- mpiwg-sessions - mailman3.cels.anl.gov

Re: [mpiwg-sessions] MPI_Buffer_attach in the sessions model
by Dan Holmes 17 Sep '22

17 Sep '22

Hi Rolf & Tony, With current text, MPI-4.0, all sessions and world model buffered send operations would use the same attached buffer. This is likely to be implemented correctly but it breaches the intent of separation of resources used by different sessions, so is not ideal. A new buffer attach API could be designed to be more in keeping with the intent of sessions. A per-session buffer and a per-communicator buffer are both options to consider. There’s no GitHub issue or PR for this new API yet. Addressing this should fall under the remit of the sessions WG and will probably target MPI-5.0 at the earliest. Thanks for raising this point. Cheers, Dan. Sent from my iPhone > On 17 Sep 2022, at 14:18, Tony Skjellum via mpiwg-sessions <mpiwg-sessions(a)lists.mpi-forum.org> wrote: > > Rolf, you’ve discovered a problem with the spec… it’s global to attach a buffer, so implicitly associated with ‘world model’ and the communicators derived from MPI_Comm_world … > > Each session is to be isolated resource wise … so, we have to make a new API when there is no world model . > > The API would take a session key to align buffer attachment to a session in my opinion. > > Use of the current api with a session appears erroneous. :-) > > I am sure it would work in real implementations nonetheless. > > Tony > > Anthony Skjellum, PhD > 205-807-4968 > > > On Sep 17, 2022, at 8:42 AM, Rolf Rabenseifner via mpiwg-sessions <mpiwg-sessions(a)lists.mpi-forum.org> wrote: > > Dear sessions-group, > > is there any information about the usage of MPI_Buffer_attach, > MPI_Buffer_detach and MPI_Bsend for applications that use the > world model together with (several) sessions. > > MPI-4.0, the function index entries of MPI_Buffer_attach and > the corresponding texts do not give an answer, or do they? > > Is it allowed or intended that two MPI_Bsend calls with two different > communicators derived not from the same session or world > may use portions of the same buffer that was attached with MPI_Buffer_attach? > > Is there any issue or pull request already for this problem prepared? > Or slides in previous meetings? > > 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 > _______________________________________________ > mpiwg-sessions mailing list > mpiwg-sessions(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions > _______________________________________________ > mpiwg-sessions mailing list > mpiwg-sessions(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions

1 0

Re: [mpiwg-sessions] MPI_Buffer_attach in the sessions model
by Tony Skjellum 17 Sep '22

17 Sep '22

Rolf, you’ve discovered a problem with the spec… it’s global to attach a buffer, so implicitly associated with ‘world model’ and the communicators derived from MPI_Comm_world … Each session is to be isolated resource wise … so, we have to make a new API when there is no world model . The API would take a session key to align buffer attachment to a session in my opinion. Use of the current api with a session appears erroneous. :-) I am sure it would work in real implementations nonetheless. Tony Anthony Skjellum, PhD 205-807-4968 On Sep 17, 2022, at 8:42 AM, Rolf Rabenseifner via mpiwg-sessions <mpiwg-sessions(a)lists.mpi-forum.org> wrote: Dear sessions-group, is there any information about the usage of MPI_Buffer_attach, MPI_Buffer_detach and MPI_Bsend for applications that use the world model together with (several) sessions. MPI-4.0, the function index entries of MPI_Buffer_attach and the corresponding texts do not give an answer, or do they? Is it allowed or intended that two MPI_Bsend calls with two different communicators derived not from the same session or world may use portions of the same buffer that was attached with MPI_Buffer_attach? Is there any issue or pull request already for this problem prepared? Or slides in previous meetings? 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 _______________________________________________ mpiwg-sessions mailing list mpiwg-sessions(a)lists.mpi-forum.org https://lists.mpi-forum.org/mailman/listinfo/mpiwg-sessions

