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

Agenda for March Meeting
by Martin Schulz 27 Feb '23

27 Feb '23

Hi all, I have updated the March meeting agenda based on the announcements. The agenda is at: https://www.mpi-forum.org/meetings/2023/03/agenda Please let me know if I missed an item. 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 Centre (LRZ) Email: schulzm(a)in.tum.de

1 0

Re: [Mpi-forum] Memory Alloc Kind Reading
by Jim Dinan 27 Feb '23

27 Feb '23

Hi All, Wanted to share an update on this reading. I added an update to the proposal yesterday (before cutoff) that introduces an explicit "default" memory kind. My hope is that this will address several concerns that were raised at the virtual meeting. Namely, 1. Use cases where users want to add a kind in addition to the default were challenging/confusing and are now easy ("default;additional_kinds") 2. The default kind must include "system" and "mpi". After re-reading Section 9.2, I believe this is not introducing a new constraint. 3. MPI may not ignore a request to support the default kind. This relieves users from checking for basic support they have relied on for many years. If you've already read the proposal in preparation for the meeting, you can find these changes in commit 80a67b3. Cheers, ~Jim. On Fri, Feb 17, 2023 at 5:38 PM Jim Dinan <james.dinan(a)gmail.com> wrote: > Hi All, > > I would like to read the Memory Allocation Kind Info proposal at the > upcoming Forum meeting: > > https://github.com/mpi-forum/mpi-issues/issues/580 > > ~Jim. >

1 0

Additional reading at MPI-Forum meeting March 13-16, 2023 (related to Topologies)
by Rolf Rabenseifner 27 Feb '23

