MPI_ISENDRECV(3)

πŸ“¦
version 5.0.10 Β· 4 versions
, in ubuntu stonking

πŸ’Œ MPI_ISENDRECV β€” bestie, you're sending AND receiving, at the same time, no blocking, no waiting πŸ’…

okay so this is the Open MPI manual page for MPI_ISENDRECV(3), and honestly? she's iconic. two messages, one call, zero patience. let's get into it πŸ¦‹


πŸŽ€ NAME πŸŽ€

MPI_Isendrecv β€” sends and receives a message. yes. both. at once. she contains multitudes ✨


✨ SYNTAX ✨

🌸 C Syntax

#include <mpi.h>

int MPI_Isendrecv(const void *sendbuf, int sendcount, MPI_Datatype sendtype,
    int dest, int sendtag, void *recvbuf, int recvcount,
    MPI_Datatype recvtype, int source, int recvtag,
    MPI_Comm comm, MPI_Request *request)

🌸 Fortran Syntax

USE MPI
! or the older form: INCLUDE 'mpif.h'
MPI_ISENDRECV(SENDBUF, SENDCOUNT, SENDTYPE, DEST, SENDTAG,
        RECVBUF, RECVCOUNT, RECVTYPE, SOURCE, RECVTAG, COMM,
        REQUEST, IERROR)
    <type>    SENDBUF(*), RECVBUF(*)
    INTEGER    SENDCOUNT, SENDTYPE, DEST, SENDTAG
    INTEGER    RECVCOUNT, RECVTYPE, SOURCE, RECVTAG, COMM
    INTEGER    REQUEST, IERROR

🌸 Fortran 2008 Syntax

USE mpi_f08
MPI_Isendrecv(sendbuf, sendcount, sendtype, dest, sendtag, recvbuf,
        recvcount, recvtype, source, recvtag, comm, request, ierror)
    TYPE(*), DIMENSION(..), INTENT(IN) :: sendbuf
    TYPE(*), DIMENSION(..) :: recvbuf
    INTEGER, INTENT(IN) :: sendcount, dest, sendtag, recvcount, source,
    recvtag
    TYPE(MPI_Datatype), INTENT(IN) :: sendtype, recvtype
    TYPE(MPI_Comm), INTENT(IN) :: comm
    TYPE(MPI_Request) :: request
    INTEGER, OPTIONAL, INTENT(OUT) :: ierror

πŸ“₯ INPUT PARAMETERS πŸ“₯

the stuff you're handing in, bestie. give her what she needs 🫢

  • sendbuf β€” initial address of send buffer (choice). where your outgoing message lives πŸ’–
  • sendcount β€” number of elements to send (integer). count 'em πŸ’…
  • sendtype β€” type of elements in send buffer (handle). what kind of little guys you're sending πŸ‘‘
  • dest β€” rank of destination (integer). who's receiving this πŸ‘€
  • sendtag β€” send tag (integer). label it so it knows where it belongs 🏷️
  • recvcount β€” maximum number of elements to receive (integer). maximum, she said. not a promise 🌸
  • recvtype β€” type of elements in receive buffer (handle)
  • source β€” rank of source (integer). who's sliding into your buffer πŸ¦‹
  • recvtag β€” receive tag (integer)
  • comm β€” communicator (handle). the group chat everyone's in 🌈

πŸ“€ OUTPUT PARAMETERS πŸ“€

what she gives back 😭

  • recvbuf β€” initial address of receive buffer (choice). the landing pad ⭐
  • request β€” communication request (handle). hold onto this one, you'll need her later πŸ”₯
  • ierror β€” Fortran only: error status (integer). classic Fortran, always hiding the error in the last argument πŸ™ˆ

