Argonne Leadership Computing Facility

10/20/2026

How can supercomputers help scientists better understand the early universe?

Using the ALCF’s Aurora exascale system, an Argonne National Laboratory-led team carried out two of the largest cosmological simulations ever performed. Each modeled the evolution of 13.4 trillion dark matter particles across a region 28.6 billion light years wide.

The massive simulations will enable researchers to investigate subtle signatures of cosmic inflation and create synthetic full-sky maps and realistic mock observations to help improve the analysis of data from DESI, SPHEREx, and other major sky surveys.

https://www.alcf.anl.gov/news/argonnes-aurora-supercomputer-powers-two-largest-cosmological-simulations-date

10/09/2026

Argonne National Laboratory's annual Postdoctoral Research and Career Symposium brings together early career researchers to showcase their work, exchange ideas, and network with peers and leaders from industry, government, and academia.

Several postdocs working with ALCF staff and computing resources were among those presenting posters at today's event, highlighting research in AI, scientific visualization, computational chemistry, materials science, and high-performance computing. Their work includes developing AI agents to automate scientific workflows, optimizing communication and task scheduling on supercomputers, and using machine learning to improve simulations of materials at the atomic scale.

10/06/2026

From agentic AI and American Science Cloud workflows to numerical methods and HPC software, students from across the country joined the ALCF this summer to gain hands-on experience working on a range of projects to advance the use of HPC and AI for science. Check out this article to learn how five of the students — Kausalya Sankaranarayanan of Northeastern University, Minyeong Kim of The University of Texas at Austin, Marija Stojanoska of DePaul University, Michael Lance of University of Wyoming, and Akhilesh Bondapalli of Clemson University — tackled real-world computing challenges and what they took away from their experiences working at Argonne National Laboratory.

https://www.alcf.anl.gov/news/alcf-summer-students-explore-hpc-ai-and-scientific-computing

09/22/2026

Researchers from the University of Michigan developed MIST, a family of open-source AI foundation models designed to accelerate molecular property prediction and materials discovery across vast chemical spaces.

Using Argonne National Laboratory's Polaris supercomputer, the team pretrained the models on large collections of molecular data and fine-tuned them for nearly 400 molecular and mixture property prediction tasks, releasing model checkpoints, datasets, training code, and tutorials to help scientists advance chemistry and materials research.

https://www.alcf.anl.gov/science/case-studies/mist-open-source-foundation-models-chemistry-and-materials-discovery

09/21/2026

By connecting Advanced Photon Source beamlines with ALCF supercomputers, scientists can analyze X-ray data in near real time as experiments unfold, allowing them to adjust experimental parameters on the fly, refine hypotheses, and make the most of their beam time.

This capability is also being extended to other data-intensive experimental facilities beyond Argonne National Laboratory while helping lay the groundwork for new efforts under the American Science Cloud, a cornerstone of DOE’s Genesis Mission.

The integration of Advanced Photon Source experiments with Argonne Leadership Computing Facility supercomputers enables scientists to analyze data in near real time, guiding their work and accelerating discoveries.

09/17/2026

"Where Data Becomes Light" returns to ART on THE MART tonight! Presented by Argonne National Laboratory and University of Illinois Chicago, the installation brings scientific visualizations of atomic interactions, turbulent flows, hurricanes, supernovae, and more to life at a massive scale.

Featuring visualizations from the ALCF and UIC’s Electronic Visualization Laboratory, the free show runs Thursdays through Sundays, 7:30–8 p.m., until Nov. 15, with viewing along the Chicago Riverwalk between Wells and Franklin streets.

Read more about the event and last week’s opening night: https://www.alcf.anl.gov/news/scientific-visualizations-light-chicago-art-mart

09/17/2026

In work recognized as a finalist for the 2026 ACM Gordon Bell Prize, a multi-institutional team used DOE’s exascale supercomputers to carry out the largest high-order multiphysics simulations of a liquid-metal fusion breeding blanket to date.

Researchers from Penn State, University of Illinois Urbana-Champaign, Argonne National Laboratory, and UK Atomic Energy Authority leveraged ALCF’s Aurora and OLCF’s Frontier to run simulations with the NekRS code that exceeded two trillion grid points and captured the coupled effects of magnetohydrodynamics, turbulent flow, heat transfer, tritium transport, and other processes across a full fusion blanket at engineering scale for the first time. The team’s work advances the use of extreme-scale simulation to support the development of future fusion power plants, demonstrating capabilities that can complement experiments and guide optimization while generating data to help pave the way for automated, AI-driven design.

https://www.alcf.anl.gov/science/case-studies/exascale-multiphysics-simulations-liquid-metal-fusion-blankets

09/16/2026

Join us on Sept. 30 for a hands-on session to learn how to connect jobs, data, computing resources, and AI to build agentic workflows for science.

As part of the ALCF’s Service-Enabled Science training series, Argonne National Laboratory's Benoit Côté, Christine Simpson, Misha Salim, Ethan Wong, and Huihuo Zheng will guide attendees through using ALCF and Globus.org services to manage jobs and data, working with the ALCF Inference Service, integrating AI models with agent environments such as Claude Code and Codex, and building a working agentic workflow they can adapt to their own research.

Register here: https://www.alcf.anl.gov/events/service-enabled-science-jobs-data-ai-agents

09/15/2026

Using ALCF’s Aurora supercomputer, Argonne National Laboratory researchers completed two of the largest cosmological simulations ever performed, each evolving approximately 13.4 trillion dark matter particles in a cube 28.6 billion light years across. The simulations will support analysis of data from DESI, SPHEREx, and other surveys, helping scientists advance our understanding of cosmic inflation and the physics of the early universe.

https://www.alcf.anl.gov/science/case-studies/simulating-universe-unprecedented-scale

09/14/2026

Visitors to the Vintage Computer Festival Midwest this weekend got a close-up look at some of the technology that has helped shape Argonne National Laboratory’s long history in scientific computing.

The “Big Iron for Big Science” exhibit featured several computing artifacts used for research at the laboratory, including hardware from ALCF’s IBM Blue Gene and Intel Xeon Phi systems, along with networking and scientific equipment.