27 Feb '23

Dear MPI forum members, dear Wesley, in addition to my previous two announcment, Christoph and I would like to have the following issues (and their solving RPs) read: 683 Question: MPI_Dims_create - is "as close to each other as possible" a requirement? - https://github.com/mpi-forum/mpi-issues/issues/683 - https://github.com/mpi-forum/mpi-standard/pull/810 - https://github.com/mpi-forum/mpi-standard/files/10842980/mpi41-report_Issue… Re-reading: 120 MPI_Cart_create_weighted / Topology aware Cartesian communicators Rolf,Christoph - https://github.com/mpi-forum/mpi-issues/issues/120 - https://github.com/mpi-forum/mpi-standard/pull/98 - https://github.com/mpi-forum/mpi-standard/files/10842552/mpi41-report_Issue… Best regards Rolf and Christoph ----- Original Message ----- > From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> > To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> > Cc: "Wesley Bland" <work(a)wesbland.com>, "Christoph Niethammer" <niethammer(a)hlrs.de>, "Puri Bangalore" > <pvbangalore(a)ua.edu> > Sent: Saturday, February 25, 2023 5:22:16 PM > Subject: Votes and reading at MPI-Forum meeting March 13-16, 2023 (related to Terms and Topologies) > Dear MPI forum members, dear Wesley, > > in addition to my previous two announcment, > I would like to have the following issues (and their solving RPs) read: > > 278 Consistent use: Section 18.1.9 - p. 661 - Replace <TYPE> Rolf,Puri > - https://github.com/mpi-forum/mpi-issues/issues/278 > - https://github.com/mpi-forum/mpi-standard/pull/802 > - > https://github.com/mpi-forum/mpi-standard/files/10831642/mpi41-report_Issue… > > 485 Fix Incorrect Usage of Rank/Task/etc. in Language Bindings Chapter > Rolf,Puri > - https://github.com/mpi-forum/mpi-issues/issues/485 > - https://github.com/mpi-forum/mpi-standard/pull/803/files > - > https://github.com/mpi-forum/mpi-standard/files/10831598/mpi41-report_Issue… > > 457 Improvements around the word "rank" in the Process Topologies chapter > Christoph,Rolf > - https://github.com/mpi-forum/mpi-issues/issues/457 > - https://github.com/mpi-forum/mpi-standard/pull/804 > - > https://github.com/mpi-forum/mpi-standard/files/10831527/mpi41-report_Issue… > > Best regards > Rolf > > > ----- Original Message ----- >> From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> >> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >> Cc: "Wesley Bland" <work(a)wesbland.com> >> Sent: Friday, February 24, 2023 8:51:27 AM >> Subject: Votes and reading at MPI-Forum meeting March 13-16, 2023 (related to >> Terms and Topologies) > >> Dear MPI forum members, dear Wesley, >> >> in addition to my previous announcment, >> I would like to have the following issues (and their common PR 805) read: >> >> 235 Bindings - Shorten Example Index Names >> 607 Bindings - Add Example Titles >> 244 Bindings - Mark Erroneous Examples >> 610 Bindings - Fix Example Environment >> 633 Bindings - Update verbatim examples to example environment >> - All done together in the PR 805 >> https://github.com/mpi-forum/mpi-standard/pull/805/files >> >> I'll do final fixes and the pdf on next Monday before the 2-weeks-deadline. >> >> Special thanks to Bill Gropp for his great work on all his better formatting of >> the >> examples since 4.0 and marking the things that had to be fixed by the chapter >> committee. >> >> Best regards >> Rolf >> >> >> ----- Original Message ----- >>> From: "Rolf Rabenseifner" <rabenseifner(a)hlrs.de> >>> To: "Main MPI Forum mailing list" <mpi-forum(a)lists.mpi-forum.org> >>> Cc: "Wesley Bland" <work(a)wesbland.com> >>> Sent: Thursday, February 16, 2023 7:50:12 PM >>> Subject: Votes and reading at MPI-Forum meeting March 13-16, 2023 (related to >>> Terms and Topologies) >> >>> Dear MPI forum members, dear Wesley, >>> >>> I would like to have the following issues (and their PRs) voted or read. >>> >>> ================== 2nd votes related to MPI 4.1 ============== >>> 641 Including Annex A.2 back into the standard Rolf >>> 493 Issue N - Progress-Section in terms-2.tex Rolf >>> 471 Fix Incorrect Usage of Rank/Task/etc. in Terms Chapter Rolf >>> 565 Refer to Fortran 2018 instead of 2008+TS Rolf+Jeff H. >>> 644 Update 'Process Topology' chapter title to use 'MPI process' Christoph >>> >>> ================== 1st votes related to MPI 4.1 ================= >>> 543 'Pending operation' not defined, pending proper definition Rolf >>> >>> ================== Reading announcment =========================== >>> 669 Add operation state 'enabled' and 'local calls' into Terms Rolf >>> - Issue https://github.com/mpi-forum/mpi-issues/issues/669 >>> - PR https://github.com/mpi-forum/mpi-standard/pull/788/files >>> - PDF >>> https://github.com/mpi-forum/mpi-standard/files/10758712/mpi41-report_Issue… >>> >>> ================== Pre-Announcments for Reading ================== >>> 586 MPI_Buffer_attach and Sessions Martin,Jeff,Rolf >>> will first be discussed next week in the virtual MPI meeting Feb 22, 2023 >>> - Issue https://github.com/mpi-forum/mpi-issues/issues/586 >>> - PR https://github.com/mpi-forum/mpi-standard/pull/736/files >>> - PDF >>> https://github.com/mpi-forum/mpi-standard/files/10744036/mpi41-report_Issue… >>> >>> Best regards >>> Rolf -- Dr. Rolf Rabenseifner . . . . . . . . . .. . . rabenseifner(a)hlrs.de . High Performance Computing Center (HLRS) . . . ++49(0)711/685-65530 . University of Stuttgart . . . . . . www.hlrs.de/people/rabenseifner . Nobelstr. 19, 70569 Stuttgart, Germany

