ECP @ ISC High Performance 2022
Hamburg, Germany | May 29–June 2
When devotees of high-performance computing, machine learning, and data analytics meet in Hamburg, Germany, May 29–June 2 for the ISC High Performance 2022 conference, the US Department of Energy’s Exascale Computing Project (ECP) will contribute at numerous sessions covering a variety of topics and provide chair and committee leadership.
Of special note with respect to conference activities, Johannes Doerfert of Argonne National Laboratory and Atmn Patel of Argonne and the University of Waterloo (Canada) won the ISC22 best research paper for “Remote OpenMP Offloading,” which explains ECP-funded work. The Research Papers committee selected the entry for the award over 18 accepted papers. The authors of the award-winning paper will present a keynote talk on their team’s activities in an exclusive session on Monday, May 30, at 4:00 p.m. The team will receive a cash prize of 3,000 Euros, along with an award certificate.
The knowledge, skills, and talents of ECP researchers were brought to bear in many ISC chair and committee roles.
Hartwig Anzt (University of Tennessee) chairs Workshops; Kathryn Mohror (Lawrence Livermore National Laboratory), Tutorials; and Piotr Luszczek (University of Tennessee), Proceedings. Committees set up the contributed program sessions, based on the call procedures.
The members of the ISC Steering committee are distinguished experts and decision makers who provide guidance on the conference program and help enhance publicity, acceptance, and growth of ISC in industry, research, and academia. ECP’s Jack Dongarra (University of Tennessee and Oak Ridge National Laboratory), Anzt, Suren Byna (Lawrence Berkeley National Laboratory) (Tutorials deputy chair), Luszczek, and Mohror are members of the Steering committee.
ECP researchers are also on the Research Papers committee. Under the Architecture, Networks, and Storage portion of the committee is Elsa Gonsiorowski (Lawrence Livermore National Laboratory); under HPC Algorithms & Applications is Johann Rudi (Argonne National Laboratory); and under Performance Modeling, Evaluation, and Analysis is Scott Pakin (Los Alamos National Laboratory).
Additionally, Anzt is on the Research Posters committee.
Mohror and Byna are on the Tutorials committee, as is ECP’s Philip Carns (Argonne National Laboratory). On the Workshops committee from ECP are Anzt, Natalie Beams (University of Tennessee), George Bosilca (University of Tennessee), Lisa Claus (Lawrence Berkeley National Laboratory), and Anthony Danalis (University of Tennessee).
| Event | Summary | Speakers/Presenters |
|---|---|---|
| Sunday, May 29, 9:00 a.m.–1:00 p.m. | ||
| Tutorial: Managing HPC Software Complexity with Spack | This tutorial provides a thorough introduction to Spack, an open source tool for HPC package management that simplifies building, installing, customizing, and sharing HPC software stacks. | Todd Gamblin, Computer Scientist, Lawrence Livermore National Laboratory; Gregory Becker, Computer Scientist, Lawrence Livermore National Laboratory; Tamara Dahlgren, Computer Scientist, Lawrence Livermore National Laboratory; Massimiliano Culpo, Computer Scientist, n.p. complete, S.r.l.; Michael Kuhn, Junior Professor, Otto von Guericke University Magdeburg; Harmen Stoppels, Computer Scientist, Swiss National Supercomputing Centre; Adam Stewart, Ph.D. Student, University of Illinois |
| Tutorial: “Modern Mixed-Precision Methods” | This tutorial will expose the audience to the rapidly expanding landscape of mixed- and multi-precision methods. | Piotr Luszczek, Research Assistant Professor, University of Tennessee; Jack Dongarra, Distinguished Professor, University of Tennessee; Hartwig Anzt, Helmholtz-Young-Investigator Group leader, Karlsruhe Institute of Technology |
| Tutorial: “Practical Hybrid Parallel Application Performance Engineering” | This tutorial presents state-of-the-art performance tools for leading-edge HPC systems founded on the community-developed Score-P instrumentation and measurement infrastructure, demonstrating how they can be used for performance engineering of effective scientific applications based on standard MPI, OpenMP, hybrid combination of both, and increasingly common usage of accelerators. | Sameer Shende, Director University of Oregon; Markus Geimer, Research Scientist, Jülich Supercomputing Centre; Bill Williams, Research Scientist, Technische Universität Dresden, Technische Universität Dresden, ZIH |
