PSC and Collaborating Organizations Recognized by Two 23rd Annual HPCwire Awards: POPri Tool for Avoiding Exposure of User Data and No-Cost-to-Users NSF HPC Collaboration

A method for generating synthetic data that represents user preferences, without sending actual user data onto the Internet, as well as a national collaboration giving academic users no-cost access to NSF-funded supercomputers have been recognized in the 23rd annual HPCwire Readers’ and Editors’ Choice Awards. The method, called POPri, greatly reduces security risks to users’ data and was developed by scientists at Carnegie Mellon University and the Pittsburgh Supercomputing Center. The collaboration, ACCESS, provides computation on systems including PSC’s Bridges-2 to academic researchers at no cost.

The awards were presented at the 2026 International Conference for High Performance Computing, Networking, Storage, and Analysis (SC26), in Chicago. The list of winners was revealed at the SC26 HPCwire booth, as well as on the HPCwire website: https://www.hpcwire.com/2026-hpcwire-awards-readers-editors-choice.

“Recognition like this reflects what PSC does best,” said Barr von Oehsen, Director of PSC. “Our team’s expertise and Bridges-2 make research possible that otherwise wouldn’t be. POPri is a good example, giving researchers a way to learn from user preferences without exposing anyone’s private data. We’re proud to see that work honored by the HPCwire community.”

PSC and its collaborating institutions were recognized with two HPCwire Reader’s Choice Awards, selected by the publication’s high performance computing professional readers:

  • Best Use of HPC/AI in Financial Services
    A team led by Carnegie Mellon University and the Pittsburgh Supercomputing Center developed a new approach called POPri (Preference Optimization for Private Client Data), which generates safe synthetic data to train large language models without exposing private user information. The work allows researchers to personalize LLMs to users’ preferences, avoiding potentially exposing financial and other data to bad actors.
  • Best HPC/AI Collaboration, Academia/Government/Industry
    The NSF ACCESS program and its fourteen integrated resource providers are working together to provide the nation’s research and education community with the computational capacity and support services needed to advance the state of the art in AI and all fields of study by bringing together government, academia, and technology partners. The effort involves a nationwide ecosystem of HPC, cloud, AI, data, and storage resources operated by leading institutions including NCSA, TACC, PSC, SDSC, Indiana University, and Purdue University. Technologies from major industry partners such as NVIDIA, AMD, HPE, Dell, Intel, and associated software ecosystems are widely used.

The coveted annual HPCwire Readers’ Choice Awards are determined through a nomination and voting process with the global HPCwire community, as well as selections from the HPCwire editors. The awards are an annual feature of the publication and constitute prestigious recognition from the HPC community. They are revealed each year to kick off the annual supercomputing conference, which showcases high performance computing, networking, storage, and data analysis.

“As HPCwire celebrates 38 years of covering high performance computing, the combined power of HPC and AI is accelerating discovery and reshaping what science and industry can achieve,” said Tom Tabor, CEO of TCI Media, publishers of HPCwire. “Our 2027 Readers’ and Editors’ Choice Awards celebrate the people, organizations, and innovations driving that progress. Throughout the world, our community is tackling grand challenges whose solutions will improve lives for generations to come. Yet too often, these accomplishments remain out of the spotlight, and the people behind them receive even less recognition for their contributions to society. These awards help bring their achievements to light.”

More information on these awards can be found at the HPCwire website (www.HPCwire.com) or on X through the following hashtag: #HPCwireRCA26.