Argonne’s data compression software honored at global symposium
A team led by the U.S. Department of Energy’s Argonne National Laboratory was recognized at the 2025 IEEE International Parallel & Distributed Processing Symposium (IPDPS) in Milan, Italy, for their achievements in scientific data compression.
A foundational element in the team’s work has been SQUEEZE (SZ), a lossy data compression framework designed to reduce the size of scientific data without losing critical information. Lossy compression is increasingly important as modern science generates massive amounts of data that must be transferred and stored for analysis.
A History of Innovation
SZ was first introduced at IPDPS in 2016 by Argonne researchers Sheng Di and Franck Cappello. It was one of the first compression tools built for high-performance computing that allowed users to set error limits for their data.
That initial work spurred numerous enhancements to the SZ suite and the growth of the Argonne compression team. Developers and users both inside and outside Argonne added new features to meet the evolving needs of scientific applications and hardware. The success of the team at the IPDPS 2025 symposium testifies to the impact the work is having in the computing world.
“In 2016, a compression paper at IPDPS was a rarity,” said Cappello, a senior computer scientist in the Mathematics and Computer Science (MCS) division at Argonne and head of the lossy compression team. “But now, in 2025, the situation has changed remarkably: around 10% of the papers presented at IPDPS were on data compression — including six papers presented by SZ team members, and one paper winning the symposium’s Best Paper Award.”
Best Paper Highlights
The winning paper, “Enabling Efficient Error-controlled Lossy Compression for Unstructured Scientific Data,” involved researchers from Argonne, the University of Kentucky and The Ohio State University. It presents a generic framework that integrates several methods for unstructured meshes, offering improved performance over leading compressors (see Fig. 1).
Di, a computer scientist in the Mathematics and Computer Science (MCS) division at Argonne, co-author of the paper and core developer of SZ, accepted the award on behalf of the team. “This was a collaborative effort in which each of the team members contributed expertise and insights into how to better handle data transmission and storage issues associated with data compressors,” he said.
Contributions and Challenges
The other papers by the SZ team were presented in a full-day track at the symposium. The researchers discussed the expanding capabilities of lossy compressors and the challenges open for future research:
-
Trade-offs between speed, energy use and compression.
-
Fast compression on CPUs and GPUs.
-
Memory-efficient compression workflows.
-
Smarter data prediction techniques.
-
Fast, accurate sampling techniques of compression ratio.
“It is exciting to see data compression become a central topic in computing,” said Robert Underwood, an assistant computer scientist in Argonne’s MCS division and co-author of three of the papers.
“Data compression is an established part of IPDPS!” Cappello declared. He hinted that more SZ papers are finalists for other top conference awards later in 2025.
For more information about the event and abstracts of the papers, see the IPDPS 2025 program. For more information about the SZ software, visit the website.
Argonne National Laboratory seeks solutions to pressing national problems in science and technology by conducting leading-edge basic and applied research in virtually every scientific discipline. Argonne is managed by UChicago Argonne, LLC for the U.S. Department of Energy’s Office of Science.
The U.S. Department of Energy’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of the most pressing challenges of our time. For more information, visit https://energy.gov/science.