Refactoring for the Future: Adapting Nuclear Software to Emerging Hardware Architectures As high-performance computing platforms evolve toward GPU accelerated and distributed architectures, nuclear energy software must adapt to remain performant and relevant. This panel will explore the challenges and strategies involved in refactoring legacy CPU-based codes for modern systems, including exascale machines and native GPU approaches. Panelists will share lessons learned from projects like ExaSMR, Celeritas, and CeMeNT, and discuss how these transitions are shaping the future of nuclear simulation and modeling.

Panelists

  • Elliott Biondo (ORNL)

    Papers

    To access paper attachments, you must be logged in and registered for the meeting.

    Register NowLog In

    Current Issues in Computational Methods: Roundtable

    1:00–2:45PM EST

    Koroush Shirvan (MIT)

    Discussion

    To join the conversation, you must be logged in and registered for the meeting.

    Register NowLog In