March 2025 - mpiwg-coll - mailman3.cels.anl.gov

Re: [mpiwg-coll] Discussing Semantics and Synchronization Requirements for Neighbourhood Collectives
by Christoph Niethammer 27 Mar '25

27 Mar '25

1 0

Re: [mpiwg-coll] [mpiwg-hw-topology] Discussing Semantics and Synchronization Requirements for Neighbourhood Collectives
by Anthony Skjellum 18 Mar '25

18 Mar '25

FYI to Christoph and Julien, I think that alltoallv/w operations with zero counts are deserving special consideration too … why would we require a barrier ? ; allowing a barrier is acceptable … with neighbors, there is little conceptual difference from zero counts for some sources and sinks, so we must be consistent. The term synchronizing collective means — must Or — may ?? That is confusing. Clearly barrier must… what about persistent allreduce implemented via 1-sided inside … is there for sure a barrier effect ? I am worried about this detail to define partitioned collectives too. Tony Anthony Skjellum, PhD skjellum(a)gmail.com Cell: +1-205-807-4968 On Tue, Mar 18, 2025 at 9:15 AM Christoph Niethammer via mpiwg-hw-topology < mpiwg-hw-topology(a)lists.mpi-forum.org> wrote: > Hello everyone, During the preparation for the FRM, a question was raised > regarding whether and how the semantics of collectives apply to > neighborhood collectives. [1] This question stemmed from remark 18) in the > summary table in Appendix A. 2, > ZjQcmQRYFpfptBannerStart > This Message Is From an External Sender > This message came from outside your organization. > > ZjQcmQRYFpfptBannerEnd > > Hello everyone, > > During the preparation for the FRM, a question was raised regarding > whether and how the semantics of collectives apply to neighborhood collectives. [1] > This question stemmed from remark 18) in the summary table in Appendix A.2, > which was used to apply the semantics of collectives to neighbourhood > collectives at this place. > > The discussions expanded during the March meeting to the question of whether remark 18, > stating that "all-to-all collectives *must* synchronize (with some restrictions)", > is correct. [2] > > I would like to invite you all to discuss these two issues in more detail during a > Topologies WG meeting on March 27th at 15:00, as part of our regular WG slot. > > Best regards > Christoph Niethammer > > [1] https://urldefense.us/v3/__https://github.com/mpi-forum/mpi-issues/issues/8… > [2] https://urldefense.us/v3/__https://github.com/mpi-forum/mpi-issues/issues/9… > > -- > > Dr.-Ing. Christoph Niethammer > High Performance Computing Center Stuttgart (HLRS)Nobelstrasse 19 > 70569 Stuttgart <https://urldefense.us/v3/__https://www.google.com/maps/search/Nobelstrasse*… > > > Tel: ++49(0)711-685-87203 > email: [email protected]://urldefense.us/v3/__https://www.hlrs.de/people/christoph-niethammer__;!!G_uCfscf7eWS!edo6G3QE7JbjtZTTusbpg5yIgnpGmlkZf6PJXEhK0BIFo_p25xzgJ-iVjVZhEuL7z3exLA62R0MjXhpEVDjC5QAigTtN4tiyVz0SIxo$ > > _______________________________________________ > mpiwg-hw-topology mailing list > mpiwg-hw-topology(a)lists.mpi-forum.org > https://urldefense.us/v3/__https://lists.mpi-forum.org/mailman/listinfo/mpi… >

1 0

Re: [mpiwg-coll] [mpiwg-hw-topology] Discussing Semantics and Synchronization Requirements for Neighbourhood Collectives
by Julien.JAEGER＠CEA.FR 18 Mar '25

18 Mar '25

Hello Christoph, it is indeed a good question. I will also take a look at the table w.r.t. inter-communicators. It may have similar caveats for collectives with inter-communicators. Best regards, Julien J. ________________________________ De : mpiwg-hw-topology <mpiwg-hw-topology-bounces(a)lists.mpi-forum.org> de la part de Christoph Niethammer via mpiwg-hw-topology <mpiwg-hw-topology(a)lists.mpi-forum.org> Envoyé : mardi 18 mars 2025 09:15:23 À : mpiwg-hw-topology; mpiwg-semantic-terms(a)lists.mpi-forum.org; MPI-3 Collective Subgroup Discussions Cc : Martin Schulz Objet : [mpiwg-hw-topology] Discussing Semantics and Synchronization Requirements for Neighbourhood Collectives Hello everyone, During the preparation for the FRM, a question was raised regarding whether and how the semantics of collectives apply to neighborhood collectives. [1] This question stemmed from remark 18) in the summary table in Appendix A. 2, ZjQcmQRYFpfptBannerStart This Message Is From an External Sender This message came from outside your organization. ZjQcmQRYFpfptBannerEnd Hello everyone, During the preparation for the FRM, a question was raised regarding whether and how the semantics of collectives apply to neighborhood collectives. [1] This question stemmed from remark 18) in the summary table in Appendix A.2, which was used to apply the semantics of collectives to neighbourhood collectives at this place. The discussions expanded during the March meeting to the question of whether remark 18, stating that "all-to-all collectives *must* synchronize (with some restrictions)", is correct. [2] I would like to invite you all to discuss these two issues in more detail during a Topologies WG meeting on March 27th at 15:00, as part of our regular WG slot. Best regards Christoph Niethammer [1] https://urldefense.us/v3/__https://github.com/mpi-forum/mpi-issues/issues/8… [2] https://urldefense.us/v3/__https://github.com/mpi-forum/mpi-issues/issues/9… -- Dr.-Ing. Christoph Niethammer High Performance Computing Center Stuttgart (HLRS) Nobelstrasse 19 70569 Stuttgart Tel: ++49(0)711-685-87203 email: christoph.niethammer(a)hlrs.de https://urldefense.us/v3/__https://www.hlrs.de/people/christoph-niethammer_…

1 0

Re: [mpiwg-coll] Discussing Semantics and Synchronization Requirements for Neighbourhood Collectives
by Michael Knobloch 18 Mar '25

18 Mar '25

1 0

Discussing Semantics and Synchronization Requirements for Neighbourhood Collectives
by Christoph Niethammer 18 Mar '25

18 Mar '25

1 0