July 2021 - mpi-forum - mailman3.cels.anl.gov

Re: [Mpi-forum] Next Virtual MPI Forum Meeting Next Week - Request for MPI 4.1 issues/topics
by Gropp, William D 22 Jul '21

22 Jul '21

I’ve been looking at modernizing and simplifying some of the LaTeX for managing the document. These do result in some changes to the appearance, which I’d like to discuss to see if this is the direction the MPI Forum would like. The biggest change is in line numbering (the numbers along the side of the page). There is a nice LaTeX package for numbering lines of text, and it is pretty accurate - numbers are aligned with the text, not off some fraction as happens with the current number “ruler”. But a disadvantage is that *only* lines are numbered - figures (and their captions) aren’t numbered (or have just one number, depending on exactly how they are done), and the nice boxes around the examples also prevent numbering, so I replaced those with simpler horizontal rules in the samples. I’ve attached examples to show what this looks like. Bill William Gropp Director, NCSA Thomas M. Siebel Chair in Computer Science University of Illinois Urbana-Champaign IEEE-CS President-Elect On Jul 14, 2021, at 8:42 AM, Martin Schulz via mpi-forum <mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]>> wrote: Hi all, As mentioned last week, we'll pick up our Virtual MPI Forum Meetings again next week on July 21st. As for the 21st: I would like to use the meeting to set up a (at least preliminary) goal-line for MPI 4.1 - we had several open items from the 4.0 process (and before), which will be the starting point. If there are any particular items you think we should address as part of a "clean up" or "minor functionality" please either send them to me and I will collect them or add them to the system as a blank ticket with the MPI 4.1 tag. Next week we'll consolidate the list and decide how we approach this. 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<mailto:[email protected]> _______________________________________________ mpi-forum mailing list mpi-forum(a)lists.mpi-forum.org<mailto:[email protected]> https://urldefense.com/v3/__https://lists.mpi-forum.org/mailman/listinfo/mp…

1 0

Reminder: Next Virtual MPI Forum Meeting today - Discussion on MPI 4.1 issues/topics
by Martin Schulz 21 Jul '21

21 Jul '21

Hi all, As mentioned last week, we'll pick up our MPI Forum meeting series again - usual time 10am CT, link as always on the forum webpage. 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 On 14.07.21, 15:42, "Martin Schulz" <schulzm(a)in.tum.de> wrote: Hi all, As mentioned last week, we'll pick up our Virtual MPI Forum Meetings again next week on July 21st. As for the 21st: I would like to use the meeting to set up a (at least preliminary) goal-line for MPI 4.1 - we had several open items from the 4.0 process (and before), which will be the starting point. If there are any particular items you think we should address as part of a "clean up" or "minor functionality" please either send them to me and I will collect them or add them to the system as a blank ticket with the MPI 4.1 tag. Next week we'll consolidate the list and decide how we approach this. 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] Next Virtual MPI Forum Meeting Next Week - Request for MPI 4.1 issues/topics
by Guillaume Mercier 21 Jul '21

21 Jul '21

Hello Martin, Sorry to answer your message so late. There are a couple of things I'd like to address for 4.1 1- all the clean-up stuff w.r.t (now) chapter 7, as I presented during my reading in the last ratification meeting. This include (see attached document): * Add process topology functions that create communicators to the various lists throughout the chapter -> Issue #497, PR yet to create * Remove/improve text for callback-copy functions -> Issue #496, PR yet to create * Improve semantic index -> Issue #237, PR yet to create * Library examples to reformat -> Issue #229 (continued), PR yet to create * Process → MPI Process where relevant -> Issue #??? → PR#590 * Fix incorrect usage of ranks/tasks/etc -> Issue #474 → PR#566 2- I'd like to resurrect issue #154 (A new function to query hardware resource names, https://github.com/mpi-forum/mpi-issues/issues/154 ) with PR#143. If you recall, people were OK with this in Albuquerque on the substance but not on the interface. Using an MPI-defined new datastruct as a parameter was suggested. That's was I should have done a while ago and that is to be proposed. 4.1 seems a good target but I might be wrong on this. I'll do my best to attend today's meeting but I will have to leave for a while because of another engagement that I can't postpone nor defer. Thanks, Guillaume On 7/14/21 3:42 PM, Martin Schulz via mpi-forum wrote: > Hi all, > > As mentioned last week, we'll pick up our Virtual MPI Forum Meetings again next week on July 21st. > > As for the 21st: I would like to use the meeting to set up a (at least preliminary) goal-line for MPI 4.1 - we had several open items from the 4.0 process (and before), which will be the starting point. If there are any particular items you think we should address as part of a "clean up" or "minor functionality" please either send them to me and I will collect them or add them to the system as a blank ticket with the MPI 4.1 tag. Next week we'll consolidate the list and decide how we approach this. > > Thanks! > > Martin >