1 0

MPI_Buffer_attach in the sessions model
by Rolf Rabenseifner 17 Sep '22

17 Sep '22

Dear sessions-group, is there any information about the usage of MPI_Buffer_attach, MPI_Buffer_detach and MPI_Bsend for applications that use the world model together with (several) sessions. MPI-4.0, the function index entries of MPI_Buffer_attach and the corresponding texts do not give an answer, or do they? Is it allowed or intended that two MPI_Bsend calls with two different communicators derived not from the same session or world may use portions of the same buffer that was attached with MPI_Buffer_attach? Is there any issue or pull request already for this problem prepared? Or slides in previous meetings? 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

PMIx allocate resources request
by Ralph Castain 14 Sep '22

14 Sep '22

Hi folks I was in a meeting earlier today where it emerged that you folks might be having some discussions related to the PMIx_Allocate_resources API? It sounded like there might be some confusion over that operation and how it works, plus there have been some developments regarding it in other forums. If you let me know when you might be discussing this topic, I'd be happy to drop in to answer questions and provide updates on it. Ralph

1 0

Sessions/FT WG meet today
by Pritchard Jr., Howard 12 Sep '22

12 Sep '22

Hello All, Today I would like to do a sanity check on various sessions/ft related issues/PRs we need to take action on for the upcoming Forum. One’s I have on my list are https://github.com/mpi-forum/mpi-issues/issues/535 - Language bindings chapter needs updating for Sessions model https://github.com/mpi-forum/mpi-issues/issues/545 - MPI_LASTUSEDCODE not covered in the sessions model https://github.com/mpi-forum/mpi-issues/issues/534 - Clarify/fix description and usage of MPI_GROUP_NULL https://github.com/mpi-forum/mpi-issues/issues/511 - expansion of error handler association to cover sessions (and clarifications) The PRs associated with these issues all had readings at the last forum. Thanks, Howard — [signature_61897647] Howard Pritchard Research Scientist HPC-ENV Los Alamos National Laboratory howardp(a)lanl.gov [signature_116643141]<https://www.instagram.com/losalamosnatlab/>[signature_2106573303]<https://twitter.com/LosAlamosNatLab>[signature_1572415991]<https://www.linkedin.com/company/los-alamos-national-laboratory/>[signature_1497523290]<https://www.facebook.com/LosAlamosNationalLab/>

1 0

Sessions WG meeting 9/5/22
by Pritchard Jr., Howard 02 Sep '22

02 Sep '22

Hi All, This coming Monday is a holiday in the US, so attendance from the left side of the Atlantic may be low. I will not be attending. Howard — [signature_61897647] Howard Pritchard Research Scientist HPC-ENV Los Alamos National Laboratory howardp(a)lanl.gov [signature_3808957771]<https://www.instagram.com/losalamosnatlab/>[signature_1636348083]<https://twitter.com/LosAlamosNatLab>[signature_1217036039]<https://www.linkedin.com/company/los-alamos-national-laboratory/>[signature_979643112]<https://www.facebook.com/LosAlamosNationalLab/>

1 0

Re: [mpiwg-sessions] Joint Sessions/FT WG zoom - meet today
by Dan Holmes 29 Aug '22

29 Aug '22

Hi all, I’m going to miss today’s call—covid has finally managed to find me and I’m not going to be any use to anyone for a few days. Cheers, Dan. Sent from my iPhone > On 29 Aug 2022, at 15:20, Pritchard Jr., Howard via mpiwg-sessions <mpiwg-sessions(a)lists.mpi-forum.org> wrote: > > > HI All, > > Picking up from where we left off four weeks ago (since two weeks ago we didn’t have a “quorum”), the agenda is currently: > follow up on discussion of Dominik’s proposal > version 1 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…) > version 2 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…) > local (or more so) alternatives to MPI_Session_finalize - https://github.com/mpiwg-sessions/sessions-issues/issues/23 > > Howard > > > > > — > > > Howard Pritchard > Research Scientist > HPC-ENV > > Los Alamos National Laboratory > howardp(a)lanl.gov > > > > > _______________________________________________ > mpiwg-ft mailing list > mpiwg-ft(a)lists.mpi-forum.org > https://lists.mpi-forum.org/mailman/listinfo/mpiwg-ft

