Argonne National Laboratory’s expertise and leadership in extreme-scale computing will again be on display at the annual International Conference for High-Performance Computing, Networking, Storage and Analysis (SC18). This year’s event will take place Nov. 11–16, 2018, in Dallas, Texas.
More than 60 Argonne researchers will be on hand to share their latest research and insights on topics ranging from quantum computing and big data analysis to machine learning and algorithms and applications for exascale.
“Argonne has had a strong presence at SC since the conference series began 30 years ago,” said Valerie Taylor, director of Argonne’s Mathematics and Computer Science division. “It is a great place to learn about new tools and approaches, share best practices and connect with peers who share a common goal to push the boundaries of what’s possible with high-performance computing.”
As the future home to one of the world’s first exascale systems, an Intel-Cray supercomputer called Aurora, Argonne is continuing its long history of developing the technologies, tools and techniques needed to enable scientific breakthroughs on the supercomputers of today and tomorrow. To fully realize the potential of exascale computing, the laboratory is developing the framework by which to harness this immense computing power to an advanced combination of simulation, data science and machine learning.
“The convergence of simulation, data and learning is reframing the way we do scientific computing,” said Michael Papka, director of the Argonne Leadership Computing Facility (ALCF), a U.S. Department of Energy (DOE) Office of Science User Facility. “At SC18, Argonne researchers will be presenting their work on exploring new ways to enable data-driven discoveries across scientific disciplines.”
The laboratory’s conference activities will include technical paper presentations, invited talks, workshops, “birds of a feather” sessions, panel discussions and tutorials. In addition, Argonne will be partnering with the other DOE national laboratories to provide talks and demos at the DOE’s conference booth (#2433). Some notable Argonne activities are highlighted below.
Argonne computer scientist Kate Keahey co-authored a technical paper on a method to help address the growing need for on-demand analysis of experimental data. To do so, the researchers evaluated a system that dynamically moves nodes between a traditional high-performance computing (HPC) cluster and an on-demand cluster configured with cloud technology.
In the quantum computing space, Argonne researchers contributed to three papers that will be presented at the Workshop on Post Moore’s Era Supercomputing (PMES18). The papers cover how supercomputers can aid small and noisy quantum computers to simulate large circuits; how a quantum machine learning algorithm combined with a portable, architecture-agnostic hybrid quantum-classical framework can be used to solve real-world problems; and how the use of lossy compression can significantly increase the size of quantum circuit simulations with modeled noise.
“Although the technology for quantum computing is not quite there yet, now is the prime time for investing, and through basic science research, Argonne is invested in helping build the foundational knowledge necessary to bring the technology within reach,” said Argonne computational scientist Yuri Alexeev, a co-author on the three PMES18 papers.
Argonne computer scientist Pete Beckman will discuss the performance race between HPC systems and cloud computing at an invited talk at the DOE’s conference booth. Beckman’s talk will also cover Argo, an effort to develop an operating system and runtime system to support extreme-scale scientific computing; how “fluid HPC” could make large-scale systems more flexible; and how the HPC community might leverage these new technologies.
At another SC18 event, Paul Messina, director of Argonne’s Computational Science division and former director of the DOE Exascale Computing Project, will serve as the moderator for a panel titled “How Can Lessons Learned in the Past 40 Years Guide Future HPC Research Strategies?” Messina will lead a discussion with other HPC luminaries on the investment strategies, programming techniques and potential disruptions that will impact the HPC, computer science and mathematics communities in the coming decades.