Denver Hosting Supercomputing 2019 (SC19); Argonne National Laboratory Researchers to Tout Quantum Computing


Full release…

The Supercomputing 2019 (SC19) conference, scheduled for November 17–22 in Denver, will bring together the global high-performance computing (HPC) community, including researchers from the U.S. Department of Energy’s (DOE) Argonne National Laboratory, to share scientific computing advances and insights with an eye toward the upcoming exascale era.

Continuing the laboratory’s long history of participation in the SC conference series, more than 90 Argonne researchers will contribute to conference activities and studies on topics ranging from exascale computing and big data analysis to artificial intelligence (AI) and quantum computing.

“SC is a tremendous venue for Argonne to showcase its innovative uses of high-performance and data-intensive computing to advance science and engineering,” said Salman Habib, director of Argonne’s Computational Science division. “We look forward to sharing our research and connecting with and learning from our peers, who are also working to push the boundaries of extreme-scale computing in new directions.”

As the future home to one of the world’s first exascale supercomputers — Aurora, an Intel-Cray machine scheduled to arrive in 2021— Argonne continues to drive the development of technologies, tools and techniques that enable scientific breakthroughs on current and future HPC systems.

To fully realize exascale’s potential, the laboratory is helping to frame the convergence of AI, machine learning and data science methods alongside traditional modeling and simulation-based research.

“We are seeing rapid advances in the application of deep learning and other forms of AI to complex science problems at Argonne and across the broader research community,” said Ian Foster, director of Argonne’s Data Science and Learning division, Argonne Distinguished Fellow and also the Arthur Holly Compton Distinguished Service Professor of Computer Science at the University of Chicago. “SC provides a forum for the community to get together and share how these methods are being used to accelerate research for a diverse set of applications.”

The laboratory’s conference activities will include technical paper presentations, talks, workshops, ​“birds of a feather” sessions, panel discussions and tutorials. In addition, Argonne will partner with other DOE national laboratories to deliver talks and demos at the DOE’s conference booth (#925). Some notable Argonne activities are highlighted below. See the full schedule of the laboratory’s conference participation here.

DOE Booth Talk: Scientific Domain-Informed Machine Learning

Argonne computer scientist Prasanna Balaprakash will deliver a talk at the DOE booth on the laboratory’s pivotal research with machine learning. His featured talk will cover Argonne’s efforts to develop and apply machine learning approaches that enable data-driven discoveries in a wide variety of scientific domains, including cosmology, cancer research and climate modeling. Balaprakash will highlight successful use cases across the laboratory, as well as some exciting avenues for future research.

In-Situ Analysis for Extreme-Scale Cosmological Simulations

Argonne physicist and computational scientist Katrin Heitmann will deliver the keynote talk at the In Situ Infrastructures for Enabling Extreme-scale Analysis and Visualization (ISAV 2019) workshop. Her talk will cover the development of in situ analysis capabilities (i.e., data analysis while a simulation is in progress) for the Hardware/Hybrid Accelerated Cosmology Code, which has been used to carry out several extreme-scale simulations on DOE supercomputers. Heitmann will discuss the current limitations of her team’s “on the fly” analysis tool suite and how they are developing solutions to prepare for the arrival of DOE’s forthcoming exascale systems.

Full-State Quantum Circuit Simulation by Using Data Compression

Researchers from Argonne and the University of Chicago will present a technical paper on their work to develop a new quantum circuit simulation technique that leverages data compression, trading computation time and fidelity to reduce the memory requirements of full-state quantum circuit simulations. Demonstrated on Argonne’s Theta supercomputer, the team’s novel approach provides researchers and developers with a platform for quantum software debugging and hardware validation for modern quantum devices that have more than 50 qubits.

Deep Learning on Supercomputers

Argonne scientists will have a strong presence at the Deep Learning on Supercomputers workshop. Co-chaired by Foster, the workshop provides a forum for researchers working at the intersection of deep learning and HPC. Argonne researchers are part of a multi-institutional team that will present “DeepDriveMD: Deep-Learning-Driven Adaptive Molecular Simulations for Protein Folding.” The study provides a quantitative basis by which to understand how coupling deep learning approaches to molecular dynamics simulations can lead to effective performance gains and reduced times-to-solution on supercomputing resources.

A research team from Argonne and the University of Chicago will present “Scaling Distributed Training of Flood-Filling Networks on HPC Infrastructure for Brain Mapping” at the Deep Learning on Supercomputers workshop. The team’s paper details an approach to improve the performance of flood-filling networks, an automated method for segmenting brain data from electron microscopy experiments. Using Argonne’s Theta supercomputer, the researchers implemented a new synchronous and data-parallel distributed training scheme that reduced the amount of time required to train a flood-filling network.

Priority Research Directions for In Situ Data Management: Enabling Scientific Discovery from Diverse Data Sources

At the 14th Workshop on Workflows in Support of Large-Scale Science (WORKS19), Argonne computer scientist Tom Peterka’s keynote talk will cover six priority research directions that highlight the components and capabilities needed for in situ data management to be successful for a wide variety of applications. In situ analysis tools can enable discoveries from a broad range of data sources — HPC simulations, experiments, scientific instruments and sensor networks — by helping researchers minimize data movement, save storage space and boost resource efficiency, often while simultaneously increasing scientific precision.

The Many Faces of Instrumentation: Debugging and Better Performance using LLVM in HPC

Argonne computational scientist Hal Finkel will deliver a keynote talk on the open-source LLVM compiler infrastructure at the Workshop on Programming and Performance Visualization Tools (ProTools 19). LLVM, winner of the 2012 ACM Software System Award, has become an integral part of the software-development ecosystem for optimizing compilers, dynamic-language execution engines, source-code analysis and transformation tools, debuggers and linkers, and a host of other programming language- and toolchain-related components. Finkel will discuss various LLVM technologies, HPC tooling use cases, challenges in using these technologies in HPC environments, and interesting opportunities for the future.

Source:  Argonne National Laboratory .  James R. Collins,  ARGONNE RESEARCHERS TO SHARE SCIENTIFIC COMPUTING INSIGHTS AT SC19…

Content may have been edited for style and clarity.