK_END\fP .UNINDENT .SS mpi4py.MPI.DISPLACEMENT_CURRENT .INDENT 0.0 .TP .B mpi4py.MPI.DISPLACEMENT_CURRENT: int = DISPLACEMENT_CURRENT \fBint\fP \fBDISPLACEMENT_CURRENT\fP .UNINDENT .SS mpi4py.MPI.DISP_CUR .INDENT 0.0 .TP .B mpi4py.MPI.DISP_CUR: int = DISP_CUR \fBint\fP \fBDISP_CUR\fP .UNINDENT .SS mpi4py.MPI.THREAD_SINGLE .INDENT 0.0 .TP .B mpi4py.MPI.THREAD_SINGLE: int = THREAD_SINGLE \fBint\fP \fBTHREAD_SINGLE\fP .UNINDENT .SS mpi4py.MPI.THREAD_FUNNELED .INDENT 0.0 .TP .B mpi4py.MPI.THREAD_FUNNELED: int = THREAD_FUNNELED \fBint\fP \fBTHREAD_FUNNELED\fP .UNINDENT .SS mpi4py.MPI.THREAD_SERIALIZED .INDENT 0.0 .TP .B mpi4py.MPI.THREAD_SERIALIZED: int = THREAD_SERIALIZED \fBint\fP \fBTHREAD_SERIALIZED\fP .UNINDENT .SS mpi4py.MPI.THREAD_MULTIPLE .INDENT 0.0 .TP .B mpi4py.MPI.THREAD_MULTIPLE: int = THREAD_MULTIPLE \fBint\fP \fBTHREAD_MULTIPLE\fP .UNINDENT .SS mpi4py.MPI.VERSION .INDENT 0.0 .TP .B mpi4py.MPI.VERSION: int = VERSION \fBint\fP \fBVERSION\fP .UNINDENT .SS mpi4py.MPI.SUBVERSION .INDENT 0.0 .TP .B mpi4py.MPI.SUBVERSION: int = SUBVERSION \fBint\fP \fBSUBVERSION\fP .UNINDENT .SS mpi4py.MPI.MAX_PROCESSOR_NAME .INDENT 0.0 .TP .B mpi4py.MPI.MAX_PROCESSOR_NAME: int = MAX_PROCESSOR_NAME \fBint\fP \fBMAX_PROCESSOR_NAME\fP .UNINDENT .SS mpi4py.MPI.MAX_ERROR_STRING .INDENT 0.0 .TP .B mpi4py.MPI.MAX_ERROR_STRING: int = MAX_ERROR_STRING \fBint\fP \fBMAX_ERROR_STRING\fP .UNINDENT .SS mpi4py.MPI.MAX_PORT_NAME .INDENT 0.0 .TP .B mpi4py.MPI.MAX_PORT_NAME: int = MAX_PORT_NAME \fBint\fP \fBMAX_PORT_NAME\fP .UNINDENT .SS mpi4py.MPI.MAX_INFO_KEY .INDENT 0.0 .TP .B mpi4py.MPI.MAX_INFO_KEY: int = MAX_INFO_KEY \fBint\fP \fBMAX_INFO_KEY\fP .UNINDENT .SS mpi4py.MPI.MAX_INFO_VAL .INDENT 0.0 .TP .B mpi4py.MPI.MAX_INFO_VAL: int = MAX_INFO_VAL \fBint\fP \fBMAX_INFO_VAL\fP .UNINDENT .SS mpi4py.MPI.MAX_OBJECT_NAME .INDENT 0.0 .TP .B mpi4py.MPI.MAX_OBJECT_NAME: int = MAX_OBJECT_NAME \fBint\fP \fBMAX_OBJECT_NAME\fP .UNINDENT .SS mpi4py.MPI.MAX_DATAREP_STRING .INDENT 0.0 .TP .B mpi4py.MPI.MAX_DATAREP_STRING: int = MAX_DATAREP_STRING \fBint\fP \fBMAX_DATAREP_STRING\fP .UNINDENT .SS mpi4py.MPI.MAX_LIBRARY_VERSION_STRING .INDENT 0.0 .TP .B mpi4py.MPI.MAX_LIBRARY_VERSION_STRING: int = MAX_LIBRARY_VERSION_STRING \fBint\fP \fBMAX_LIBRARY_VERSION_STRING\fP .UNINDENT .SS mpi4py.MPI.DATATYPE_NULL .INDENT 0.0 .TP .B mpi4py.MPI.DATATYPE_NULL: Datatype = DATATYPE_NULL \fBDatatype\fP \fBDATATYPE_NULL\fP .UNINDENT .SS mpi4py.MPI.UB .INDENT 0.0 .TP .B mpi4py.MPI.UB: Datatype = UB \fBDatatype\fP \fBUB\fP .UNINDENT .SS mpi4py.MPI.LB .INDENT 0.0 .TP .B mpi4py.MPI.LB: Datatype = LB \fBDatatype\fP \fBLB\fP .UNINDENT .SS mpi4py.MPI.PACKED .INDENT 0.0 .TP .B mpi4py.MPI.PACKED: Datatype = PACKED \fBDatatype\fP \fBPACKED\fP .UNINDENT .SS mpi4py.MPI.BYTE .INDENT 0.0 .TP .B mpi4py.MPI.BYTE: Datatype = BYTE \fBDatatype\fP \fBBYTE\fP .UNINDENT .SS mpi4py.MPI.AINT .INDENT 0.0 .TP .B mpi4py.MPI.AINT: Datatype = AINT \fBDatatype\fP \fBAINT\fP .UNINDENT .SS mpi4py.MPI.OFFSET .INDENT 0.0 .TP .B mpi4py.MPI.OFFSET: Datatype = OFFSET \fBDatatype\fP \fBOFFSET\fP .UNINDENT .SS mpi4py.MPI.COUNT .INDENT 0.0 .TP .B mpi4py.MPI.COUNT: Datatype = COUNT \fBDatatype\fP \fBCOUNT\fP .UNINDENT .SS mpi4py.MPI.CHAR .INDENT 0.0 .TP .B mpi4py.MPI.CHAR: Datatype = CHAR \fBDatatype\fP \fBCHAR\fP .UNINDENT .SS mpi4py.MPI.WCHAR .INDENT 0.0 .TP .B mpi4py.MPI.WCHAR: Datatype = WCHAR \fBDatatype\fP \fBWCHAR\fP .UNINDENT .SS mpi4py.MPI.SIGNED_CHAR .INDENT 0.0 .TP .B mpi4py.MPI.SIGNED_CHAR: Datatype = SIGNED_CHAR \fBDatatype\fP \fBSIGNED_CHAR\fP .UNINDENT .SS mpi4py.MPI.SHORT .INDENT 0.0 .TP .B mpi4py.MPI.SHORT: Datatype = SHORT \fBDatatype\fP \fBSHORT\fP .UNINDENT .SS mpi4py.MPI.INT .INDENT 0.0 .TP .B mpi4py.MPI.INT: Datatype = INT \fBDatatype\fP \fBINT\fP .UNINDENT .SS mpi4py.MPI.LONG .INDENT 0.0 .TP .B mpi4py.MPI.LONG: Datatype = LONG \fBDatatype\fP \fBLONG\fP .UNINDENT .SS mpi4py.MPI.LONG_LONG .INDENT 0.0 .TP .B mpi4py.MPI.LONG_LONG: Datatype = LONG_LONG \fBDatatype\fP \fBLONG_LONG\fP .UNINDENT .SS mpi4py.MPI.UNSIGNED_CHAR .INDENT 0.0 .TP .B mpi4py.MPI.UNSIGNED_CHAR: Datatype = UNSIGNED_CHAR \fBDatatype\fP \fBUNSIGNED_CHAR\fP .UNINDENT .SS mpi4py.MPI.UNSIGNED_SHORT .INDENT 0.0 .TP .B mpi4py.MPI.UNSIGNED_SHORT: Datatype = UNSIGNED_SHORT \fBDatatype\fP \fBUNSIGNED_SHORT\fP .UNINDENT .SS mpi4py.MPI.UNSIGNED .INDENT 0.0 .TP .B mpi4py.MPI.UNSIGNED: Datatype = UNSIGNED \fBDatatype\fP \fBUNSIGNED\fP .UNINDENT .SS mpi4py.MPI.UNSIGNED_LONG .INDENT 0.0 .TP .B mpi4py.MPI.UNSIGNED_LONG: Datatype = UNSIGNED_LONG \fBDatatype\fP \fBUNSIGNED_LONG\fP .UNINDENT .SS mpi4py.MPI.UNSIGNED_LONG_LONG .INDENT 0.0 .TP .B mpi4py.MPI.UNSIGNED_LONG_LONG: Datatype = UNSIGNED_LONG_LONG \fBDatatype\fP \fBUNSIGNED_LONG_LONG\fP .UNINDENT .SS mpi4py.MPI.FLOAT .INDENT 0.0 .TP .B mpi4py.MPI.FLOAT: Datatype = FLOAT \fBDatatype\fP \fBFLOAT\fP .UNINDENT .SS mpi4py.MPI.DOUBLE .INDENT 0.0 .TP .B mpi4py.MPI.DOUBLE: Datatype = DOUBLE \fBDatatype\fP \fBDOUBLE\fP .UNINDENT .SS mpi4py.MPI.LONG_DOUBLE .INDENT 0.0 .TP .B mpi4py.MPI.LONG_DOUBLE: Datatype = LONG_DOUBLE \fBDatatype\fP \fBLONG_DOUBLE\fP .UNINDENT .SS mpi4py.MPI.C_BOOL .INDENT 0.0 .TP .B mpi4py.MPI.C_BOOL: Datatype = C_BOOL \fBDatatype\fP \fBC_BOOL\fP .UNINDENT .SS mpi4py.MPI.INT8_T .INDENT 0.0 .TP .B mpi4py.MPI.INT8_T: Datatype = INT8_T \fBDatatype\fP \fBINT8_T\fP .UNINDENT .SS mpi4py.MPI.INT16_T .INDENT 0.0 .TP .B mpi4py.MPI.INT16_T: Datatype = INT16_T \fBDatatype\fP \fBINT16_T\fP .UNINDENT .SS mpi4py.MPI.INT32_T .INDENT 0.0 .TP .B mpi4py.MPI.INT32_T: Datatype = INT32_T \fBDatatype\fP \fBINT32_T\fP .UNINDENT .SS mpi4py.MPI.INT64_T .INDENT 0.0 .TP .B mpi4py.MPI.INT64_T: Datatype = INT64_T \fBDatatype\fP \fBINT64_T\fP .UNINDENT .SS mpi4py.MPI.UINT8_T .INDENT 0.0 .TP .B mpi4py.MPI.UINT8_T: Datatype = UINT8_T \fBDatatype\fP \fBUINT8_T\fP .UNINDENT .SS mpi4py.MPI.UINT16_T .INDENT 0.0 .TP .B mpi4py.MPI.UINT16_T: Datatype = UINT16_T \fBDatatype\fP \fBUINT16_T\fP .UNINDENT .SS mpi4py.MPI.UINT32_T .INDENT 0.0 .TP .B mpi4py.MPI.UINT32_T: Datatype = UINT32_T \fBDatatype\fP \fBUINT32_T\fP .UNINDENT .SS mpi4py.MPI.UINT64_T .INDENT 0.0 .TP .B mpi4py.MPI.UINT64_T: Datatype = UINT64_T \fBDatatype\fP \fBUINT64_T\fP .UNINDENT .SS mpi4py.MPI.C_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.C_COMPLEX: Datatype = C_COMPLEX \fBDatatype\fP \fBC_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.C_FLOAT_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.C_FLOAT_COMPLEX: Datatype = C_FLOAT_COMPLEX \fBDatatype\fP \fBC_FLOAT_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.C_DOUBLE_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.C_DOUBLE_COMPLEX: Datatype = C_DOUBLE_COMPLEX \fBDatatype\fP \fBC_DOUBLE_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.C_LONG_DOUBLE_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.C_LONG_DOUBLE_COMPLEX: Datatype = C_LONG_DOUBLE_COMPLEX \fBDatatype\fP \fBC_LONG_DOUBLE_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.CXX_BOOL .INDENT 0.0 .TP .B mpi4py.MPI.CXX_BOOL: Datatype = CXX_BOOL \fBDatatype\fP \fBCXX_BOOL\fP .UNINDENT .SS mpi4py.MPI.CXX_FLOAT_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.CXX_FLOAT_COMPLEX: Datatype = CXX_FLOAT_COMPLEX \fBDatatype\fP \fBCXX_FLOAT_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.CXX_DOUBLE_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.CXX_DOUBLE_COMPLEX: Datatype = CXX_DOUBLE_COMPLEX \fBDatatype\fP \fBCXX_DOUBLE_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.CXX_LONG_DOUBLE_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.CXX_LONG_DOUBLE_COMPLEX: Datatype = CXX_LONG_DOUBLE_COMPLEX \fBDatatype\fP \fBCXX_LONG_DOUBLE_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.SHORT_INT .INDENT 0.0 .TP .B mpi4py.MPI.SHORT_INT: Datatype = SHORT_INT \fBDatatype\fP \fBSHORT_INT\fP .UNINDENT .SS mpi4py.MPI.INT_INT .INDENT 0.0 .TP .B mpi4py.MPI.INT_INT: Datatype = INT_INT \fBDatatype\fP \fBINT_INT\fP .UNINDENT .SS mpi4py.MPI.TWOINT .INDENT 0.0 .TP .B mpi4py.MPI.TWOINT: Datatype = TWOINT \fBDatatype\fP \fBTWOINT\fP .UNINDENT .SS mpi4py.MPI.LONG_INT .INDENT 0.0 .TP .B mpi4py.MPI.LONG_INT: Datatype = LONG_INT \fBDatatype\fP \fBLONG_INT\fP .UNINDENT .SS mpi4py.MPI.FLOAT_INT .INDENT 0.0 .TP .B mpi4py.MPI.FLOAT_INT: Datatype = FLOAT_INT \fBDatatype\fP \fBFLOAT_INT\fP .UNINDENT .SS mpi4py.MPI.DOUBLE_INT .INDENT 0.0 .TP .B mpi4py.MPI.DOUBLE_INT: Datatype = DOUBLE_INT \fBDatatype\fP \fBDOUBLE_INT\fP .UNINDENT .SS mpi4py.MPI.LONG_DOUBLE_INT .INDENT 0.0 .TP .B mpi4py.MPI.LONG_DOUBLE_INT: Datatype = LONG_DOUBLE_INT \fBDatatype\fP \fBLONG_DOUBLE_INT\fP .UNINDENT .SS mpi4py.MPI.CHARACTER .INDENT 0.0 .TP .B mpi4py.MPI.CHARACTER: Datatype = CHARACTER \fBDatatype\fP \fBCHARACTER\fP .UNINDENT .SS mpi4py.MPI.LOGICAL .INDENT 0.0 .TP .B mpi4py.MPI.LOGICAL: Datatype = LOGICAL \fBDatatype\fP \fBLOGICAL\fP .UNINDENT .SS mpi4py.MPI.INTEGER .INDENT 0.0 .TP .B mpi4py.MPI.INTEGER: Datatype = INTEGER \fBDatatype\fP \fBINTEGER\fP .UNINDENT .SS mpi4py.MPI.REAL .INDENT 0.0 .TP .B mpi4py.MPI.REAL: Datatype = REAL \fBDatatype\fP \fBREAL\fP .UNINDENT .SS mpi4py.MPI.DOUBLE_PRECISION .INDENT 0.0 .TP .B mpi4py.MPI.DOUBLE_PRECISION: Datatype = DOUBLE_PRECISION \fBDatatype\fP \fBDOUBLE_PRECISION\fP .UNINDENT .SS mpi4py.MPI.COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.COMPLEX: Datatype = COMPLEX \fBDatatype\fP \fBCOMPLEX\fP .UNINDENT .SS mpi4py.MPI.DOUBLE_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.DOUBLE_COMPLEX: Datatype = DOUBLE_COMPLEX \fBDatatype\fP \fBDOUBLE_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.LOGICAL1 .INDENT 0.0 .TP .B mpi4py.MPI.LOGICAL1: Datatype = LOGICAL1 \fBDatatype\fP \fBLOGICAL1\fP .UNINDENT .SS mpi4py.MPI.LOGICAL2 .INDENT 0.0 .TP .B mpi4py.MPI.LOGICAL2: Datatype = LOGICAL2 \fBDatatype\fP \fBLOGICAL2\fP .UNINDENT .SS mpi4py.MPI.LOGICAL4 .INDENT 0.0 .TP .B mpi4py.MPI.LOGICAL4: Datatype = LOGICAL4 \fBDatatype\fP \fBLOGICAL4\fP .UNINDENT .SS mpi4py.MPI.LOGICAL8 .INDENT 0.0 .TP .B mpi4py.MPI.LOGICAL8: Datatype = LOGICAL8 \fBDatatype\fP \fBLOGICAL8\fP .UNINDENT .SS mpi4py.MPI.INTEGER1 .INDENT 0.0 .TP .B mpi4py.MPI.INTEGER1: Datatype = INTEGER1 \fBDatatype\fP \fBINTEGER1\fP .UNINDENT .SS mpi4py.MPI.INTEGER2 .INDENT 0.0 .TP .B mpi4py.MPI.INTEGER2: Datatype = INTEGER2 \fBDatatype\fP \fBINTEGER2\fP .UNINDENT .SS mpi4py.MPI.INTEGER4 .INDENT 0.0 .TP .B mpi4py.MPI.INTEGER4: Datatype = INTEGER4 \fBDatatype\fP \fBINTEGER4\fP .UNINDENT .SS mpi4py.MPI.INTEGER8 .INDENT 0.0 .TP .B mpi4py.MPI.INTEGER8: Datatype = INTEGER8 \fBDatatype\fP \fBINTEGER8\fP .UNINDENT .SS mpi4py.MPI.INTEGER16 .INDENT 0.0 .TP .B mpi4py.MPI.INTEGER16: Datatype = INTEGER16 \fBDatatype\fP \fBINTEGER16\fP .UNINDENT .SS mpi4py.MPI.REAL2 .INDENT 0.0 .TP .B mpi4py.MPI.REAL2: Datatype = REAL2 \fBDatatype\fP \fBREAL2\fP .UNINDENT .SS mpi4py.MPI.REAL4 .INDENT 0.0 .TP .B mpi4py.MPI.REAL4: Datatype = REAL4 \fBDatatype\fP \fBREAL4\fP .UNINDENT .SS mpi4py.MPI.REAL8 .INDENT 0.0 .TP .B mpi4py.MPI.REAL8: Datatype = REAL8 \fBDatatype\fP \fBREAL8\fP .UNINDENT .SS mpi4py.MPI.REAL16 .INDENT 0.0 .TP .B mpi4py.MPI.REAL16: Datatype = REAL16 \fBDatatype\fP \fBREAL16\fP .UNINDENT .SS mpi4py.MPI.COMPLEX4 .INDENT 0.0 .TP .B mpi4py.MPI.COMPLEX4: Datatype = COMPLEX4 \fBDatatype\fP \fBCOMPLEX4\fP .UNINDENT .SS mpi4py.MPI.COMPLEX8 .INDENT 0.0 .TP .B mpi4py.MPI.COMPLEX8: Datatype = COMPLEX8 \fBDatatype\fP \fBCOMPLEX8\fP .UNINDENT .SS mpi4py.MPI.COMPLEX16 .INDENT 0.0 .TP .B mpi4py.MPI.COMPLEX16: Datatype = COMPLEX16 \fBDatatype\fP \fBCOMPLEX16\fP .UNINDENT .SS mpi4py.MPI.COMPLEX32 .INDENT 0.0 .TP .B mpi4py.MPI.COMPLEX32: Datatype = COMPLEX32 \fBDatatype\fP \fBCOMPLEX32\fP .UNINDENT .SS mpi4py.MPI.UNSIGNED_INT .INDENT 0.0 .TP .B mpi4py.MPI.UNSIGNED_INT: Datatype = UNSIGNED_INT \fBDatatype\fP \fBUNSIGNED_INT\fP .UNINDENT .SS mpi4py.MPI.SIGNED_SHORT .INDENT 0.0 .TP .B mpi4py.MPI.SIGNED_SHORT: Datatype = SIGNED_SHORT \fBDatatype\fP \fBSIGNED_SHORT\fP .UNINDENT .SS mpi4py.MPI.SIGNED_INT .INDENT 0.0 .TP .B mpi4py.MPI.SIGNED_INT: Datatype = SIGNED_INT \fBDatatype\fP \fBSIGNED_INT\fP .UNINDENT .SS mpi4py.MPI.SIGNED_LONG .INDENT 0.0 .TP .B mpi4py.MPI.SIGNED_LONG: Datatype = SIGNED_LONG \fBDatatype\fP \fBSIGNED_LONG\fP .UNINDENT .SS mpi4py.MPI.SIGNED_LONG_LONG .INDENT 0.0 .TP .B mpi4py.MPI.SIGNED_LONG_LONG: Datatype = SIGNED_LONG_LONG \fBDatatype\fP \fBSIGNED_LONG_LONG\fP .UNINDENT .SS mpi4py.MPI.BOOL .INDENT 0.0 .TP .B mpi4py.MPI.BOOL: Datatype = BOOL \fBDatatype\fP \fBBOOL\fP .UNINDENT .SS mpi4py.MPI.SINT8_T .INDENT 0.0 .TP .B mpi4py.MPI.SINT8_T: Datatype = SINT8_T \fBDatatype\fP \fBSINT8_T\fP .UNINDENT .SS mpi4py.MPI.SINT16_T .INDENT 0.0 .TP .B mpi4py.MPI.SINT16_T: Datatype = SINT16_T \fBDatatype\fP \fBSINT16_T\fP .UNINDENT .SS mpi4py.MPI.SINT32_T .INDENT 0.0 .TP .B mpi4py.MPI.SINT32_T: Datatype = SINT32_T \fBDatatype\fP \fBSINT32_T\fP .UNINDENT .SS mpi4py.MPI.SINT64_T .INDENT 0.0 .TP .B mpi4py.MPI.SINT64_T: Datatype = SINT64_T \fBDatatype\fP \fBSINT64_T\fP .UNINDENT .SS mpi4py.MPI.F_BOOL .INDENT 0.0 .TP .B mpi4py.MPI.F_BOOL: Datatype = F_BOOL \fBDatatype\fP \fBF_BOOL\fP .UNINDENT .SS mpi4py.MPI.F_INT .INDENT 0.0 .TP .B mpi4py.MPI.F_INT: Datatype = F_INT \fBDatatype\fP \fBF_INT\fP .UNINDENT .SS mpi4py.MPI.F_FLOAT .INDENT 0.0 .TP .B mpi4py.MPI.F_FLOAT: Datatype = F_FLOAT \fBDatatype\fP \fBF_FLOAT\fP .UNINDENT .SS mpi4py.MPI.F_DOUBLE .INDENT 0.0 .TP .B mpi4py.MPI.F_DOUBLE: Datatype = F_DOUBLE \fBDatatype\fP \fBF_DOUBLE\fP .UNINDENT .SS mpi4py.MPI.F_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.F_COMPLEX: Datatype = F_COMPLEX \fBDatatype\fP \fBF_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.F_FLOAT_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.F_FLOAT_COMPLEX: Datatype = F_FLOAT_COMPLEX \fBDatatype\fP \fBF_FLOAT_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.F_DOUBLE_COMPLEX .INDENT 0.0 .TP .B mpi4py.MPI.F_DOUBLE_COMPLEX: Datatype = F_DOUBLE_COMPLEX \fBDatatype\fP \fBF_DOUBLE_COMPLEX\fP .UNINDENT .SS mpi4py.MPI.REQUEST_NULL .INDENT 0.0 .TP .B mpi4py.MPI.REQUEST_NULL: Request = REQUEST_NULL \fBRequest\fP \fBREQUEST_NULL\fP .UNINDENT .SS mpi4py.MPI.MESSAGE_NULL .INDENT 0.0 .TP .B mpi4py.MPI.MESSAGE_NULL: Message = MESSAGE_NULL \fBMessage\fP \fBMESSAGE_NULL\fP .UNINDENT .SS mpi4py.MPI.MESSAGE_NO_PROC .INDENT 0.0 .TP .B mpi4py.MPI.MESSAGE_NO_PROC: Message = MESSAGE_NO_PROC \fBMessage\fP \fBMESSAGE_NO_PROC\fP .UNINDENT .SS mpi4py.MPI.OP_NULL .INDENT 0.0 .TP .B mpi4py.MPI.OP_NULL: Op = OP_NULL \fBOp\fP \fBOP_NULL\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.MAX .INDENT 0.0 .TP .B mpi4py.MPI.MAX: Op = MAX \fBOp\fP \fBMAX\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.MIN .INDENT 0.0 .TP .B mpi4py.MPI.MIN: Op = MIN \fBOp\fP \fBMIN\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.SUM .INDENT 0.0 .TP .B mpi4py.MPI.SUM: Op = SUM \fBOp\fP \fBSUM\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.PROD .INDENT 0.0 .TP .B mpi4py.MPI.PROD: Op = PROD \fBOp\fP \fBPROD\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.LAND .INDENT 0.0 .TP .B mpi4py.MPI.LAND: Op = LAND \fBOp\fP \fBLAND\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.BAND .INDENT 0.0 .TP .B mpi4py.MPI.BAND: Op = BAND \fBOp\fP \fBBAND\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.LOR .INDENT 0.0 .TP .B mpi4py.MPI.LOR: Op = LOR \fBOp\fP \fBLOR\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.BOR .INDENT 0.0 .TP .B mpi4py.MPI.BOR: Op = BOR \fBOp\fP \fBBOR\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.LXOR .INDENT 0.0 .TP .B mpi4py.MPI.LXOR: Op = LXOR \fBOp\fP \fBLXOR\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.BXOR .INDENT 0.0 .TP .B mpi4py.MPI.BXOR: Op = BXOR \fBOp\fP \fBBXOR\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.MAXLOC .INDENT 0.0 .TP .B mpi4py.MPI.MAXLOC: Op = MAXLOC \fBOp\fP \fBMAXLOC\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.MINLOC .INDENT 0.0 .TP .B mpi4py.MPI.MINLOC: Op = MINLOC \fBOp\fP \fBMINLOC\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.REPLACE .INDENT 0.0 .TP .B mpi4py.MPI.REPLACE: Op = REPLACE \fBOp\fP \fBREPLACE\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.NO_OP .INDENT 0.0 .TP .B mpi4py.MPI.NO_OP: Op = NO_OP \fBOp\fP \fBNO_OP\fP .INDENT 7.0 .TP .B Parameters .INDENT 7.0 .IP \(bu 2 \fBx\fP (\fIAny\fP) – .IP \(bu 2 \fBy\fP (\fIAny\fP) – .UNINDENT .TP .B Return type Any .UNINDENT .UNINDENT .SS mpi4py.MPI.GROUP_NULL .INDENT 0.0 .TP .B mpi4py.MPI.GROUP_NULL: Group = GROUP_NULL \fBGroup\fP \fBGROUP_NULL\fP .UNINDENT .SS mpi4py.MPI.GROUP_EMPTY .INDENT 0.0 .TP .B mpi4py.MPI.GROUP_EMPTY: Group = GROUP_EMPTY \fBGroup\fP \fBGROUP_EMPTY\fP .UNINDENT .SS mpi4py.MPI.INFO_NULL .INDENT 0.0 .TP .B mpi4py.MPI.INFO_NULL: Info = INFO_NULL \fBInfo\fP \fBINFO_NULL\fP .UNINDENT .SS mpi4py.MPI.INFO_ENV .INDENT 0.0 .TP .B mpi4py.MPI.INFO_ENV: Info = INFO_ENV \fBInfo\fP \fBINFO_ENV\fP .UNINDENT .SS mpi4py.MPI.ERRHANDLER_NULL .INDENT 0.0 .TP .B mpi4py.MPI.ERRHANDLER_NULL: Errhandler = ERRHANDLER_NULL \fBErrhandler\fP \fBERRHANDLER_NULL\fP .UNINDENT .SS mpi4py.MPI.ERRORS_RETURN .INDENT 0.0 .TP .B mpi4py.MPI.ERRORS_RETURN: Errhandler = ERRORS_RETURN \fBErrhandler\fP \fBERRORS_RETURN\fP .UNINDENT .SS mpi4py.MPI.ERRORS_ARE_FATAL .INDENT 0.0 .TP .B mpi4py.MPI.ERRORS_ARE_FATAL: Errhandler = ERRORS_ARE_FATAL \fBErrhandler\fP \fBERRORS_ARE_FATAL\fP .UNINDENT .SS mpi4py.MPI.COMM_NULL .INDENT 0.0 .TP .B mpi4py.MPI.COMM_NULL: Comm = COMM_NULL \fBComm\fP \fBCOMM_NULL\fP .UNINDENT .SS mpi4py.MPI.COMM_SELF .INDENT 0.0 .TP .B mpi4py.MPI.COMM_SELF: Intracomm = COMM_SELF \fBIntracomm\fP \fBCOMM_SELF\fP .UNINDENT .SS mpi4py.MPI.COMM_WORLD .INDENT 0.0 .TP .B mpi4py.MPI.COMM_WORLD: Intracomm = COMM_WORLD \fBIntracomm\fP \fBCOMM_WORLD\fP .UNINDENT .SS mpi4py.MPI.WIN_NULL .INDENT 0.0 .TP .B mpi4py.MPI.WIN_NULL: Win = WIN_NULL \fBWin\fP \fBWIN_NULL\fP .UNINDENT .SS mpi4py.MPI.FILE_NULL .INDENT 0.0 .TP .B mpi4py.MPI.FILE_NULL: File = FILE_NULL \fBFile\fP \fBFILE_NULL\fP .UNINDENT .SS mpi4py.MPI.pickle .INDENT 0.0 .TP .B mpi4py.MPI.pickle: Pickle = \fBPickle\fP \fBpickle\fP .UNINDENT .SH CITATION .sp If MPI for Python been significant to a project that leads to an academic publication, please acknowledge that fact by citing the project. .INDENT 0.0 .IP \(bu 2 L. Dalcin and Y.\-L. L. Fang, \fImpi4py: Status Update After 12 Years of Development\fP, Computing in Science & Engineering, 23(4):47\-54, 2021. \fI\%https://doi.org/10.1109/MCSE.2021.3083216\fP .IP \(bu 2 L. Dalcin, P. Kler, R. Paz, and A. Cosimo, \fIParallel Distributed Computing using Python\fP, Advances in Water Resources, 34(9):1124\-1139, 2011. \fI\%https://doi.org/10.1016/j.advwatres.2011.04.013\fP .IP \(bu 2 L. Dalcin, R. Paz, M. Storti, and J. D’Elia, \fIMPI for Python: performance improvements and MPI\-2 extensions\fP, Journal of Parallel and Distributed Computing, 68(5):655\-662, 2008. \fI\%https://doi.org/10.1016/j.jpdc.2007.09.005\fP .IP \(bu 2 L. Dalcin, R. Paz, and M. Storti, \fIMPI for Python\fP, Journal of Parallel and Distributed Computing, 65(9):1108\-1115, 2005. \fI\%https://doi.org/10.1016/j.jpdc.2005.03.010\fP .UNINDENT .SH INSTALLATION .SS Requirements .sp You need to have the following software properly installed in order to build \fIMPI for Python\fP: .INDENT 0.0 .IP \(bu 2 A working MPI implementation, preferably supporting MPI\-3 and built with shared/dynamic libraries. .sp \fBNOTE:\fP .INDENT 2.0 .INDENT 3.5 If you want to build some MPI implementation from sources, check the instructions at building\-mpi in the appendix. .UNINDENT .UNINDENT .IP \(bu 2 Python 2.7, 3.5 or above. .sp \fBNOTE:\fP .INDENT 2.0 .INDENT 3.5 Some MPI\-1 implementations \fBdo require\fP the actual command line arguments to be passed in \fBMPI_Init()\fP\&. In this case, you will need to use a rebuilt, MPI\-enabled, Python interpreter executable. \fIMPI for Python\fP has some support for alleviating you from this task. Check the instructions at python\-mpi in the appendix. .UNINDENT .UNINDENT .UNINDENT .SS Using \fBpip\fP .sp If you already have a working MPI (either if you installed it from sources or by using a pre\-built package from your favourite GNU/Linux distribution) and the \fBmpicc\fP compiler wrapper is on your search path, you can use \fBpip\fP: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python \-m pip install mpi4py .ft P .fi .UNINDENT .UNINDENT .sp \fBNOTE:\fP .INDENT 0.0 .INDENT 3.5 If the \fBmpicc\fP compiler wrapper is not on your search path (or if it has a different name) you can use \fBenv\fP to pass the environment variable \fBMPICC\fP providing the full path to the MPI compiler wrapper executable: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ env MPICC=/path/to/mpicc python \-m pip install mpi4py .ft P .fi .UNINDENT .UNINDENT .UNINDENT .UNINDENT .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 \fBpip\fP keeps previouly built wheel files on its cache for future reuse. If you want to reinstall the \fBmpi4py\fP package using a different or updated MPI implementation, you have to either first remove the cached wheel file with: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python \-m pip cache remove mpi4py .ft P .fi .UNINDENT .UNINDENT .sp or ask \fBpip\fP to disable the cache: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python \-m pip install \-\-no\-cache\-dir mpi4py .ft P .fi .UNINDENT .UNINDENT .UNINDENT .UNINDENT .SS Using \fBdistutils\fP .sp The \fIMPI for Python\fP package is available for download at the project website generously hosted by GitHub. You can use \fBcurl\fP or \fBwget\fP to get a release tarball. .INDENT 0.0 .IP \(bu 2 Using \fBcurl\fP: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C $ curl \-O https://github.com/mpi4py/mpi4py/releases/download/X.Y.Z/mpi4py\-X.Y.Z.tar.gz .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 Using \fBwget\fP: .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C $ wget https://github.com/mpi4py/mpi4py/releases/download/X.Y.Z/mpi4py\-X.Y.Z.tar.gz .ft P .fi .UNINDENT .UNINDENT .UNINDENT .sp After unpacking the release tarball: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ tar \-zxf mpi4py\-X.Y.Z.tar.gz $ cd mpi4py\-X.Y.Z .ft P .fi .UNINDENT .UNINDENT .sp the package is ready for building. .sp \fIMPI for Python\fP uses a standard distutils\-based build system. However, some distutils commands (like \fIbuild\fP) have additional options: .INDENT 0.0 .TP .B \-\-mpicc= Lets you specify a special location or name for the \fBmpicc\fP compiler wrapper. .UNINDENT .INDENT 0.0 .TP .B \-\-mpi= Lets you pass a section with MPI configuration within a special configuration file. .UNINDENT .INDENT 0.0 .TP .B \-\-configure Runs exhaustive tests for checking about missing MPI types, constants, and functions. This option should be passed in order to build \fIMPI for Python\fP against old MPI\-1 or MPI\-2 implementations, possibly providing a subset of MPI\-3. .UNINDENT .sp If you use a MPI implementation providing a \fBmpicc\fP compiler wrapper (e.g., MPICH, Open MPI), it will be used for compilation and linking. This is the preferred and easiest way of building \fIMPI for Python\fP\&. .sp If \fBmpicc\fP is located somewhere in your search path, simply run the \fIbuild\fP command: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python setup.py build .ft P .fi .UNINDENT .UNINDENT .sp If \fBmpicc\fP is not in your search path or the compiler wrapper has a different name, you can run the \fIbuild\fP command specifying its location: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python setup.py build \-\-mpicc=/where/you/have/mpicc .ft P .fi .UNINDENT .UNINDENT .sp Alternatively, you can provide all the relevant information about your MPI implementation by editing the file called \fBmpi.cfg\fP\&. You can use the default section \fB[mpi]\fP or add a new, custom section, for example \fB[other_mpi]\fP (see the examples provided in the \fBmpi.cfg\fP file as a starting point to write your own section): .