| Tutorial: “Developing Robust and Scalable Next Generation Workflows Applications and Systems” | This tutorial will present the ExaWorks SDK, and its constituent components: Flux, Parsl, RADICAL-Cybertools (RCT), and Swift/T. These components are widely used, and available tools for developing workflow applications. | Daniel Laney, Computational Scientist, Lawrence Livermore National Laboratory; Justin Wozniak, Computer Scientist, Argonne National Laboratory; Andre Merzky; for more information, contact Daniel Laney ([email protected]) |
| Sunday, May 29, 2:00 p.m.–6:00 p.m. | ||
| Tutorial: “Using Containers to Accelerate HPC” | This hands-on tutorial aims to train users on the use of containers on HPC resources. It provides a background on Linux containers, along with introductory hands-on experience building a container image, sharing the container and running it on a HPC cluster. Furthermore, the tutorial will provide more advanced information on how to run MPI-based and GPU-enabled HPC applications, how to optimize I/O intensive workflows, and other best practices. | Sameer Shende, Director, University of Oregon; Shane Canon, Project Engineer, Lawrence Berkeley National Lab; Eduardo Arango Gutierrez, Software Engineer, Redhat; Alexis Espinosa, Supercomputing Applications Specialist, Pawsey Supercomputing Center |
| Tutorial: “Mastering Tasking with OpenMP” | Mastering the tasking concept of OpenMP requires a change in the way developers reason about the structure of their code and how to expose the parallelism of it. This tutorial addresses this critical aspect by examining the tasking concept in detail and presenting patterns as solutions to many common problems. | Christian Terboven, HPC group lead, RWTH Aachen University; Michael Klemm, Principal Member of Technical Staff, OpenMP ARB; Bronis R. de Supinski, Chief Technology Officer for Livermore Computing, Lawrence Livermore National Laboratory; Xavier Teruel, Researcher, Barcelona Supercomputing Center |
| Tutorial: “Advanced MPI Programming” | This is an advanced-level tutorial that will provide an overview of various powerful features in MPI, especially with MPI-2 and MPI-3, and will present a brief introduction on MPI-4 new features. | Pavan Balaji, Applied Research Scientist, Facebook AI; Torsten Hoefler, Associate Professor, ETH Zurich; Antonio Peña, Sr. Researcher and Team Lead, Barcelona Supercomputing Center; Yanfei Guo, Assistant Computer Scientist, Argonne National Laboratory |
| Sunday, May 29, 2:00 p.m.–6:00 p.m. | ||
| Tutorial: "Better Scientific Software" | This tutorial will provide information and hands-on experience with software practices, processes, and tools explicitly tailored for computational science and engineering (CSE). Goals are improving the productivity of those who develop CSE software and increasing the sustainability of software artifacts. The presenters will discuss practices that are relevant for projects of all sizes, with emphasis on small teams, and on aggregate teams composed of small teams. | Anshu Dubey, Computational Scientist, Argonne National Laboratory; Rinku Gupta, Research Software Specialist, Argonne National Laboratory; Greg Watson, Group Leader, Application Engineering, Oak Ridge National Laboratory; David Rogers, Computational Scientist, Oak Ridge National Laboratory |
| Monday, May 30, 4:00 p.m.–5:15 p.m. | ||
| Focus Session: System Architecture of "ExaFLOPS" Supercomputers | Sato Mitsuhisa, Deputy Director of R-CCS, RIKEN; Doug Kothe, Oak Ridge National Laboratory; | |
| Monday, May 30, 4:20 p.m.–4:40 p.m. | ||
| Focus Session: System Architecture of "ExaFLOPS" Supercomputers; Presentation: TBD | Mike Heroux, Senior Research Scientist, Sandia National Laboratories | |
| Monday, May 30, 5:00 p.m.–5:15 p.m. | ||
| Focus Session: System Architecture of "ExaFLOPS" Supercomputers; Live Q&A | Mike Heroux, Senior Research Scientist, Sandia National Laboratories; Martin Schulz, Professor, Technical University of Munich; Doug Kothe, Oak Ridge National Laboratory | |
