MPI_WAITALL(3)

📦
version 5.0.11 · 17 versions
, in arch rolling

🎀 MPI_Waitall — the "wait for ALL ur little comms to finish" girlie 🎀

okay so this is Open MPI and we're looking at MPI_Waitall(3) — yes bestie, the one from the Open MPI manual, she's iconic, she's patient, she waits for everything.

✨ SYNTAX ✨

💅 C Syntax

#include <mpi.h>

int MPI_Waitall(int count, MPI_Request array_of_requests[],
     MPI_Status *array_of_statuses)

💅 Fortran Syntax

USE MPI
! or the older form: INCLUDE 'mpif.h'
MPI_WAITALL(COUNT, ARRAY_OF_REQUESTS, ARRAY_OF_STATUSES, IERROR)
     INTEGER COUNT, ARRAY_OF_REQUESTS(*)
     INTEGER ARRAY_OF_STATUSES(MPI_STATUS_SIZE,*), IERROR

💅 Fortran 2008 Syntax

USE mpi_f08
MPI_Waitall(count, array_of_requests, array_of_statuses, ierror)
     INTEGER, INTENT(IN) :: count
     TYPE(MPI_Request), INTENT(INOUT) :: array_of_requests(count)
     TYPE(MPI_Status) :: array_of_statuses(*)
     INTEGER, OPTIONAL, INTENT(OUT) :: ierror

🌸 INPUT PARAMETERS 🌸

  • count — the length of the list, bestie. It's an integer. It's giving "how many things am I waiting on".
  • array_of_requests — the array of requests (array of handles). literally the guest list of comms u want to see finish 💌

🦋 OUTPUT PARAMETERS 🦋

  • array_of_statuses — the array of status objects (array of status). the receipts 📋
  • ierror — Fortran only: error status (integer). yes it's a Fortran-exclusive lil guy 🫶

👑 DESCRIPTION 👑

so here's the tea: MPI_Waitall blocks until all the communication operations associated with the active handles in the list are complete, and then hands u back the status of all those operations. and yes, that includes the case where literally no handle in the list is active — she still shows up, she still does her job 💅

Both arrays have the same number of valid entries. No cap. The i-th entry in array_of_statuses gets set to the return status of the i-th operation. Matchy-matchy, the way that we love it ✨

Requests that were created by nonblocking communication operations get deallocated, and the corresponding handles in the array get set to MPI_REQUEST_NULL. so they're gone, bestie, released into the void 🦋

The list may contain null or inactive handles — that's totally fine, she's not gonna be weird about it. The call sets to empty the status of each such entry.

okay here's the equivalence moment, write this down:

The error-free execution of MPI_Waitall(count, array_of_requests, array_of_statuses) has the same effect as the execution of MPI_Wait(&array_of_request[i], &array_of_statuses[i]), for i=0,…,count-1, in some arbitrary order. MPI_Waitall with an array of length 1 is equivalent to MPI_Wait.

(notice she wrote array_of_request[i] with one s in that snippet — that's how the original spells it, so we're keeping it exactly, we don't fix, we don't touch, we preserve 💖)

Now the messy part 😭 — when one or more of the communications completed by a call to MPI_Waitall fail, it's desirable to return specific information on each communication. In that case MPI_Waitall returns the error code MPI_ERR_IN_STATUS and sets the error field of each status to a specific error code, like so:

  • it'll be MPI_SUCCESS if that specific communication completed ✅
  • it'll be another specific error code if it failed 💀
  • it can be MPI_ERR_PENDING if it has neither failed nor completed (she's still ✨processing✨)

MPI_Waitall returns MPI_SUCCESS if no request had an error, or it returns another error code if it failed for other reasons (such as invalid arguments). In those other-reasons cases, it will not update the error fields of the statuses. so don't go looking, they're untouched 🙈

and if ur application doesn't need to examine the array_of_statuses field? u can save resources by using the predefined constant MPI_STATUSES_IGNORE as a special value for the array_of_statuses argument. it's giving minimalism, it's giving efficiency, we stan 💜

🧚 ERRORS 🧚

Almost all MPI routines return an error value; C routines as the return result of the function and Fortran routines in the last argument. classic.

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. okay that was a lot but we got through it together 🫶

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

  • MPI_ERRORS_ARE_FATAL — Causes the program to abort all connected MPI processes. dramatic but honestly valid 💀
  • MPI_ERRORS_ABORT — An error handler that can be invoked on a communicator, window, file, or session. When called on a communicator, it acts as if MPI_Abort was called on that communicator. If called on a window or file, acts as if MPI_Abort was called on a communicator containing the group of processes in the corresponding window or file. If called on a session, aborts only the local process. she's versatile, she contains multitudes 👑
  • MPI_ERRORS_RETURN — Returns an error code to the application. the chill one 🌈

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

DIY queen behavior, honestly 💅

Note that MPI does not guarantee that an MPI program can continue past an error. fr fr, take that seriously.

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.

For each invocation of MPI_Waitall, if one or more requests generate an MPI error, only the first MPI request that caused an error will be passed to its corresponding error handler. No other error handlers will be invoked (even if multiple requests generated errors). However, all requests that generate an error will have a relevant error code set in the corresponding status.MPI_ERROR field (unless MPI_STATUSES_IGNORE was used). so like. one handler gets the message, but everybody gets the receipt 📬

If the invoked error handler allows MPI_Waitall to return to the caller, the value MPI_ERR_IN_STATUS will be returned in the C and Fortran bindings.

🍓 SEE ALSO 🍓

the squad, the friend group, the whole ensemble:

  • MPI_Comm_set_errhandler
  • MPI_File_set_errhandler
  • MPI_Test
  • MPI_Testall
  • MPI_Testany
  • MPI_Testsome
  • MPI_Wait
  • MPI_Waitany
  • MPI_Waitsome
  • MPI_Win_set_errhandler

go read them, iykyk 💖

2003-2026, The Open MPI Community

okay that's the whole page bestie, ily, go wait on ur requests responsibly 💜✨ this was MPI_WAITALL(3) from Open MPI, dated Sep 11, 2026 — mwah, slay, periodt 🎀🦋

📦 where this page comes from