October 2023 - mpi-forum - mailman3.cels.anl.gov

Errata reading for December 23 meeting
by Pritchard Jr., Howard 30 Oct '23

30 Oct '23

HI All, We had a PR that got stuck and did not get into the current MPI-41-RC. As a plan B, I’d like to do an errata reading at the December ’23 meeting for a change-log related item: Pull request https://github.com/mpi-forum/mpi-standard/pull/899 Related issue: https://github.com/mpi-forum/mpi-issues/issues/778 Thanks, Howard ------- [signature_61897647] Howard Pritchard Research Scientist HPC-ENV Los Alamos National Laboratory howardp(a)lanl.gov<mailto:[email protected]> [signature_568045215]<https://www.instagram.com/losalamosnatlab/>[signature_3647202021]<https://twitter.com/LosAlamosNatLab>[signature_254260304]<https://www.linkedin.com/company/los-alamos-national-laboratory/>[signature_3176230336]<https://www.facebook.com/LosAlamosNationalLab/>

1 0

Reminder: Special Voting Meeting this week
by Martin Schulz 29 Oct '23

29 Oct '23

Hi all, Just as a quick reminder, we have our special voting meeting for MPI 4.1 this week. If you plan to attend, but haven’t registered, yet, please do so as soon as possible The link is, as usual, on the forum logistics page. Also, please note: As usual, we meet from 9am to 1pm US central time (zoom link from the agenda) As daylight savings time ended in Europe over the weekend, this is now 3-7pm European Standard Time We meet Tuesday-Friday (not Monday-Thursday) See you all on Tuesday, 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

Reading Announcement
by Jim Dinan 19 Oct '23

19 Oct '23

Hi All, If time permits, the Hybrid WG would like to do an informal reading of the memory allocation kinds side document at the upcoming meeting. https://github.com/mpi-forum/mpi-issues/issues/809 Cheers, ~Jim.

1 0

Reading announcement for Reinit (slice 1)
by Ignacio Laguna 17 Oct '23

17 Oct '23

The MPI FT group would like to read the Coarse-grained fault-tolerance interface (Reinit, slice 1). This is for MPI v5: Issue: https://github.com/mpi-forum/mpi-issues/issues/578 PR: https://github.com/mpi-forum/mpi-standard/pull/711 I recommend reading the PDF chapter posted at the top of the issue: https://github.com/mpi-forum/mpi-issues/files/12703802/reinit.pdf Right now it's a separate chapter, but once approved we plan to integrate it in the already voted-in FT chapter (from ULFM slice 1). -- Ignacio Laguna Center for Applied Scientific Computing (CASC) Lawrence Livermore National Laboratory Phone: 925-422-7308, Fax: 925-422-6287

1 0

MPI 4.1 RC2 Document and Upcoming Meeting
by Wes Bland 17 Oct '23

17 Oct '23

Hi folks, Sorry about the delay on getting the new version of the 4.1 RC document out for the upcoming meeting. It was very helpful to have the last virtual meeting to make a better document that is more likely to be adopted at the next meeting. The newest version of the document is now posted on the docs page: https://www.mpi-forum.org/docs/ Direct link: https://www.mpi-forum.org/docs/mpi-4.1/mpi-41-rc2.pdf Please take a look and let us (Bill, Martin, and/or myself) know if you see any showstoppers. Anything that isn’t a complete blocker at this point will be deferred to either an errata or a future version of the standard if the proposed document is adopted at the upcoming meeting. Speaking of the upcoming meeting, the full agenda with all of the votes and readings is now posted on the website: https://www.mpi-forum.org/meetings/2023/10/agenda. If you feel that I missed anything, please let me know. The only thing that may change at this point is that some of the readings might be marked as getting an errata vote, but need to be changed into a no-no vote. Remember that everything that was on Martin’s list (https://docs.google.com/spreadsheets/d/1ZpsyVTnfNHv85K9Mzi_3niNXdmQ67yV2iwS…) will be voted on with errata rules and anything that isn’t will be voted on with no-no rules. I’ll go through the list soon to make sure I have them categorized correctly, but feel free to let me know if you see something wrong. The only things that are exempted from being voted on are the “editor changes” from Bill to fix whitespace, formatting, etc. You can find those by looking for any of Bill’s recent pull requests, but voting on those would be difficult as they are often needed after merging any other PR after it reflows the text. Today is also the deadline for officer nominations and any other readings for the October meeting. For anyone who is interested in being an officer, please have a friend nominate you and you can still be on the ballot. Lastly, please make sure you’ve registered for the upcoming meeting. The current list of registrants is here: https://www.mpi-forum.org/meetings/2023/10/attendance. Please double check that you’re on there and sign up here if not: https://forms.gle/tPvZNmT7ob7rNtqJ6 Thanks, Wes