1 0

MPI Languages Working Group Announcement
by Ruefenacht, Martin 19 Jul '21

19 Jul '21

Hello all, We are announcing the start of our Languages WG meetings for Thursday 22nd of July at 15:00 CEST, 09:00 EST, 08:00 CST continuing every other week. Please join with this link: https://meet.lrz.de/MPILanguagesWG We have a GitHub organization at https://github.com/mpiwg-languages. Please ping me directly to be added to the slack channel. We will be considering a set of MPI-4.1 and MPI-5 proposals, so please bring your topics and ideas to this meeting. Thank you, Martin Ruefenacht

1 0

Next Virtual MPI Forum Meeting Next Week - Request for MPI 4.1 issues/topics
by Martin Schulz 14 Jul '21

14 Jul '21

Hi all, As mentioned last week, we'll pick up our Virtual MPI Forum Meetings again next week on July 21st. As for the 21st: I would like to use the meeting to set up a (at least preliminary) goal-line for MPI 4.1 - we had several open items from the 4.0 process (and before), which will be the starting point. If there are any particular items you think we should address as part of a "clean up" or "minor functionality" please either send them to me and I will collect them or add them to the system as a blank ticket with the MPI 4.1 tag. Next week we'll consolidate the list and decide how we approach this. 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] MPI_Get_elements_c
by George Bosilca 13 Jul '21

13 Jul '21

As the _c API provides exactly the same functionality as the _x when are we deprecating the _X API ? George. On Wed, Jul 7, 2021 at 1:23 PM Dan Holmes via mpi-forum < mpi-forum(a)lists.mpi-forum.org> wrote: > Hi Ken, > > Looks like an oversight to me. > > It needs a fix in "chap-ei/ei-2.tex” on line 575 of 668 > from "XFER_NUM_ELEM_SMALL”, > to “POLYXFER_NUM_ELEM_NNI", > > Cheers, > Dan. > — > Dr Daniel Holmes PhD > Executive Director > Chief Technology Officer > CHI Ltd > danholmes(a)chi.scot > > > > On 7 Jul 2021, at 15:37, Raffenetti, Kenneth J. via mpi-forum < > mpi-forum(a)lists.mpi-forum.org> wrote: > > Hi, > > This issue was raised in the MPICH GitHub > https://github.com/pmodels/mpich/issues/5413. > > In MPI-3.1, we have C bindings for: > > int MPI_Get_elements(const MPI_Status *status, MPI_Datatype datatype, int > *count) > int MPI_Get_elements_x(const MPI_Status *status, MPI_Datatype datatype, > MPI_Count *count) > int MPI_Status_set_elements(MPI_Status *status, MPI_Datatype datatype, int > count) > int MPI_Status_set_elements_x(MPI_Status *status, MPI_Datatype datatype, > MPI_Count count) > > MPI-4.0 added: > > int MPI_Get_elements_c(const MPI_Status *status, MPI_Datatype datatype, > MPI_Count *count) > > But no binding was added for an _c version of MPI_Status_set_elements. Was > this intended? > > Thanks, > Ken > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum > > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum >

1 0

Re: [Mpi-forum] MPI_Get_elements_c
by Dan Holmes 07 Jul '21

07 Jul '21

Hi Ken, Looks like an oversight to me. It needs a fix in "chap-ei/ei-2.tex” on line 575 of 668 from "XFER_NUM_ELEM_SMALL”, to “POLYXFER_NUM_ELEM_NNI", Cheers, Dan. — Dr Daniel Holmes PhD Executive Director Chief Technology Officer CHI Ltd danholmes(a)chi.scot > On 7 Jul 2021, at 15:37, Raffenetti, Kenneth J. via mpi-forum <mpi-forum(a)lists.mpi-forum.org> wrote: > > Hi, > > This issue was raised in the MPICH GitHub https://github.com/pmodels/mpich/issues/5413. > > In MPI-3.1, we have C bindings for: > > int MPI_Get_elements(const MPI_Status *status, MPI_Datatype datatype, int *count) > int MPI_Get_elements_x(const MPI_Status *status, MPI_Datatype datatype, MPI_Count *count) > int MPI_Status_set_elements(MPI_Status *status, MPI_Datatype datatype, int count) > int MPI_Status_set_elements_x(MPI_Status *status, MPI_Datatype datatype, MPI_Count count) > > MPI-4.0 added: > > int MPI_Get_elements_c(const MPI_Status *status, MPI_Datatype datatype, MPI_Count *count) > > But no binding was added for an _c version of MPI_Status_set_elements. Was this intended? > > Thanks, > Ken > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum

1 0

Re: [Mpi-forum] MPI_Get_elements_c
by Jeff Hammond 07 Jul '21

07 Jul '21

Related: do we have an F08 overload for this or not? Jeff > On Jul 7, 2021, at 5:37 PM, Raffenetti, Kenneth J. via mpi-forum <mpi-forum(a)lists.mpi-forum.org> wrote: > > Hi, > > This issue was raised in the MPICH GitHub https://github.com/pmodels/mpich/issues/5413. > > In MPI-3.1, we have C bindings for: > > int MPI_Get_elements(const MPI_Status *status, MPI_Datatype datatype, int *count) > int MPI_Get_elements_x(const MPI_Status *status, MPI_Datatype datatype, MPI_Count *count) > int MPI_Status_set_elements(MPI_Status *status, MPI_Datatype datatype, int count) > int MPI_Status_set_elements_x(MPI_Status *status, MPI_Datatype datatype, MPI_Count count) > > MPI-4.0 added: > > int MPI_Get_elements_c(const MPI_Status *status, MPI_Datatype datatype, MPI_Count *count) > > But no binding was added for an _c version of MPI_Status_set_elements. Was this intended? > > Thanks, > Ken > > _______________________________________________ > mpi-forum mailing list > mpi-forum(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpi-forum

1 0

MPI_Get_elements_c
by Raffenetti, Kenneth J. 07 Jul '21

07 Jul '21

Hi, This issue was raised in the MPICH GitHub https://github.com/pmodels/mpich/issues/5413. In MPI-3.1, we have C bindings for: int MPI_Get_elements(const MPI_Status *status, MPI_Datatype datatype, int *count) int MPI_Get_elements_x(const MPI_Status *status, MPI_Datatype datatype, MPI_Count *count) int MPI_Status_set_elements(MPI_Status *status, MPI_Datatype datatype, int count) int MPI_Status_set_elements_x(MPI_Status *status, MPI_Datatype datatype, MPI_Count count) MPI-4.0 added: int MPI_Get_elements_c(const MPI_Status *status, MPI_Datatype datatype, MPI_Count *count) But no binding was added for an _c version of MPI_Status_set_elements. Was this intended? Thanks, Ken

1 0

Next Virtual MPI Forum Meeting: July 21st
by Martin Schulz 07 Jul '21

07 Jul '21

Hi all, Due to vacation time and other constraints, we'll pick up our Virtual MPI Forum Meetings again on July 21st. If, in the meantime or after that day, a WG would like to use the slot either for internal/extended WG meetings or for a general group discussions, please let me know. As for the 21st: I would like to use the meeting to set up a (at least preliminary) goal-line for MPI 4.1 - we had several open items from the 4.0 process (and before), which will be the starting point. If there are any particular items you think we should address as part of a "clean up" or "minor functionality" please either send them to me and I will collect them or add them to the system as a blank ticket with the MPI 4.1 tag. In two weeks we'll consolidate the list and decide how we approach this. 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