| Monday, May 30, 4:00 p.m.–5:00 p.m. | ||
| BOF Session: Accelerating HPC and AI with DAOS Storage | This BoF will be the opportunity for members of the DAOS community to share experience running and using DAOS and also brainstorm on possible enhancements (e.g., new OS support, storage management, GPU acceleration, AI framework integration) that should be considered for future DAOS versions. The current roadmap will be presented to the audience and short lightning talks from various community members will be used to spark the discussion on a broad set of topics. | Johann Lombardi, Senior Principal Engineer, Intel Corporation; Kevin Harms, Performance Engineering Team Lead, ALCF; Michael Hennecke, Principal Engineer, Intel (Deutschland) GmbH |
This year’s Hans Meuer Award winning title is: Remote OpenMP Offloading, which is on the topic: Programming Environments & Systems Software. The Research Papers Committee selected this paper over 18 accepted papers to receive this award. ECP funded the work explained in the paper. | Paper co-authors Johannes Doerfert (Argonne National Lab, USA) and Atmn Patel (Argonne National Lab, USA/University of Waterloo, Canada) will present a keynote talk on their team’s work in an exclusive session on Monday, May 30 at 4 pm. The team will also receive a cash prize of 3,000 Euros, along with an award certificate. | |
| Monday, May 30, 5:15 p.m.–6:30 p.m. | ||
| Focus Session: "Advances in OpenMP" | Barbara Chapman, Professor of Applied Mathematics and Statistics, and of Computer Science, HPE and Stony Brook University | |
| Focus Session: "Advances in OpenMP"; Presentation: "The LLVM/OpenMP Ecosystem—Optimizations, Features, and Outlook" | This talk looks at the revolution in the LLVM/OpenMP development and discusses the significance beyond OpenMP. It introduces the non-conventional features already available through LLVM/OpenMP but also provides an outlook for what is to come. The goal of the talk is to introduce compilers, parallelism, and OpenMP in particular, as approachable technologies with often underestimated potential. | Johannes Doerfert, Assistant Computer Scientist, Argonne National Lab |
| Monday, May 30, 6:15 p.m.–6:30 p.m. | ||
| Focus Session: "Advances in OpenMP"; "Live Q&A: Advances in OpenMP" | Johannes Doerfert, Assistant Computer Scientist, Argonne National Lab; Barbara Chapman, Professor of Applied Mathematics and Statistics, and of Computer Science, HPE and Stony Brook University; Joachim Protze, Research Staff, RWTH Aachen University, ITC; Jeffrey Sandoval, OpenMP Compiler Tech Lead Hewlett Packard Enterprise | |
| Tuesday, May 31, 9:00 a.m.–10:00 a.m. | ||
| BOF Session: "Unified Communication X (UCX) Community | UCX is a collaboration between industry, national labs and academia that consolidates multiple technologies that provides a unified open source framework.UCX has already been integrated with upstream of Open MPI project and OpenSHMEM, being used with MPICH and more. UCX is now being deployed in many supercomputers worldwide and US DOE systems. The session will enable a dialog on the future plans for UCX and review the operations of the UCX consortium. It will include performance evaluation based on testing results conducted on several supercomputing platforms. | Dhabaleswar Panda, Professor and Distinguished Scholar of Computer Science, The Ohio State University; Gilad Shainer, SVP Networking, NVIDIA; Yanfei Guo, Assistant Computer Scientist, Argonne National Laboratory; Steve Poole, Chief Architect, Los Alamos National Laboratory; Jeff Kuehn, Program Manager, Los Alamos National Laboratory; Pavel Shamis, Principal Research Engineer, Arm; Perry Schmidt, Software Development Manager, IBM |
