Argonne scientists are finalists for several prestigious Gordon Bell Prizes, which celebrate major achievements in HPC. In addition, the Aurora supercomputer was used in another finalist project led by University of Southern California and Lawrence Berkeley National Laboratory that achieved a new level of accuracy in simulating the behavior of quantum materials.
Argonne is developing cutting-edge artificial intelligence (AI) tools and methods to help enhance scientific discovery. The research takes several different avenues — from large language models to generative AI. Regardless of the road taken, the challenges faced are numerous. To help users and developers, Argonne researchers will offer three half-day tutorials: one on middleware for supporting agentic
Combining the power of AI and HPC is a hot topic. Rick Stevens, associate laboratory director of Argonne’s Computing, Environment and Life Science directorate, will participate in a panel on next-generation computing architectures — the status, challenges and opportunities for integrating these technologies into today’s computing infrastructure and industry standards. In another panel session Argonne
Leonardo da Vinci once said: “Study the science of art. Study the art of science.” In that vein, SC25 is showcasing Art of HPC. Chaired by Joseph Insley, Visualization and Data Analytics Team lead at the Argonne Leadership Computing Facility, Art of HPC will feature exhibits and presentations at the intersection of art expression and
The team behind the 2005 paper “The Globus Striped GridFTP Framework and Server” will receive this year’s Test of Time Award at the SC25 supercomputing conference. The paper is being honored for its foundational role in high performance computing (HPC) and its lasting influence on research and practice in the field. The Test of Time
Technical demonstrations in the DOE exhibit booth highlight several Argonne projects including a path toward more powerful, sustainable and collaborative scientific software ecosystems. Other sessions feature portable visualization tools and a framework that allows scientific applications to integrate AI/ML engines in the workflow. Katrin Heitmann gives a featured talk on the OpenCosmo project which shares
Reproducibility is essential in science — results must be reliable. Several Argonne colleagues will join forces to discuss addressing reproducibility challenges in HPC. Anshu Dubey will co-lead a half-day tutorial on practices, processes and tools to increase the sustainability of software artifacts and enhance trustworthiness in their use. Kyle Chard, a research associate professor at
Gaining improved performance from computers has always been a major concern. One approach is to use devices such as graphics processing units; another approach involves AI accelerator chips to speed up machine learning models. Using accelerators effectively, however, remains a challenge. Argonne scientists will describe their work on developing a RISC-V Accelerator, which can be
Simulations, experiments and AI computations are producing enormous volumes of data that must be transferred, stored and analyzed. In a three-part tutorial, researchers from Argonne’s Mathematics and Computer Science division and Lawrence Livermore National Laboratory (LLNL) will discuss will discuss how to use lossy compression to reduce file sizes without losing important details. Moreover, in
To get the most out of supercomputers, scientists often need to coordinate across many systems. Researchers from Argonne and Lawrence Berkeley National Laboratory will collaborate in a workshop on experiment-in-the-loop computing, in which they will describe an infrastructure that advances light source science through multifacility HPC workflows. Argonne will also collaborate with researchers from LLNL