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C [mpi] include_dirs = /usr/local/mpi/include libraries = mpi library_dirs = /usr/local/mpi/lib runtime_library_dirs = /usr/local/mpi/lib [other_mpi] include_dirs = /opt/mpi/include ... libraries = mpi ... library_dirs = /opt/mpi/lib ... runtime_library_dirs = /op/mpi/lib ... \&... .ft P .fi .UNINDENT .UNINDENT .sp and then run the \fIbuild\fP command, perhaps specifying you custom configuration section: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python setup.py build \-\-mpi=other_mpi .ft P .fi .UNINDENT .UNINDENT .sp After building, the package is ready for install. .sp If you have root privileges (either by log\-in as the root user of by using \fBsudo\fP) and you want to install \fIMPI for Python\fP in your system for all users, just do: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python setup.py install .ft P .fi .UNINDENT .UNINDENT .sp The previous steps will install the \fBmpi4py\fP package at standard location \fIprefix\fP\fB/lib/python\fP\fIX\fP\fB\&.\fP\fIX\fP\fB/site\-packages\fP\&. .sp If you do not have root privileges or you want to install \fIMPI for Python\fP for your private use, just do: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python setup.py install \-\-user .ft P .fi .UNINDENT .UNINDENT .SS Testing .sp To quickly test the installation: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ 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. .ft P .fi .UNINDENT .UNINDENT .sp If you installed from source, issuing at the command line: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 5 python demo/helloworld.py .ft P .fi .UNINDENT .UNINDENT .sp or (in the case of ancient MPI\-1 implementations): .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpirun \-np 5 python \(gapwd\(ga/demo/helloworld.py .ft P .fi .UNINDENT .UNINDENT .sp will launch a five\-process run of the Python interpreter and run the test script \fBdemo/helloworld.py\fP from the source distribution. .sp You can also run all the \fIunittest\fP scripts: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 5 python test/runtests.py .ft P .fi .UNINDENT .UNINDENT .sp or, if you have \fI\%nose\fP unit testing framework installed: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 5 nosetests \-w test .ft P .fi .UNINDENT .UNINDENT .sp or, if you have \fI\%py.test\fP unit testing framework installed: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ mpiexec \-n 5 py.test test/ .ft P .fi .UNINDENT .UNINDENT .SH APPENDIX .SS MPI\-enabled Python interpreter .INDENT 0.0 .INDENT 3.5 .sp \fBWARNING:\fP .INDENT 0.0 .INDENT 3.5 These days it is no longer required to use the MPI\-enabled Python interpreter in most cases, and, therefore, it is not built by default anymore because it is too difficult to reliably build a Python interpreter across different distributions. If you know that you still \fBreally\fP need it, see below on how to use the \fBbuild_exe\fP and \fBinstall_exe\fP commands. .UNINDENT .UNINDENT .UNINDENT .UNINDENT .sp Some MPI\-1 implementations (notably, MPICH 1) \fBdo require\fP the actual command line arguments to be passed at the time \fBMPI_Init()\fP is called. In this case, you will need to use a re\-built, MPI\-enabled, Python interpreter binary executable. A basic implementation (targeting Python 2.X) of what is required is shown below: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C #include #include int main(int argc, char *argv[]) { int status, flag; MPI_Init(&argc, &argv); status = Py_Main(argc, argv); MPI_Finalized(&flag); if (!flag) MPI_Finalize(); return status; } .ft P .fi .UNINDENT .UNINDENT .sp The source code above is straightforward; compiling it should also be. However, the linking step is more tricky: special flags have to be passed to the linker depending on your platform. In order to alleviate you for such low\-level details, \fIMPI for Python\fP provides some pure\-distutils based support to build and install an MPI\-enabled Python interpreter executable: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ cd mpi4py\-X.X.X $ python setup.py build_exe [\-\-mpi=|\-\-mpicc=/path/to/mpicc] $ [sudo] python setup.py install_exe [\-\-install\-dir=$HOME/bin] .ft P .fi .UNINDENT .UNINDENT .sp After the above steps you should have the MPI\-enabled interpreter installed as \fIprefix\fP\fB/bin/python\fP\fIX\fP\fB\&.\fP\fIX\fP\fB\-mpi\fP (or \fB$HOME/bin/python\fP\fIX\fP\fB\&.\fP\fIX\fP\fB\-mpi\fP). Assuming that \fIprefix\fP\fB/bin\fP (or \fB$HOME/bin\fP) is listed on your \fBPATH\fP, you should be able to enter your MPI\-enabled Python interactively, for example: .INDENT 0.0 .INDENT 3.5 .sp .nf .ft C $ python2.7\-mpi Python 2.7.8 (default, Nov 10 2014, 08:19:18) [GCC 4.9.2 20141101 (Red Hat 4.9.2\-1)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> sys.executable \(aq/usr/bin/python2.7\-mpi\(aq >>> .ft P .fi .UNINDENT .UNINDENT .SS Building MPI from sources .sp In the list below you have some executive instructions for building some of the open\-source MPI implementations out there with support for shared/dynamic libraries on POSIX environments. .INDENT 0.0 .IP \(bu 2 \fIMPICH\fP .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C $ tar \-zxf mpich\-X.X.X.tar.gz $ cd mpich\-X.X.X $ ./configure \-\-enable\-shared \-\-prefix=/usr/local/mpich $ make $ make install .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 \fIOpen MPI\fP .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C $ tar \-zxf openmpi\-X.X.X tar.gz $ cd openmpi\-X.X.X $ ./configure \-\-prefix=/usr/local/openmpi $ make all $ make install .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 \fIMPICH 1\fP .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C $ tar \-zxf mpich\-X.X.X.tar.gz $ cd mpich\-X.X.X $ ./configure \-\-enable\-sharedlib \-\-prefix=/usr/local/mpich1 $ make $ make install .ft P .fi .UNINDENT .UNINDENT .UNINDENT .sp Perhaps you will need to set the \fBLD_LIBRARY_PATH\fP environment variable (using \fBexport\fP, \fBsetenv\fP or what applies to your system) pointing to the directory containing the MPI libraries . In case of getting runtime linking errors when running MPI programs, the following lines can be added to the user login shell script (\fB\&.profile\fP, \fB\&.bashrc\fP, etc.). .INDENT 0.0 .IP \(bu 2 \fIMPICH\fP .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C MPI_DIR=/usr/local/mpich export LD_LIBRARY_PATH=$MPI_DIR/lib:$LD_LIBRARY_PATH .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 \fIOpen MPI\fP .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C MPI_DIR=/usr/local/openmpi export LD_LIBRARY_PATH=$MPI_DIR/lib:$LD_LIBRARY_PATH .ft P .fi .UNINDENT .UNINDENT .IP \(bu 2 \fIMPICH 1\fP .INDENT 2.0 .INDENT 3.5 .sp .nf .ft C MPI_DIR=/usr/local/mpich1 export LD_LIBRARY_PATH=$MPI_DIR/lib/shared:$LD_LIBRARY_PATH: export MPICH_USE_SHLIB=yes .ft P .fi .UNINDENT .UNINDENT .sp \fBWARNING:\fP .INDENT 2.0 .INDENT 3.5 MPICH 1 support for dynamic libraries is not completely transparent. Users should set the environment variable \fBMPICH_USE_SHLIB\fP to \fByes\fP in order to avoid link problems when using the \fBmpicc\fP compiler wrapper. .UNINDENT .UNINDENT .UNINDENT .SH AUTHOR Lisandro Dalcin .SH COPYRIGHT 2021, Lisandro Dalcin .\" Generated by docutils manpage writer. .