1 0

Joint Sessions/FT WG zoom - meet today
by Pritchard Jr., Howard 29 Aug '22

29 Aug '22

HI All, Picking up from where we left off four weeks ago (since two weeks ago we didn’t have a “quorum”), the agenda is currently: * follow up on discussion of Dominik’s proposal * version 1 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…<https://urldefense.com/v3/__https:/raw.githubusercontent.com/wiki/mpiwg-ses…>) * version 2 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…<https://urldefense.com/v3/__https:/raw.githubusercontent.com/wiki/mpiwg-ses…>) * local (or more so) alternatives to MPI_Session_finalize - https://github.com/mpiwg-sessions/sessions-issues/issues/23<https://urldefense.com/v3/__https:/github.com/mpiwg-sessions/sessions-issue…> Howard — [signature_61897647] Howard Pritchard Research Scientist HPC-ENV Los Alamos National Laboratory howardp(a)lanl.gov [signature_1415175627]<https://www.instagram.com/losalamosnatlab/>[signature_3364980106]<https://twitter.com/LosAlamosNatLab>[signature_3909117046]<https://www.linkedin.com/company/los-alamos-national-laboratory/>[signature_2453085042]<https://www.facebook.com/LosAlamosNationalLab/>

1 0

Cancel today's Sessions WG meeting
by Pritchard Jr., Howard 22 Aug '22

22 Aug '22

Hi All, Per discussion at last week’s joint meeting with FT group, we’ll cancel today’s Sessions WG meeting owing to numerous folks being on vacation. We’ll meet again next week to pick up with the agenda we didn’t cover last Monday: * follow up on discussion of Dominik’s proposal * version 1 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…<https://urldefense.com/v3/__https:/raw.githubusercontent.com/wiki/mpiwg-ses…>) * version 2 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…<https://urldefense.com/v3/__https:/raw.githubusercontent.com/wiki/mpiwg-ses…>) * local (or more so) alternatives to MPI_Session_finalize - https://github.com/mpiwg-sessions/sessions-issues/issues/23<https://urldefense.com/v3/__https:/github.com/mpiwg-sessions/sessions-issue…> Thanks, Howard — [signature_61897647] Howard Pritchard Research Scientist HPC-ENV Los Alamos National Laboratory howardp(a)lanl.gov [signature_1109091345]<https://www.instagram.com/losalamosnatlab/>[signature_749480697]<https://twitter.com/LosAlamosNatLab>[signature_855853447]<https://www.linkedin.com/company/los-alamos-national-laboratory/>[signature_3022872670]<https://www.facebook.com/LosAlamosNationalLab/>

1 0

Session/FT WGs meet today
by Pritchard Jr., Howard 15 Aug '22

15 Aug '22

Hi All, Let’s meet today. Current agenda items: * follow up on discussion of Dominik’s proposal * version 1 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…) * version 2 (https://raw.githubusercontent.com/wiki/mpiwg-sessions/sessions-issues/dynam…) * local (or more so) alternatives to MPI_Session_finalize - https://github.com/mpiwg-sessions/sessions-issues/issues/23 We can add other items to the agenda if time permits. Howard — [signature_61897647] Howard Pritchard Research Scientist HPC-ENV Los Alamos National Laboratory howardp(a)lanl.gov [signature_2626573852]<https://www.instagram.com/losalamosnatlab/>[signature_232057090]<https://twitter.com/LosAlamosNatLab>[signature_3339681486]<https://www.linkedin.com/company/los-alamos-national-laboratory/>[signature_4187432442]<https://www.facebook.com/LosAlamosNationalLab/>

1 0