| BOF Session: "Spack Community BOF" | Spack is a package manager for scientific computing, with a rapidly growing open source community. At this BOF, Spack core developers will give updates on the state of the Spack community, and on new features such as Spack container generation, Spack's new concretization algorithm, new compiler dependency models, and gitlab continuous integration pipelines. They will also conduct an interactive survey with the audience and open the floor for guided discussion as we review the results. All are invited to provide feedback, request features, and participate in the discussion! Users will be encouraged to present specific use-cases they think may be relevant to the larger community. | Todd Gamblin, Computer Scientist, Lawrence Livermore National Laboratory; Gregory Becker, Computer Scientist, Lawrence Livermore National Laboratory; Tamara Dahlgren, Computer Scientist, Lawrence Livermore National Laboratory; Massimiliano Culpo, Computer Scientist, n.p. complete, S.r.l.; Michael Kuhn, Junior Professor, Otto von Guericke University Magdeburg; Harmen Stoppels, Computer Scientist, Swiss National Supercomputing Centre; Adam Stewart, Ph.D. Student, University of Illinois |
| Tuesday, May 31, 9:15 a.m.–10:15 a.m. | ||
| Focus Sessions: "Mixed Precision in Low-Rank Approximation and Randomization"; Presentation: "Mixed Precision Randomised Nyström Method" | This work considers a single pass version of a randomised Nyström method for approximating a symmetric positive semidefinite matrix A. This method accesses A only once to compute a matrix-matrix product. The presenters perform this product in lower than the working precision and examine the finite precision error. The Nyström approximation is then used to construct a limited memory preconditioner for systems with coefficient matrices with spectral decay, and the presenters solve these via conjugate gradient method. | Ieva Dauzickaite, Charles University; for more information, contact Erin Carson ([email protected]) |
| Wednesday, June 1, 9:00 a.m.–10:00 a.m. | ||
| BOF Session: "LLVM in HPC—Where We are Where We Need to be" | In this BoF, participants from academia and science discuss the current status of LLVM in HPC and how it can best be improved to benefit a broader community. The presenters plan to have LLVM developers additionally highlight existing and emerging features of interest for scientific software and HPC applications. Anyone experienced or simply interested in the LLVM compiler framework or any of its subprojects is invited to participate in the discussion as a stepping stone to get involved. | Johannes Doerfert, Researcher, Argonne National Laboratory; Sameer Shende, Director of the Performance Research Lab, University of Oregon; Anja Gerbes, Researcher, ZIH, Technische Universität Dresden; Jeremy Bennett, Founder and Chief Executive, Embecosm |
| Wednesday, June 1, 2:30 p.m.–3:30 p.m. | ||
| BOF Session: "Software Engineering and Reuse in Modeling, Simulation, and Data Analytics for Science and Engineering" | This BOF provides an opportunity for people concerned about the growing challenges in software engineering for computational science and the pace of change and level of diversity in platform architectures to share existing activities, discuss how they can be expanded and improved, and share the results. Presentations and discussion notes will be made available to the community at the BoF series website, http://bit.ly/swe-cse-bof. | Anshu Dubey, Computational Scientist, Argonne National Laboratory; Weronika Filinger, HPC Application Consultant, EPCC; Mozhgan Kabiri chimeh, GPU Developer Advocate, NVIDIA; Jonathan Dursi, Staff Scientist, University Health Network; Marion Weinzierl, Research Software Engineer, Durham University |
| Wednesday, June 1, 4:00 p.m.–5:15 p.m. | ||
| Focus Session: "HPC in the Cloud, Accessibility, and Performance" | Sameer Shende, Director, University of Oregon Dhabaleswar Panda, Professor and Distinguished Scholar of Computer Science, The Ohio State University; Sameer Shende, Director, University of Oregon; Ewa Deelman, University of Southern California; Ilkay Altintas, UC San Diego | |
| Focus Session: "Extreme-Scale Parallelism and Mathematical Libraries" | Hartwig Anzt (ICL, UTK); Hatem Ltaief (KAUST), Nils Kohl (Friedrich-Alexander-University Erlangen-Nuermberg), Katazyna Swirydowicz (Pacific Northwest National Laboratory); for more information contact Hartwig Anzt ([email protected]) | |
| Thursday, June 2, 9:00 a.m.–1:00 p.m. | ||
