February 2019 - mpi-forum - mailman3.cels.anl.gov

Re: [Mpi-forum] Request for Reading: Clarify alignment of memory provided to applications
by Joseph Schuchart 25 Feb '19

25 Feb '19

Dear MPI forum members, I have made some minor changes to the proposal mentioned below, after the 2 week deadline (adding entries to the index and minor changes to the wording, no essential changes). I would be grateful if this update was considered for the reading, replacing the old document. The updated PDF can be found at https://github.com/mpi-forum/mpi-issues/files/2893889/mpi-report_issue-121_…, the PR has been updated accordingly and the issue has been updated to reference the new PDF version. Best regards, Joseph Schuchart On 2/18/19 1:17 PM, Joseph Schuchart via mpi-forum wrote: > Dear MPI forum members, > > [Apologies for double-posting, I did not receive a notification for my > previous mail and it does not show up in the archives] > > I would like to request a reading of the changes proposed in issue #121 > "Clarify alignment of memory provided to applications": > https://github.com/mpi-forum/mpi-issues/issues/121. > > The changes to the standard text are detailed in PR #96 and include a > requirement for proper alignment suitable for load/store accesses of > predefined MPI types and adds a new info key to allow users to request > bigger alignments: > https://github.com/mpi-forum/mpi-standard/pull/96/files. > > A PDF file with the proposed changes highlighted can be found here: > https://github.com/mpi-forum/mpi-issues/files/2874936/mpi-report.pdf > > This issue has been discussed in the RMA-WG before and I would like to > propose this for MPI 4.0. I am not familiar with the internal processes > of the MPI forum so please let me know if there are any issues with this > request. > > Many thanks in advance, > Joseph Schuchart

1 0

Agenda for March Meeting Posted
by Martin Schulz 20 Feb '19

20 Feb '19

Hi all, I just posted the draft agenda for the March forum meeting. Please let me know if missed anything. Thanks! Martin — Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching Member of the Board of Directors at the Leibniz Supercomputing Center (LRZ) Email: schulzm(a)in.tum.de

1 0

Virtual Meeting today cancelled
by Martin Schulz 20 Feb '19

20 Feb '19

Hi all, The virtual meeting that was (additionally) scheduled for today (in a few minutes) is cancelled. We will pick up the topic at the meeting in Chattanooga. Sorry for any inconvenience, Martin — Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching Member of the Board of Directors at the Leibniz Supercomputing Center (LRZ) Email: schulzm(a)in.tum.de

1 0

Re: [Mpi-forum] Fwd: FINAL REMINDER: Call for items for March Forum meetings
by Rolf Rabenseifner 19 Feb '19

19 Feb '19

----- Original Message ----- > From: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Cc: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de>, "Purushotham V. Bangalore" <puri(a)uab.edu> > Sent: Monday, February 18, 2019 11:20:25 AM > Subject: [Mpi-forum] Fwd: FINAL REMINDER: Call for items for March Forum meetings > Puri asked me to send his email again to the MPI Forum, because soemthing didn't > worked with the sending out yesterday. > > ________________________________ > From: Bangalore, Purushotham > Sent: Sunday, February 17, 2019 10:22 PM > To: Main MPI Forum mailing list > Cc: Martin Schulz > Subject: Re: [Mpi-forum] FINAL REMINDER: Call for items for March Forum meetings > > Hi All, > > We would like to do the first reading on the updates to the Semantic Terms > section. Please find the corresponding issue and PR info: > > > Issue: https://github.com/mpi-forum/mpi-issues/issues/96 > > PR: https://github.com/mpi-forum/mpi-standard/pull/97 > > You can find the corresponding complete PDF file of the standard at the issue or > at https://github.com/mpi-forum/mpi-issues/files/2874029/mpi32-report.pdf. > > I have attached the PDF files that Rolf created which summarize these changes > along with the appendix. > > Thanks, > Puri > > ________________________________ > From: mpi-forum <mpi-forum-bounces(a)lists.mpi-forum.org> on behalf of Martin > Schulz via mpi-forum <mpi-forum(a)lists.mpi-forum.org> > Sent: Sunday, February 17, 2019 6:03 AM > To: mpi-forum(a)lists.mpi-forum.org > Cc: Martin Schulz > Subject: [Mpi-forum] FINAL REMINDER: Call for items for March Forum meetings > > Hi all, > > As a reminder: tomorrow (Monday) is the deadline for the announcements of > reading and ballots for the MPI forum meeting in Chattanooga. Please send any > announcements, together with the PR number and a reference to the matching PDF > to the MPI forum list. > > If there other agenda items aside from readings, please send them to me as well > so I can add them to the agenda. > > Thanks! > > Martin > > > > Sent from my iPad > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum > -- > Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de . > 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) . > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de . 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) .

1 0

Request for Reading: Errata for Passive Target RMA on shared memory #61
by Rolf Rabenseifner 19 Feb '19

19 Feb '19

Hi all, I would like to read the errata for passive target RMA on shared memory. The issue is #61: https://github.com/mpi-forum/mpi-issues/issues/61 PR and pdf are still missing, but the ticket shows the change very clear, and they should come soon. I hope that this is okay. Although the problem and solution is well known since Sep. 2016, this errate was not yet pushed into the standard. I hope that it will work now (still on Monday, Hawaiian time) for MPI-4.0. Best regards Rolf -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de . 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) .

