MPIX_Waitall_enqueue(3)

πŸ“¦
version 5.0.1 Β· 4 versions
, in nixos 26.05, ubuntu stonking

πŸ’– MPIX_Waitall_enqueue β€” the tea β˜•

okay so bestie, welcome to the MPIX_Waitall_enqueue(3) page, section 3 of the manual, living up in the MPI neighborhood 🫢. This is the glow-up of a waitall that gets enqueued onto a GPU stream β€” specifically, one that's associated with the local stream. Yeah. Main character energy, honestly. πŸ”₯


✨ NAME ✨

MPIX_Waitall_enqueue β€” enqueue a waitall operation to a GPU stream that is associated with the local stream πŸ¦‹


✨ SYNOPSIS ✨

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

✨ INPUT PARAMETERS ✨

  • count β€” lists length (a non-negative integer, no cap πŸ’…)

✨ INPUT/OUTPUT PARAMETERS ✨

  • array_of_requests β€” array of requests (handle)

✨ OUTPUT PARAMETERS ✨

  • array_of_statuses β€” array of status objects (Status)

🧡 THREAD AND INTERRUPT SAFETY 🧡

So this routine is thread-safe, periodt. That means multiple threads can use it at the same time, and you do not need to bring your own thread locks. Slay. ✨

BUT β€” and this is the important bit, bestie β€” the routine is not interrupt safe. 😭 Typically that's because it uses memory allocation routines like malloc, or other non-MPICH runtime routines, and those cuties are themselves not interrupt-safe. It's giving "not my fault, blame malloc".

🌈 NOTES FOR FORTRAN 🌈

Every MPI routine in Fortran (except for MPI_WTIME and MPI_WTICK, the two rebels πŸ’€) gets an extra argument ierr tacked onto the end of the argument list. ierr is an integer and it means the exact same thing as the return value of the routine in C. iykyk.

In Fortran, MPI routines are subroutines, and you invoke them with the call statement. πŸŽ€

And also: all MPI objects (e.g., MPI_Datatype, MPI_Comm) are of type INTEGER in Fortran.


πŸ’€ ERRORS πŸ’€

Every MPI routine (again, except MPI_Wtime and MPI_Wtick, the iconoclasts) returns an error value β€” C routines as the value of the function, and Fortran routines in the last argument.

Before that value comes back to you, the current MPI error handler gets called. By default, this error handler aborts the MPI job. 😳 Dramatic but true.

You can swap the error handler out with:

  • MPI_Comm_set_errhandler β€” for communicators
  • MPI_File_set_errhandler β€” for files
  • MPI_Win_set_errhandler β€” for RMA windows

The MPI-1 routine MPI_Errhandler_set may be used, but ngl its use is deprecated. πŸ™ˆ

And the predefined error handler MPI_ERRORS_RETURN may be used to cause error values to be returned instead.

Note that MPI does not guarantee that an MPI program can continue past an error. However, MPI implementations will attempt to continue whenever possible. So like… bestie, no promises. 🫢

Here's the full error-code lineup:

  • MPI_SUCCESS β€” No error; MPI routine completed successfully. πŸ’–
  • MPI_ERR_COUNT β€” Invalid count argument. Count arguments must be non-negative; a count of zero is often valid. (Yes, zero counts. She said "often valid" and she meant it.)
  • MPI_ERR_REQUEST β€” Invalid MPI_Request. Either null or, in the case of a MPI_Start or MPI_Startall, not a persistent request. πŸ’€
  • MPI_ERR_OTHER β€” Other error; use MPI_Error_string to get more information about this error code. The catch-all. The "it's giving mystery" option. πŸ”₯

and that's the whole page, bestie!! πŸŽ€ you ate this documentation fr fr ✨ no cap 🌸

signing off from the MPIX_Waitall_enqueue(3) page β€” dated 4/10/2026 (that's right, future girlie πŸ’…) β€” mwah, go enqueue something iconic πŸ‘‘πŸ¦‹πŸ’œ

πŸ“¦ where this page comes from