| Workshop: "The Third Workshop on LLVM Compiler and Tools for HPC" | The LLVM framework is a vast ecosystem surrounding a compiler core that enabled various advances in source-code tools, debuggers, linkers, and a host of programming-language and toolchain-related components. In addition to the open-source components in its framework, LLVM also serves as a foundation for a majority of vendor compilers and toolchains. Just as its already established cousin LLVM-HPC @ Super Computing (SC), this workshop will bring together the different groups interested in LLVM: compiler and tooling developers, researchers, and application specialists. The workshop features compiler and tool developers that present new features to make the community aware, researchers who discuss new ideas to gather feedback and build collaborations, and application specialists with experience reports and requests. | Tim Cramer, Postdoctoral Researcher, RWTH Aachen University; Roger Ferrer Ibanez, Research Leader of the Compilers and Toolchains for HPC group at BSC Barcelona Supercomputing Center; William Moses, Researcher, Massachusetts Institute of Technology; Hal Finkel, Program Manager, US Department of Energy; Ludger Paehler, Research Associate, Technical University of Munich; Michal Paszkowski, Graphics Compiler Engineer, Intel; Anastasia Stulova, Compiler Engineer, ARM; Konrad Trifunovic, Compiler Optimization Engineer, Intel; Alex Zinenko, Senior Research Engineer, Google; Lewis Revill, Tool Chain Engineer, Embecosm |
| Focus Session: "Extreme-Scale Parallelism and Mathematical Libraries"; Presentation: "Achieving Portable Performance on Extreme-Scale Systems with the PETSc Numerical Toolkit" | Richard Tran Mills will present some of the algorithmic choices that have been made in the PETSc numerical toolkit over the years to target extreme-scale systems as they have evolved, then focus on the strategies that are being employed to prepare PETSc to efficiently utilize the upcoming exascale generation of supercomputers. | Richard Tran Mills, Argonne National Laboratory; for questions about the session, contact Hartwig Anzt ([email protected]); for questions about the PETsc talk, contact Richard Tran Mills ([email protected]) |
| Thursday, June 2, 2:00 p.m.–6:00 p.m. | ||
| Workshop: "HPC on Heterogeneous Hardware (H3)" http://www.icl.utk.edu/~luszczek/conf/2022/h3/ | The HPC on Heterogeneous Hardware (H3) Workshop is intended to complement the broader scope of the program of ISC High Performance 2022 in Hamburg, Germany. It does so by providing a platform for pioneering work on algorithmic research, software library design, programming models, and workflow development for increasingly heterogeneous hardware. In the workshop context, such hardware spans from ARM processors featuring long-vector extensions through GPU-accelerated systems to architectures deploying special function units, FPGAs, or deep learning processors. The workshop will compose of a well-balanced mix of invited talks, peer-reviewed conference contributions, and a panel bringing together worldwide experts in heterogeneous computing. | Bilel Hadri (KAUST), Hatem Ltaief (KAUST), Piotr Luszczek (University of Tennessee), Hartwig Anzt (Karlsruhe Institute of Technology); for more information, contact Hartwig Anzt ([email protected]) |
| Workshop: "Portable Heterogeneous Programming with SYCL" | In keeping with the broader ISC theme, the workshop will focus on the experience of programming next-generation heterogeneous systems. It aims to be a platform for HPC code developers, users, and providers of systems, languages, and tools to converge and share their experiences, successes, and challenges in using SYCL for programming heterogeneous systems. The workshop will feature a mix of peer-reviewed talks, lightning talks, and an invited keynote talk. The intention is to create a deeper understanding of vendor-neutral high-performance heterogeneous programming through talks and demonstrations using SYCL and related tools, plus interaction with academic and industry experts from Europe and around the world. | Tobias Weinzierl, Professor, Durham University; Sameer Shende, Director of the Performance Research Lab, University of Oregon; Igor Vorobtsov, Sr. Compiler Technical Consulting Engineer, Intel; Igor Baratta, Research Associate, University of Cambridge; Peter Žužek, Principal Software Engineer, ComputeCpp Product Owner, Codeplay Software; Szilárd Páll, Researcher, KTH |
