- mpiwg-hybridpm - mailman3.cels.anl.gov

Re: [mpiwg-hybridpm] Joint RMA & HACC WG meeting this week
by Joseph Schuchart 25 Mar '26

25 Mar '26

1 0

Joint RMA & HACC WG meeting this week
by Joseph Schuchart 24 Mar '26

24 Mar '26

1 0

Call for topics & Joint Meeting with RMA WG
by Joseph Schuchart 25 Feb '26

25 Feb '26

1 0

Arxiv Pre-print of MPI Stream Triggering Paper
by Patrick Bridges 17 Feb '26

17 Feb '26

As you all know, we’ve been working on an MPI stream triggering API that relies on persistent operations and proposes relatively modest changes to the MPI standard to enable high-performance one-sided data movement using persistent operations. We have completed an open source, high-performance CPU-free implementation of the point-to-point portion of this API for HPE Slingshot 11 systems that leverages the triggering operations includes in OFI and the GPU interfaces to them in the HPE CXI libfabric provider. The arxiv draft paper below describes the API, its implementation on Slingshot 11, and its evaluation up to 8,192 GPUs of ORNL Frontier and 4,06 GPUs of LLNL Tuolumne. The submission of this paper for peer reviewed publication, with some additional planned work no included in this paper, is imminent. This implementation, which is certainly still research-ware, is available as part of MPI Advance, and both our benchmarks, run scripts, and raw data are also linked from the paper. <https://urldefense.us/v3/__https://arxiv.org/abs/2602.15356__;!!G_uCfscf7eW… > [arxiv-logo-fb.png] Co-Design and Evaluation of a CPU-Free MPI GPU Communication Abstraction and Implementation<https://urldefense.us/v3/__https://arxiv.org/abs/2602.15356__;!!G_uCfscf7eW… > arxiv.org<https://urldefense.us/v3/__https://arxiv.org/abs/2602.15356__;!!G_uCfscf7eW… > Comments and suggestions welcome; we would certainly be interested in discussing the potential options and pitfalls for its implementation on other networking systems or its integration into full MPI implementations. Finally, we have scheduled working meetings begnning next week to start drafting potential language for this API for inclusion in the MPI standard. Thanks, Patrick

1 0

Meeting this week: cMPI: Using CXL Memory Sharing for MPI One-Sided and Two-Sided Inter-Node Communications
by Joseph Schuchart 02 Dec '25

02 Dec '25

1 0

Meeting this week
by Joseph Schuchart 28 Oct '25

28 Oct '25

1 0

Re: [mpiwg-hybridpm] Call for topics
by Rohit Zambre 14 Oct '25

14 Oct '25

When I had looked at endpoints, I used this ticket from Jim as a reference for the Endpoints’ proposal — https://urldefense.us/v3/__https://github.com/mpi-forum/mpi-issues/issues/5… I think the endpoints solution seamlessly addresses a variety of use cases in hybrid MPI+X programming models . Regards, Rohit From: mpiwg-hybridpm <mpiwg-hybridpm-bounces(a)lists.mpi-forum.org> on behalf of Joseph Schuchart via mpiwg-hybridpm <mpiwg-hybridpm(a)lists.mpi-forum.org> Date: Monday, October 13, 2025 at 6:58 AM To: Hybrid working group mailing list <mpiwg-hybridpm(a)lists.mpi-forum.org> Cc: Joseph Schuchart <joseph.schuchart(a)stonybrook.edu> Subject: Re: [mpiwg-hybridpm] Call for topics External email: Use caution opening links or attachments This Message Is From an External Sender This message came from outside your organization. Multi-device collectives are something we need to look into. Endpoints seem a good starting point. Was there ever a full proposal for it? A quick search only brought up Jim's 2014 paper [1] and some related papers. I put it on the list of topics on the wiki [2]. We'll go over it this Wednesday. Cheers Joseph [1] https://urldefense.us/v3/__https://dl.acm.org/doi/abs/10.1177/1094342014548… [2] https://urldefense.us/v3/__https://github.com/mpiwg-hybrid/hybrid-issues/wi… On 10/11/25 04:55, Jeff Hammond wrote: > I am curious if people have any appetite to bring back endpoints, not > for the original purpose, but in order to support GPU ranks. This is > one of the features that MPI lacks relative to NCCL that limit its use > in some AI applications. Some AI users do not want to create >1 > process per node just to manage a GPU, because this makes their > pre/post-processing much harder. They have multithreaded code for > that already and don't want to rewrite it or play silly games with > affinity to try to make the application efficient while idling GPU > management processes. > > Jeff > > On Fri, Oct 10, 2025 at 5:34 PM Joseph Schuchart via mpiwg-hybridpm > <mpiwg-hybridpm(a)lists.mpi-forum.org> wrote: >> Hi all, With the September Forum meeting behind us we should pick us the meetings again. Please let me know if you have any topics you want to discuss. We plan to call for joint meeting with the RMA WG to discuss device-side operations but the >> ZjQcmQRYFpfptBannerStart >> This Message Is From an External Sender >> This message came from outside your organization. >> >> ZjQcmQRYFpfptBannerEnd >> >> Hi all, >> >> With the September Forum meeting behind us we should pick us the >> meetings again. Please let me know if you have any topics you want to >> discuss. >> >> We plan to call for joint meeting with the RMA WG to discuss device-side >> operations but the RMA WG first needs to go through the feedback from >> the Forum meeting. Given that people are busy with SC prep, it will >> probably slip into December. >> >> Other topics are of course welcome :) >> >> Thanks >> Joseph >> >> _______________________________________________ >> mpiwg-hybridpm mailing list >> mpiwg-hybridpm(a)lists.mpi-forum.org >> https://urldefense.us/v3/__https://protect.checkpoint.com/v2/r01/___https:/… > >

1 0

Meeting this week
by Joseph Schuchart 13 Oct '25

13 Oct '25

1 0

Re: [mpiwg-hybridpm] Call for topics
by Joseph Schuchart 13 Oct '25

13 Oct '25

1 0

Re: [mpiwg-hybridpm] Call for topics
by Jeff Hammond 11 Oct '25

11 Oct '25

1 0