Argonne Leadership Computing Facility

05/29/2026

Interested in learning how DAOS, Aurora’s high-performance storage system, can enhance your research workflows? Watch the video of our recent webinar for a deep dive into using DAOS, including creating containers, setting up a mount point, running applications with PBS scripts, analyzing output data, and transferring data to Flare.

Watch this webinar for an overview of DAOS on Aurora and then movin...

05/28/2026

Quantum technologies have the potential to transform computing, communication, and sensing, but advancing these capabilities requires the discovery of stable, scalable materials for qubits.

Using DOE supercomputers, including ALCF’s Polaris system, researchers from Argonne National Laboratory, The University of Chicago, and Linköping University identified a promising defect in magnesium oxide that could serve as a spin qubit. The team carried out large-scale calculations to predict the defect’s electronic and optical properties, providing a foundation for experimental synthesis and characterization. By combining high-throughput screening with advanced electronic structure calculations, they demonstrated a powerful protocol for accelerating the search for viable quantum defects across a wide range of materials.

https://www.alcf.anl.gov/science/case-studies/large-scale-simulations-materials-energy-and-quantum-information-science

05/27/2026

Congratulations to ALCF Director Michael Papka and longtime ALCF users Martin Berzins, Peter Coveney, Amanda Randles, Jackie Chen, and Gerbrand Ceder for being named to the inaugural SCW75 list from Scientific Computing World , honoring leaders who are redefining what is possible at the intersection of computing and research - https://www.alcf.anl.gov/news/alcf-director-and-users-named-inaugural-scw75-cohort

See the SCW75 website to learn more: https://www.scientific-computing.com/scientific-computing-world-75

05/26/2026

Now available for scientific research, the ALCF Inference Service provides researchers with access to a range of large language models and foundation models, helping accelerate the path from data to insight to discovery.

A key enabler for U.S. Department of Energy Genesis Mission, the service supports parallel inference workloads across multiple AI models. It is powered by Argonne National Laboratory's high-performance computing systems, including Sophia, an NVIDIA DGX cluster, and Metis, a SambaNova platform designed for high-throughput, low-latency AI inference workloads.

https://www.alcf.anl.gov/news/alcf-launches-first-large-scale-ai-inference-service-open-science

05/22/2026

Argonne National Laboratory researchers will have a strong presence at the 2026 IEEE International Parallel and Distributed Processing Symposium (IPDPS) next week, serving in leadership roles, organizing workshops, presenting tutorials, and sharing research on topics spanning AI for science, large language models, accelerators, scheduling, storage, reproducibility, and edge computing.

Among the conference activities, ALCF staff will contribute to the following co-located workshops:

📅 Heterogeneity in Computing Workshop (HCW) — Murali Emani, Technical Program Committee member

📅 29th Workshop on Job Scheduling Strategies for Parallel Processing (JSSPP 2026)
- "Fusing System Data to Navigate Power Saving Opportunities" — Melanie Cornelius, Greg Cross (ALCF), Shilpika (ALCF), Zhiling Lan (joint appointment with ALCF), Michael Papka (ALCF)
- "From Petascale to Exascale: Evolving User Behavior on the Polaris and Aurora Systems" — Niccolò Brembilla, Zhiling Lan, Michael Papka
- "SafeWall: A Risk Aware Job Walltime Recommendation for HPC Scheduling" — Kanglin Xu, Zhiling Lan, Michael Papka

📅 HPC for AI Foundation Models & LLMs for Science (HPAI4S’26)
- "Prefill/Decode-Aware Evaluation of LLM Inference on Emerging AI Accelerators" — Shun Usami, Venkatram Vishwanath (ALCF), E. Wes Bethel

For more on Argonne's contributions, see: https://www.alcf.anl.gov/events/40th-ieee-international-parallel-distributed-processing-symposium

05/21/2026

Using AI and the exascale power of ALCF's Aurora supercomputer, an Argonne National Laboratory team developed the AERIS foundation model to extend reliable weather forecasts beyond today’s 10-day limit, providing a tool that can help utilities and infrastructure operators prepare for storms and other disruptive events weeks in advance. Read the case study from Hewlett Packard Enterprise to learn how AERIS is enabling researchers to generate high-resolution forecast scenarios in minutes instead of days.

https://www.hpe.com/psnow/doc/a50015062enw

05/20/2026

From exascale supercomputing to powerful AI systems and services, the ALCF provides computing resources, capabilities, and expertise to help researchers accelerate discovery and innovation across science and engineering domains. Explore our annual report to learn how the facility and our user community are advancing HPC and AI for science: https://ar25.alcf.anl.gov

05/19/2026

At the Advanced Photon Source at Argonne National Laboratory, rapid data analysis powered by ALCF supercomputers is enabling scientists to adjust experiments on the fly, refine hypotheses, and make the most of their beam time. This capability, supported by tools like Globus, is also helping pave the way for new efforts under the American Science Cloud, a cornerstone of DOE's Genesis Mission.

Learn how the ALCF-APS integration is impacting research at the X-ray Photon Correlation Spectroscopy (XPCS) beamline: https://www.alcf.anl.gov/news/data-analysis-speed-light-source-experiments

05/18/2026

Researchers from the University of Dayton Research Institute and the Air Force Research Laboratory - AFRL leveraged ALCF supercomputers at Argonne National Laboratory to perform large-scale molecular simulations of hypersonic flight environments. By combining quantum chemistry with particle-based flow models, their work captures the complex interactions behind shock waves, intense heating, and reactive chemistry, providing insights to help advance the development of next-generation hypersonic vehicles.

https://www.alcf.anl.gov/science/case-studies/first-principles-simulation-hypersonic-flight

05/15/2026

Researchers from the University of Illinois Chicago and Argonne National Laboratory are collaborating on a new DOE-funded project to advance AI-powered scientific visualization. The effort brings together expertise and resources from ALCF and UIC's Electronic Visualization Laboratory, with ALCF's Victor Mateevitsi and UIC's Luc Renambot among the project collaborators. The team is developing an AI assistant that can compress massive scientific datasets more than 100-fold without losing fidelity, enabling more intuitive, interactive analysis to accelerate discovery.

Read the UIC announcement to learn more: https://cs.uic.edu/news-stories/grant-funds-ai-visualization-collaboration-with-argonne/