πŸ’– DESCRIPTION πŸ’–

okay so the non-blocking send-receive operations combine, in one single call, the sending of a message to one destination and the receiving of another message from another process. it's giving multitasking main character energy. the two β€” source and destination β€” are possibly the same. so yes, she might be texting herself. no judgement πŸ’…

this operation is useful for executing a shift operation across a chain of processes. the send-receive operation can be used in conjunction with the functions described in the β€œProcess Topologies” chapter of the MPI Standard in order to perform shifts on various logical topologies. topologies? shifting? she's that girl, periodt ✨

a message sent by a send-receive operation can be received by a regular receive operation or probed by a probe operation; a send-receive operation can receive a message sent by a regular send operation. so she plays nice with everybody, which is the vibe fr fr 🫢

MPI_Isendrecv executes a non-blocking send and receive operation. both send and receive use the same communicator, but possibly different tags. the send buffer and receive buffers must be disjoint, and may have different lengths and datatypes. disjoint, bestie!! do not overlap those buffers, that's not cute 😭

A non-blocking send-receive request can be determined to be completed by calling the MPI_Wait, MPI_Waitany, MPI_Test, or MPI_Testany with the request returned by this function. so when you're ready to find out if she's done, that's how. go check on her πŸ’–


πŸ’€ ERRORS πŸ’€

almost all MPI routines return an error value; C routines as the return result of the function and Fortran routines in the last argument. (yes, that last-argument thing again. Fortran is consistent if nothing else πŸ™ˆ)

before the error value is returned, the current MPI error handler associated with the communication object (e.g., communicator, window, file) is called. if no communication object is associated with the MPI call, then the call is considered attached to MPI_COMM_SELF and will call the associated MPI error handler. when MPI_COMM_SELF is not initialized (i.e., before MPI_Init/MPI_Init_thread, after MPI_Finalize, or when using the Sessions Model exclusively) the error raises the initial error handler. the initial error handler can be changed by calling MPI_Comm_set_errhandler on MPI_COMM_SELF when using the World model, or the mpi_initial_errhandler CLI argument to mpiexec or info key to MPI_Comm_spawn/MPI_Comm_spawn_multiple. if no other appropriate error handler has been set, then the MPI_ERRORS_RETURN error handler is called for MPI I/O functions and the MPI_ERRORS_ABORT error handler is called for all other MPI functions.

Open MPI includes three predefined error handlers that can be used:

  • MPI_ERRORS_ARE_FATAL β€” causes the program to abort all connected MPI processes. she said everybody πŸ’€πŸ”₯
  • MPI_ERRORS_ABORT β€” an error handler that can be invoked on a communicator, window, file, or session. when called on a communicator, it acts as if MPI_Abort was called on that communicator. if called on a window or file, acts as if MPI_Abort was called on a communicator containing the group of processes in the corresponding window or file. if called on a session, aborts only the local process. so yeah. dramatic but localized πŸ’…
  • MPI_ERRORS_RETURN β€” returns an error code to the application. the chill one. hands you the receipt and moves on 🧾

MPI applications can also implement their own error handlers by calling:

  • MPI_Comm_create_errhandler then MPI_Comm_set_errhandler
  • MPI_File_create_errhandler then MPI_File_set_errhandler
  • MPI_Session_create_errhandler then MPI_Session_set_errhandler or at MPI_Session_init
  • MPI_Win_create_errhandler then MPI_Win_set_errhandler

note that MPI does not guarantee that an MPI program can continue past an error. that's a real one, bestie. no cap. don't build your whole flow on "it'll be fine" 😭

see the MPI man page for a full list of MPI error codes. see the Error Handling section of the MPI-3.1 standard for more information. πŸ“š

πŸ‘€ See also:

  • MPI_Isendrecv_replace
  • MPI_Sendrecv
  • MPI_Sendrecv_replace

2003-2026, The Open MPI Community πŸ’–


and that's the whole thing, bestie πŸŽ€βœ¨ this page came to you from the Open MPI crew (their MPI_ISENDRECV(3) corner of the universe), dated Sep 10, 2026. go forth, make those buffers disjoint, and remember to MPI_Wait on your request. mwah πŸ’…πŸ’–πŸ¦‹

πŸ“¦ where this page comes from