1 0

Re: [Mpi-forum] MPI Forum Virtual Meeting Wednesday on MPI_Psend_init_c & MPI_Precv_init_c
by Martin Schulz 11 Oct '23

11 Oct '23

Hi all, Quick reminder – talk to you all in 8 minutes. 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 From: Martin Schulz <schulzm(a)in.tum.de> Date: Tuesday, 10. October 2023 at 02:05 To: Benson Muite via mpi-forum <mpi-forum(a)lists.mpi-forum.org> Subject: MPI Forum Virtual Meeting Wednesday on MPI_Psend_init_c & MPI_Precv_init_c Hi all, We are getting close to the final version of MPI 4.1, but there is still the ongoing discussion on what to do with the mistake/decision that we took on the count parameters in MPI_Psend_init_c & MPI_Precv_init_c. For those who have not been following this, here is the link to the github thread: https://github.com/mpi-forum/mpi-issues/issues/765 We need to settle this one way or the other, but a growing email thread does not seem to be the most effective way. Wesley and I are therefore suggesting to use our traditional meeting slot on Wednesday to hold a virtual meeting on this (Wednesday Oct. 11th, 10am US CT). Call-in information will be, as usual, on the forum website. I hope this works for most of you and hopefully we’ll find an easy way to resolve this. Thanks! Martin (and Wesley) -- 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

MPI Forum Virtual Meeting Wednesday on MPI_Psend_init_c & MPI_Precv_init_c
by Martin Schulz 09 Oct '23

09 Oct '23

Hi all, We are getting close to the final version of MPI 4.1, but there is still the ongoing discussion on what to do with the mistake/decision that we took on the count parameters in MPI_Psend_init_c & MPI_Precv_init_c. For those who have not been following this, here is the link to the github thread: https://github.com/mpi-forum/mpi-issues/issues/765 We need to settle this one way or the other, but a growing email thread does not seem to be the most effective way. Wesley and I are therefore suggesting to use our traditional meeting slot on Wednesday to hold a virtual meeting on this (Wednesday Oct. 11th, 10am US CT). Call-in information will be, as usual, on the forum website. I hope this works for most of you and hopefully we’ll find an easy way to resolve this. Thanks! Martin (and Wesley) -- 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

Reading announcement Dec '23 meeting
by Benson Muite 06 Oct '23

06 Oct '23

Have done an initial conversion of all figures to use TikZ. Would like to discuss these changes in the December meeting. Can also do chapter meetings as needed for chapters 2 (Terms), 6 (Collectives), 7 (Groups), 8 (Topologies), 9 (Environments), 11 (Processes), 12 (One-Sided), 14 (I/O) and 19 (Language Bindings). Issue: https://github.com/mpi-forum/mpi-issues/issues/712 Draft: https://github.com/mpi-forum/mpi-standard/pull/861 PDF with updated figures: https://github.com/mpi-forum/mpi-standard/files/12834910/mpi41-report.pdf Benson

1 0