May 2018 - mpi-forum - mailman3.cels.anl.gov

Reading announcement for June 2018 - issue 83
by Skjellum, Anthony 27 May '18

27 May '18

Hi all, The Persistent WG would like to read issue 83 (PR 35), which adds new a function to persistent collective communication - one communicator-scoped info key. This is an addendum to Issue #25. Issue: <https://github.com/mpi-forum/mpi-issues/issues/> https://github.com/mpi-forum/mpi-issues/issues/83 PR: <https://github.com/mpi-forum/mpi-standard/pull/> https://github.com/mpi-forum/mpi-standard/pull/35 PDF: https://github.com/mpi-forum/mpi-standard/files/2043271/mpi32-report-ticket… Thank you, Tony Skjellum

1 0

Tools WG Topics for face-to-face meetings
by Marc-Andre Hermanns 25 May '18

25 May '18

Dear all, the tools WG would like to have Ticket #67 Refined references to MPI processes throughout the Tools chapter up for a 'no-no' and subsequent 'first-vote'. Relevant links: https://github.com/mpi-forum/mpi-standard/pull/30 https://github.com/mpi-forum/mpi-standard/pull/30 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

Draft agenda (hopefully) complete
by Martin Schulz 24 May '18

24 May '18

Hi all, I complete the draft agenda for the Austin meeting. It is available at the usual place: https://www.mpi-forum.org/meetings/2018/06/agenda <https://www.mpi-forum.org/meetings/2018/06/agenda> First, please take a look at it and please let me know as soon as possible if something is missing or incorrect. Also, there are a lot of readings and ballots and I think all of them have been announced separately, but just in case - I will consider what is on the agenda now as officially announced in time. If you want to add something, please keep in mind that the deadline is Monday! As you can see, I did get a lot of requests for WG time and for plenary items this time, which is a great sign and fits with the discussion on the timeline for the next standard, which I want to have in Austin, but it also makes the schedule very compressed. When we decided to reduce to basically 3 days, we all agreed to, worst case if needed, lengthen the days a bit and I want to make use of that this time. Consequently, please note a) that we will start at 8:30 on Wed/Thu (I will leave it to the WGs on Tuesdays if the want to do that as well), b) we will run to 6pm or 6:30pm as needed on Tue/Wed, c) I moved general discussions like locations to the lunch slot, since these conversations end up in a more social discussion anyway, d) we have some limited plenaries already on Tuesday after the WGs, but also more WG time on Wednesday morning and e) I do expect that we run right up the end time of 12:30 on Thursday. Please plan accordingly. As for the plenary, we need to get through all the readings this time as to not stall - we will try to get the updates done as well, but if we run out of time, we will move them to virtual meetings. Thanks and looking forward to a packed and productive meeting! See many of you soon, Martin PS: Please keep the discussion on future locations in mind and come with location suggestions as well as date constraints. — Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching Email: schulzm(a)in.tum.de

1 0

Reading announcement for June 2018 - issue 94
by HOLMES Daniel 23 May '18

23 May '18

Hi all, The Sessions WG would like to read issue 94 (PR 45), which adds new a function to open a file from parent group without needing a parent communicator. Issue: https://github.com/mpi-forum/mpi-issues/issues/94 PR: https://github.com/mpi-forum/mpi-standard/pull/45 PDF: https://github.com/mpi-forum/mpi-issues/files/2031275/mpi-report-ticket94.p… 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 3415, JCMB, The King’s Buildings, Edinburgh, EH9 3FD — 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. _______________________________________________ mpi-forum mailing list mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> https://lists.mpi-forum.org/mailman/listinfo/mpi-forum The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Reading announcement for June 2018 - issue 93
by HOLMES Daniel 23 May '18

23 May '18

Hi all, The Sessions WG would like to read issue 93 (PR 44), which adds new functions to create communicators from parent groups without needing a parent communicator. Issue: https://github.com/mpi-forum/mpi-issues/issues/93 PR: https://github.com/mpi-forum/mpi-standard/pull/44 PDF: https://github.com/mpi-forum/mpi-issues/files/2031167/mpi-report-ticket93.p… 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 3415, JCMB, The King’s Buildings, Edinburgh, EH9 3FD — 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. _______________________________________________ mpi-forum mailing list mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> https://lists.mpi-forum.org/mailman/listinfo/mpi-forum The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

1 0

Announcement of issue 72, was Re: Reading announcement for June 2018 - issue 87
by schulzm＠in.tum.de 22 May '18