1 0

Reading: IO Chapter, Example environments
by Benson Muite 27 Feb '23

27 Feb '23

Hi, Would like to do a reading of pull request associated to: https://github.com/mpi-forum/mpi-issues/issues/606 It is labelled as chapter committee change, but can do reading just in case. Regards, Benson

1 0

MPI-4.1 announcements for Mar'23 voting meeting
by Holmes, Daniel John 27 Feb '23

27 Feb '23

Hi all, I'd like to announce a (re-)reading for the upcoming Mar'23 meeting. Issue: https://github.com/mpi-forum/mpi-issues/issues/639 PR: https://github.com/mpi-forum/mpi-standard/pull/750 Changes since last reading (basically a complete re-write of the affected sentence): https://github.com/mpi-forum/mpi-standard/pull/750/files/78353ac06a93a01db2… PDF (with highlights, p68): https://github.com/mpi-forum/mpi-issues/files/10785518/mpi41-report_Issue63… I would also like to request a 2nd vote for the following issues that have passed a 1st vote: Issue: https://github.com/mpi-forum/mpi-issues/issues/188 PR: https://github.com/mpi-forum/mpi-standard/pull/203 Summary: state that Example 3.10 is unsafe Issue: https://github.com/mpi-forum/mpi-issues/issues/83 PR: https://github.com/mpi-forum/mpi-standard/pull/666 Summary: new Info key for strict start order of persistent collective operations Issue: https://github.com/mpi-forum/mpi-issues/issues/472 PR: https://github.com/mpi-forum/mpi-standard/pull/606 Summary: disambiguate "rank" in the point-to-point chapter Issue: https://github.com/mpi-forum/mpi-issues/issues/117 PR: https://github.com/mpi-forum/mpi-standard/pull/748 Summary: "logically concurrent" might not be - advice All these are unchanged since their 1st vote state. Best wishes, Dan.

1 0

Reading announcement of revamped side-documents concept addition
by Ruefenacht, Martin 27 Feb '23

27 Feb '23

Dear all, I’d like to announce a new reading of the side-documents additions: Introduce the concept of side-documents Issue: https://github.com/mpi-forum/mpi-issues/issues/562 PR: https://github.com/mpi-forum/mpi-standard/pull/694 We have chosen to present this as a new reading, because the changes were in our view outside of a no-no vote. The changes include nicer flowing text, addition of where to find the side-documents, and an explicit permission mechanism requirement (as discussed with Jim Dinan in the previous MPI Forum voting meeting, reinforced by Joseph Schuchart). Cheers, Martin -- Dr. Martin Ruefenacht Quantum Computing and Technologies Leibniz Supercomputing Centre of the Bavarian Academy of Sciences and Humanities Boltzmannstr. 1 85748 Garching e-mail: martin.ruefenacht(a)lrz.de<mailto:[email protected]>

1 0

Reading + Votes for _x deprecation
by Ruefenacht, Martin 27 Feb '23

27 Feb '23

Dear all, I’d like to announce a no-no reading and no-no vote, and a first vote for the _x deprecation ticket. Changes made since last reading have been the fixing of the referencing errors. Deprecate all _x procedures Issue: https://github.com/mpi-forum/mpi-issues/issues/518 PR: https://github.com/mpi-forum/mpi-standard/pull/623 Cheers, Martin -- Dr. Martin Ruefenacht Quantum Computing and Technologies Leibniz Supercomputing Centre of the Bavarian Academy of Sciences and Humanities Boltzmannstr. 1 85748 Garching e-mail: martin.ruefenacht(a)lrz.de<mailto:[email protected]>

