SC17 Denver, CO

Open MPI State of the Union XI


Authors: Jeffrey Squyres (Cisco Systems)

BP
Abstract: Open MPI continues to drive the state of the art in HPC. This year, we've added new features, fixed bugs, improved performance, and collaborated with many across the HPC community. We'll discuss what Open MPI has accomplished over the past year and present a roadmap for the next year.

One of Open MPI's strength lies in its diversity: we represent many different viewpoints across the HPC ecosystem. To that end, many developers from the community will be present to discuss and answer your questions both during and after the BoF.

Long Description: The Open MPI community has been hard at work on new and useful features, tweaking performance improvements, and overall improving the structure and clarity of the code. Our emphasis over the last year has been scalability of MPI jobs in the context of full MPI-3.1 conformance (as we creep closer to exascale), full, fine-grained MPI_THREAD_MULTIPLE support, and other miscellaneous stability and performance improvements.

Over the past few years, we have found that the BOF format makes it difficult to solicit questions and feedback from the community who attends our BOF. However, we consider such questions and feedback vital to the continuation of Open MPI -- our users and community are a huge portion of what helps Open MPI keep evolving in a direction that is genuinely useful. This year, we are soliciting questions from the web before the BOF; send us your questions, comments, and feedback and we'll discuss them during the BOF:

http://www.open-mpi.org/sc17

In this BOF, we'll present where Open MPI currently is and where it is going. Here are some of the highlights of what will be covered during the discussions:

- Continue the 2.x and 3.x release series - Performance improvements - Scalability improvements - Multithreading and asynchronous progress support - Stability - Improvements on one-sided operations - Supporting more hardware platforms and networks - Support for new collective algorithms, including direct hardware support - Reliability

Be part of the discussion: submit your questions ahead of time and come hear where Open MPI is going, and how you can (and should!) join our efforts.

Conference Presentation: pdf


Birds of a Feather Index