22 May '18

Hi Dan, Thanks for picking up this issue. All, In the process of cleaning up older issue, there is a second older issue (72), which falls into a similar category as 87, which Dan just listed. This issue https://github.com/mpi-forum/mpi-issues/issues/72 <https://github.com/mpi-forum/mpi-issues/issues/72> with PR https://github.com/mpi-forum/mpi-standard/pull/41 <https://github.com/mpi-forum/mpi-standard/pull/41> on the topic "Behavior of MPI_Dims_create when ndims is zero” is also ready and I will read it at the Austin forum meeting. For both issues, though it is not quite clear (I.e., there are different opinions) if they are ticket 0 clarifications, errata or actual changes requiring two votes. Unless I hear objections, I will therefore schedule both tickets for both a reading and an errata ballot, so we can pick in Austin what to do with them. If anyone has any comments on this topic, please comment on the two issues on github. Thanks! Martin — Prof. Dr. Martin Schulz, Chair of Computer Architecture and Parallel Systems Department of Informatics, TU-Munich, Boltzmannstraße 3, D-85748 Garching Email: schulzm(a)in.tum.de > On 22. May 2018, at 15:13, HOLMES Daniel <d.holmes(a)epcc.ed.ac.uk> wrote: > > Hi all, > > The Collectives WG would like to read issue 87 (PR 43), which is an errata that adds a clarification regarding the use of MPI_PROC_NULL as a valid rank value for graph and dist graph topologies. > > Issue: https://github.com/mpi-forum/mpi-issues/issues/87 <https://github.com/mpi-forum/mpi-issues/issues/87> > PR: https://github.com/mpi-forum/mpi-standard/pull/43 <https://github.com/mpi-forum/mpi-standard/pull/43> > PDF: https://github.com/mpi-forum/mpi-standard/files/2026783/mpi-report-ticket87… <https://github.com/mpi-forum/mpi-standard/files/2026783/mpi-report-ticket87…> > > 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 3415, JCMB, The King’s Buildings, Edinburgh, EH9 3FD > — > 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. > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum

1 0

Reading announcement for June 2018 - issue 87
by HOLMES Daniel 22 May '18

22 May '18

Hi all, The Collectives WG would like to read issue 87 (PR 43), which is an errata that adds a clarification regarding the use of MPI_PROC_NULL as a valid rank value for graph and dist graph topologies. Issue: https://github.com/mpi-forum/mpi-issues/issues/87 PR: https://github.com/mpi-forum/mpi-standard/pull/43 PDF: https://github.com/mpi-forum/mpi-standard/files/2026783/mpi-report-ticket87… 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 3415, JCMB, The King’s Buildings, Edinburgh, EH9 3FD — 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

Reading announcement for June 2018 - issue 86
by HOLMES Daniel 22 May '18

22 May '18

Hi all, The Point-to-point WG would like to read issue 86 (PR 42), which is an errata that adds a clarification regarding the minimum size of the array arguments passed to the MPI_Waitsome and MPI_Testsome functions. Issue: https://github.com/mpi-forum/mpi-issues/issues/86 PR: https://github.com/mpi-forum/mpi-standard/pull/42 PDF: https://github.com/mpi-forum/mpi-standard/files/2026613/mpi-report-ticket86… 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 3415, JCMB, The King’s Buildings, Edinburgh, EH9 3FD — 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

Reading for MPI issue #51: Deprecate MPI_SIZEOF in MPI-4.0
by Jeff Squyres (jsquyres) 21 May '18

21 May '18

There will be a (re)reading of issue #51 (deprecate MPI_SIZEOF Fortran interface). This issue was actually read about 2 years ago and had strong support in the Forum (straw vote of 25-0-2 in favor), but a) there were a few ticket 0 changes during the reading, and b) we let the issue drop and never actually voted on it. It's been so long that it should just be read again to refresh everyone's memory. I have rebased the PDF on the top of the mpi-3.1 branch. It will be read again at the upcoming Austin meeting. Here's the issue: https://github.com/mpi-forum/mpi-issues/issues/51 -- Jeff Squyres jsquyres(a)cisco.com

1 0

Reading for MPI Issue #91 - MPI_STATUS_SET_CANCELED F08 binding: flag param intent
by Bland, Wesley 21 May '18

21 May '18

Hi all, I'll be reading the very small mpi-forum/mpi-issues#91<https://github.com/mpi-forum/mpi-issues/issues/91> on behalf of the Fortran working group at the upcoming meeting in Austin. The accompanying pull request can be found at mpi-forum/mpi-standard#36<https://github.com/mpi-forum/mpi-standard/pull/36>. Thanks, Wesley

1 0