Argonne Leadership Computing Facility

A University of Michigan team, led by Venkat Viswanathan, is harnessing Argonne National Laboratory's Polaris supercomputer to develop AI-driven large language models to accelerate the discovery of new materials for energy storage and conversion applications. Read the story from ASCR Discovery: https://lnkd.in/gWRitkhM   In 2025, the team plans to create a large-scale crystal dataset to train a foundation model capable of accurately predicting the thermodynamic and mechanical properties of potential battery electrode materials. Check out the 2025 INCITE project fact sheets to learn more about this and other projects set to leverage ALCF supercomputers next year: https://lnkd.in/gAhXrBf3

Science isn’t slowing down. We collect more data and need answers to more complex questions than ever before. That’s why we need fast and powerful tools like the Aurora supercomputer. Aurora has again been named top system in the world for AI performance, earning the #1 spot the HPL-MxP benchmark for mixed precision computations at #SC24. In a new video, Katherine Riley, Director of Science for the Argonne Leadership Computing Facility, explains how the exascale system’s speed and computational power are transforming the questions we can answer and helping researchers make breakthroughs in everything from human health to energy resilience to our understanding of the cosmos. Watch the interview - https://bit.ly/4i0Q0Tm #Supercomputing #AI #HPC

Using ALCF supercomputing resources, Argonne National Laboratory's Venkat Viswanathan and Filippo Simini collaborated with researchers from Swift to explore the generation of high-utility synthetic datasets of cross-border payment transactions with transformer models, advancing AI-driven fraud prevention capabilities. https://lnkd.in/guR34Ju3

As a founding partner of the international Trillion Parameter Consortium (TPC), Argonne National Laboratory is collaborating with researchers and organizations from across the world to tackle the challenges of building, training, and using large-scale generative AI models for science, as well as designing and operating the large-scale computing systems that enable the development of these models.   To learn about TPC’s latest advances, join us at #SC24 on Friday for “Accelerating the Development and Use of Generative AI for Science and Engineering: The Trillion Parameter Consortium (TPC)” - https://lnkd.in/ghTSh76e Argonne’s Rick Stevens, Valerie Taylor, and Charlie Catlett are among the session chairs for the half-day workshop, which includes several talks from Argonne researchers and collaborators. To learn more about TPC, visit: https://tpc.dev/ For details on all of Argonne's SC24 activities, visit: https://lnkd.in/g4-PdPdR

From accelerating the discovery of new materials to advancing our understanding of the cosmos and the human brain, the ALCF's #HPC and #AI resources support a wide range of groundbreaking research projects. Check out our 2024 Science Report to learn about some of the ALCF user community's recent breakthroughs in science and engineering. ➡️ https://lnkd.in/gy6ceiTm

Congratulations to Argonne National Laboratory researchers and collaborators for being recognized with six 2024 HPCwire Awards at #SC24 this week!   - Editors’ Choice: Best Use of HPC in Physical Sciences Scientists at Argonne National Laboratory, University of Chicago, University of Illinois Urbana-Champaign, National Center for Supercomputing Applications, and University of Minnesota used Globus.org for data management, ALCF's Polaris supercomputer, and NCSA's Delta and DeltaAI supercomputers to develop a physics-informed transformer model to predict gravitational wave evolution for spinning binary black hole mergers, including higher-order modes. This AI approach dramatically reduces simulation time from days to seconds, handling terabyte-scale datasets with high accuracy.   - Editors’ Choice: Best HPC Response to Societal Plight The Open Science Platform OSPREY aims to enhance pandemic response by enabling health officials to utilize HPC resources and data-driven decision-making. With support from Argonne and University of Chicago, the researchers used Globus.org, Parsl, and EMEWS to integrate automated workflows, data curation, and model management to facilitate rapid collaboration and development during health crises.   - Readers’ Choice: Best HPC In the Cloud Globus.org enabled near-real-time data analysis at Argonne by connecting instruments at the Advanced Photon Source with ALCF supercomputers. This automated pipeline allows scientists to adjust experiments on the fly, potentially accelerating scientific breakthroughs by delivering rapid results while researchers still have facility access.   - Readers’ Choice: Best Use of High Performance Data Analytics & AI Argonne and Dow developed a framework combining computational fluid dynamics with an active machine learning optimizer (ActivO) for efficient turbulent jet mixer design. This novel approach optimized jet-mixing technology, potentially reducing reliability issues and costs associated with traditional agitators, with estimated savings of up to $6.1 million per year per plant.   - Readers’ Choice: Best Use of HPC in Industry Argonne and RTX Technology Research Center used HPC and CFD modeling to simulate gas turbine film cooling with surface roughness defects. They employed Argonne’s GPU-accelerated NekRS solver on DOE supercomputers, providing high-fidelity data for developing surrogate models to optimize thermal management in next-generation aircraft engines.   - Readers’ Choice: Top Supercomputing Achievement Argonne's Aurora supercomputer showcases AI’s growing impact in supercomputing. With its massive GPU cluster and advanced interconnect, Aurora enables AI-driven research across fields like neuroscience, particle physics, and drug discovery. This system demonstrates how AI can be effectively implemented for high-performance computing, accelerating scientific breakthroughs.   https://lnkd.in/gqZMNgTb

Tip of the Day: Dress for the job you want! Aurora is dressing for scientific success. This breathtaking design will soon appear on our Aurora supercomputer's new doors at the Argonne Leadership Computing Facility. The images illustrate the bold research initiatives Aurora is ushering in, spanning across areas like energy, health, materials science, manufacturing, and more. The design represents a creative collaboration between Argonne’s researchers, visualization experts, communications specialists, and graphic artists. These teams work hard every day to discover, learn, and share groundbreaking discoveries. Aurora: Dressed for success, ready for science. Big, bold, beautiful science. #SC24 #Supercomputing #AI #HPC

Argonne National Laboratory researchers are contributing to the open-source NVIDIA BioNeMo Framework, a collection of accelerated computing tools designed to exponentially scale AI models for biomolecular research.   “Argonne has contributed billion-parameter biological models that require specialized software to train across high-performance computing environments,” said Argonne’s Arvind R.. “With BioNeMo, Argonne and the broader biotech community gain an enterprise-level, open-source solution that enables researchers to easily scale the training of large biological foundation models — in labs that otherwise wouldn’t have the computational expertise to do so.” https://lnkd.in/dE_eVUvt

Explore the intersection of art and science at #SC24’s Art of HPC Display! Argonne National Laboratory's Joseph Insley and the ALCF Visualization and Data Analytics Team have four amazing images on display in the B301 gallery, including this visualization from a 3D supernova simulation carried out on ALCF supercomputers. https://lnkd.in/gvHG62jh

Catch ALCF researchers at #SC24 today to learn how Argonne National Laboratory and our fellow national laboratories are driving innovation in HPC and AI for science.   - ALCF’s Venkatram Vishwanath and Riccardo Balin will contribute the BOF session “Two Worlds Collide: Trustworthiness and Energy Efficiency for Coupled HPC+AI Simulation” at 12:15pm (EST) - https://lnkd.in/g-6NJBbp   - ALCF’s David Martin will contribute to the featured talk “HPC4EI: Bringing National Lab Scale Supercomputing to US Industry” at 1:45pm (EST) at the DOE booth (#3401) - https://lnkd.in/gMc4mTYh   For a full list of Argonne’s conference activities, visit: https://lnkd.in/g4-PdPdR