1 0

Re: [Mpi-forum] Request for Reading: Enhancement to Intro for nonblocking
by Rolf Rabenseifner 18 Feb '19

18 Feb '19

And an annotated version, see https://github.com/mpi-forum/mpi-issues/files/2876926/mpi-report-ticket122-… ----- Original Message ----- > From: "HOLMES Daniel" <d.holmes(a)epcc.ed.ac.uk> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Cc: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > Sent: Monday, February 18, 2019 6:16:01 PM > Subject: Re: [Mpi-forum] Request for Reading: Enhancement to Intro for nonblocking > Hi all, > > The Point-to-point WG would like to read a proposal to enhance the introduction > to nonblocking communication. > > The issue is #122: [ https://github.com/mpi-forum/mpi-issues/issues/122 | > https://github.com/mpi-forum/mpi-issues/issues/122 ] > The PR is #99: [ https://github.com/mpi-forum/mpi-standard/pull/99 | > https://github.com/mpi-forum/mpi-standard/pull/99 ] > The PDF is here: [ > https://github.com/mpi-forum/mpi-issues/files/2876759/mpi-report-ticket122.… > | > https://github.com/mpi-forum/mpi-issues/files/2876759/mpi-report-ticket122.… > ] > > Cheers, > Dan. > — > Dr Daniel Holmes PhD > Applications Consultant in HPC Research > [ mailto:[email protected] | d.holmes(a)epcc.ed.ac.uk ] > Phone: +44 (0) 131 651 3465 > Mobile: +44 (0) 7940 524 088 > Address: Room 2.09, Bayes Centre, 47 Potterrow, Central Area, Edinburgh, EH8 9BT > — > The University of Edinburgh is a charitable body, registered in Scotland, with > registration number SC005336. > — > > The University of Edinburgh is a charitable body, registered in > Scotland, with registration number SC005336. -- Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner(a)hlrs.de . 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) .

1 0

Request for Reading: Callback-driven event interface for MPI_T
by Marc-Andre Hermanns 18 Feb '19

18 Feb '19

Hi all, The Tools WG would like to read the MPI_T events proposal. The Issue is #79: https://github.com/mpi-forum/mpi-issues/issues/79 The PR is #33: https://github.com/mpi-forum/mpi-standard/pull/33 The PDF is here: https://github.com/mpi-forum/mpi-issues/files/2876870/mpi32-report-20190218… Cheers, Marc-Andre -- Dr. Marc-Andre Hermanns Jülich Aachen Research Alliance, High Performance Computing (JARA-HPC) Jülich Supercomputing Centre (JSC) Wilhelm-Johnen-Str. 52425 Jülich Germany Phone: +49 2461 61 2509 | +49 241 80 24381 Fax: +49 2461 80 6 99753 www.jara.org/jara-hpc email: hermanns(a)jara.rwth-aachen.de

1 0

Re: [Mpi-forum] Request for Reading: Enhancement to Intro for nonblocking
by HOLMES Daniel 18 Feb '19

18 Feb '19

Hi all, The Point-to-point WG would like to read a proposal to enhance the introduction to nonblocking communication. The issue is #122: https://github.com/mpi-forum/mpi-issues/issues/122 The PR is #99: https://github.com/mpi-forum/mpi-standard/pull/99 The PDF is here: https://github.com/mpi-forum/mpi-issues/files/2876759/mpi-report-ticket122.… Cheers, Dan. — Dr Daniel Holmes PhD Applications Consultant in HPC Research d.holmes(a)epcc.ed.ac.uk<mailto:[email protected]> Phone: +44 (0) 131 651 3465 Mobile: +44 (0) 7940 524 088 Address: Room 2.09, Bayes Centre, 47 Potterrow, Central Area, Edinburgh, EH8 9BT — The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. — The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Re: [Mpi-forum] Request for Reading: The Peer to Peer Process Creation Model (AKA Sessions)
by HOLMES Daniel 18 Feb '19

18 Feb '19

Hi all, The Sessions WG would like to read the sessions proposal (again). The issue is #103: https://github.com/mpi-forum/mpi-issues/issues/103 The PR is #55: https://github.com/mpi-forum/mpi-standard/pull/55 The PDF is here: https://github.com/mpi-forum/mpi-issues/files/2876679/mpi-report-ticket103-… Cheers, Dan. — Dr Daniel Holmes PhD Applications Consultant in HPC Research d.holmes(a)epcc.ed.ac.uk<mailto:[email protected]> Phone: +44 (0) 131 651 3465 Mobile: +44 (0) 7940 524 088 Address: Room 2.09, Bayes Centre, 47 Potterrow, Central Area, Edinburgh, EH8 9BT — The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. — The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

FTWG Readings Announcements
by Bland, Wesley 18 Feb '19

18 Feb '19

Hi all, The Fault Tolerance Working Group will be doing 2 readings at the upcoming meeting. 1. Define error/failure behavior in MPI_INIT/FINALIZE * Issue - https://github.com/mpi-forum/mpi-issues/issues/102 * Pull Request - https://github.com/mpi-forum/mpi-standard/pull/50 This issue pulls out the definitions around error behavior for `MPI_INIT` and `MPI_FINALIZE`. This comes from #20, but is not directly related to the fault tolerance parts of that proposal. 2. Clarify what MPI_SUCCESS guarantees, and what is left undefined when MPI_ERR_XXX happens * Issue - https://github.com/mpi-forum/mpi-issues/issues/119 * Pull Request - https://github.com/mpi-forum/mpi-standard/pull/95 Add text to clarify what is implied about the status of MPI and user visible buffers when MPI functions return `MPI_SUCCESS` or other error codes. Thanks, Wesley

1 0