Dear Dan and Bill, Yes, Bill, and more than that: Fortran INTENT(OUT) means that the Compiler may through the Input away, i.e., if a field should not be changed, e.g. because there is no such Output of Status (e.g. WAIT for a send request), then the status my be changed into some undefined values. Therefore, we ware very careful for the mpi_f08 module never to add a risky INTENT(OUT). Best regards Rolf ----- Original Message -----
From: "MPI WG Point To Point Communications working group" <[email protected]> To: "MPI WG Point To Point Communications working group" <[email protected]> Cc: "wgropp" <[email protected]> Sent: Saturday, January 25, 2020 6:04:12 PM Subject: Re: [mpiwg-p2p] Review of Pythonization changes for Chapter 3
I believe this choice is because of MPI_STATUS_IGNORE - there is, in fact, both an IN value (don’t fill in the status) and an OUT value (place the status value here). We do note in the intro that the MPI IN and OUT do not exactly match Fortran, for these reasons.
Bill
William Gropp Director and Chief Scientist, NCSA Thomas M. Siebel Chair in Computer Science University of Illinois Urbana-Champaign
On Jan 22, 2020, at 5:07 AM, HOLMES Daniel via mpiwg-p2p < [ mailto:[email protected] | [email protected] ] > wrote:
3) The F08 bindings for all variants of the MPI_TEST, MPI_WAIT, and MPI_PROBE functions and the MPI_REQUEST_GET_STATUS function do not specify an INTENT for the status(es) parameter. We could fix this by including “INTENT(OUT)” in all cases (by choosing the correct Python setting).
_______________________________________________ mpiwg-p2p mailing list [email protected] https://lists.mpi-forum.org/mailman/listinfo/mpiwg-p2p
-- Dr. Rolf Rabenseifner . . . . . . . . . .. email [email protected] . High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530 . University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832 . Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner . Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307) .