Nine research projects have been selected to run on the forthcoming Discovery supercomputer on its first day of operation at Oak Ridge National Laboratory, ensuring immediate scientific output upon deployment. The projects, developed through the Discovery Center for Accelerated Application Readiness program, are designed to fully utilize the capabilities of the nation’s next flagship supercomputer and contribute to the ambitious Genesis Mission, a Department of Energy initiative connecting high-performance computing with emerging quantum computing technologies. “Our goal is to develop a suite of leadership-class computing applications that are ready to run science campaigns on day one,” said Reuben Budiardja, CAAR programming lead and group leader for ORNL’s Advanced Computing for Nuclear, Particle, and Astrophysics Group. Technical support for optimizing these applications comes from the ORNL HPE/AMD Center of Excellence, a collaborative team including staff directly from AMD and HPE working alongside OLCF scientists.
This collaboration aims to accelerate solutions to complex global challenges. Beyond code optimization, CAAR projects are incorporating advanced methodologies like surrogate models and mixed-precision computing, and are exploring agentic AI to navigate complex simulations. The selected projects, spanning astrophysics to aerospace engineering, must demonstrate a performance increase of three to five times over OLCF’s current exascale supercomputer, Frontier. “We’re looking forward to showing the world what Discovery can do,” said Matt Sieger, project director for the Discovery supercomputer.
We’re always seeking opportunities to collaborate with industry and agency partners. Organizations like NASA and GE operate at the leading edge of science and technology, driving innovation across critical fields.
Tom Beck, section head of Science Engagement at ORNL’s National Center for Computational Sciences
Discovery’s Architecture Integrates AI, HPC, and Quantum Technologies
The convergence of high-performance computing, artificial intelligence, and quantum technologies is no longer a distant prospect; the incoming Discovery supercomputer at Oak Ridge National Laboratory exemplifies this integration. As a key component of the Department of Energy’s Genesis Mission, Discovery is designed to connect AI and HPC capabilities with emerging quantum computing technologies, aiming to accelerate solutions to previously intractable problems. This isn’t simply about adding AI as an afterthought; many of the nine initial research projects selected for “day one” operation are expected to incorporate novel methodologies like surrogate models and agentic AI to enhance simulations and explore complex datasets. This deep partnership signifies a move beyond traditional vendor relationships, fostering a more integrated approach to hardware and software co-design. The CAAR program, responsible for preparing these applications, will also focus on strengthening user environments and developing training resources for the broader scientific community. “We’re laying the groundwork for the next generation of supercomputers that will feature a convergence of AI, HPC and quantum computing.”
This is a really special time for computing. We’re laying the groundwork for the next generation of supercomputers that will feature a convergence of AI, HPC and quantum computing. And, getting to see how all of this comes together when Discovery is deployed and ready to run early science and deliver results is something we’re very excited about.
Matt Sieger, project director for the Discovery supercomputer
The pursuit of enhanced computational performance is exemplified by nine research projects poised to immediately utilize the forthcoming Discovery supercomputer; this proactive approach to application readiness is unusual for a system of this scale. The selected applications represent a broad range of disciplines, from astrophysics to aerospace engineering, and include partnerships with industry leaders like General Electric and NASA.
Our goal is to develop a suite of leadership-class computing applications that are ready to run science campaigns on day one.
Reuben Budiardja, CAAR programming lead and group leader for ORNL’s Advanced Computing for Nuclear, Particle, and Astrophysics Group
See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