1 0

March Reading and Vote Announcements
by Joseph Schuchart 26 Feb '23

26 Feb '23

Dear Wes, all, On behalf of the RMA working group, I would like to request the following readings and votes. All PRs should have an auto-generated PDF attached. Reading: *RMA accumulate operations don't distinguish between throughput and latency sensitive applications* Issue: https://github.com/mpi-forum/mpi-issues/issues/640 PR: https://github.com/mpi-forum/mpi-issues/issues/640 This was read at the December'22 meeting but had substantial changes since so a new reading is required. First vote: *Example 12.4: RMA communication calls may not block in PSCW* Issue: https://github.com/mpi-forum/mpi-issues/issues/555 PR: https://github.com/mpi-forum/mpi-standard/pull/671 No-no-votes and first vote: *Make all RMA communication procedures local* Issue: https://github.com/mpi-forum/mpi-issues/issues/653 PR: https://github.com/mpi-forum/mpi-standard/pull/769 Changes since last readings: https://github.com/mpi-forum/mpi-standard/pull/769/files/2f1cd6274e30fdf8fc… The bulk of this change is reverting the accidental removal of text. *RMA: Examples 12.4 and 12.5 contain advice to implementors* Issue: https://github.com/mpi-forum/mpi-issues/issues/652 PR: https://github.com/mpi-forum/mpi-standard/pull/768 Changes since the last reading: https://github.com/mpi-forum/mpi-standard/pull/768/commits/9e55f4c666b5d60e… *Allow MPI_WIN_SHARED_QUERY on created and allocated windows* Issue: https://github.com/mpi-forum/mpi-issues/issues/23 PR: https://github.com/mpi-forum/mpi-standard/pull/708 Changes since the last reading: https://github.com/mpi-forum/mpi-standard/pull/708/files/56a8aaf855b890cbf0… Second votes: *WTIME and WTICK should be symbols* Issue: https://github.com/mpi-forum/mpi-issues/issues/646 PR: https://github.com/mpi-forum/mpi-standard/pull/771 *Handle conversions should be symbols* Issue: https://github.com/mpi-forum/mpi-issues/issues/647 PR: https://github.com/mpi-forum/mpi-standard/pull/770 Second vote with no-no-vote: *Deprecate mpif.h* Issue: https://github.com/mpi-forum/mpi-issues/issues/561 PR: https://github.com/mpi-forum/mpi-standard/pull/722 Changes since the last vote: https://github.com/mpi-forum/mpi-standard/pull/722/files/e9629890a6cba32a0d… Many thanks Joseph

1 0

Reading Announcements Buffer Send Handling
by Martin Schulz 26 Feb '23

26 Feb '23

Hi all, Adding to Rolf’s previous announcement of the reading of Issue 586 (MPI_Buffer_attach and Sessions): The issue/PR has been updated to reflect the discussion from last time Zero buffer size behavior is now the same as in MPI 4.0 Automatic buffering is enabled by using a constant for the buffer argument Small text updates No changes in default behavior Further, there are now two companion issues with PRs, which I would like to read as well: Issue 684: make the automatic buffering the default behavior Issue 685: adds iflush procedures Note: the PRs for those two issue contain all changes for 586 as well and will have to properly merged should they pass. Also, should 586 fail, these issues will be retracted. 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 Centre (LRZ) Email: schulzm(a)in.tum.de

1 0

Ballot Announcements (408, 411, 480 and 541)
by Martin Schulz 25 Feb '23

25 Feb '23

Hi all, I would like to announce the following ballots (and associated no/no readings) for the March meeting. Thanks, Martin 1st Vote on issue 408 Topic: added reference to the fact that hints must be adhered to by the app https://github.com/mpi-forum/mpi-issues/issues/408 No/No and 1st Vote on issue 411 Topic: Fix comment on Info object in incomplete functions https://github.com/mpi-forum/mpi-issues/issues/411 No/No and 1st Vote on issue 480 Topic: Fix Incorrect Usage of Rank/Task/etc. in External Interfaces Chapter https://github.com/mpi-forum/mpi-issues/issues/480 1st Vote on issue 541 Topic: Consistency Updates for MPI_Info Chapter for MPI 4.1 https://github.com/mpi-forum/mpi-issues/issues/541 -- 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 Centre (LRZ) Email: schulzm(a)in.tum.de

1 0