Project Grant 2504592

This Project Grant award, funded by the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070), supports the development of a unified framework to reduce data movement overheads in heterogeneous high-performance computing (HPC) systems. The $300,000 award to the University of Utah will integrate three core components: a cross-layer monitoring and learning framework to characterize data transfer patterns and predict contention; a heterogeneity-aware data movement scheduler to coordinate bandwidth usage; and a collaborative caching and prefetching architecture to anticipate future data needs across workflows. This research aims to treat data movement as a first-class task, parallel to computation, and use analytical and machine learning techniques to reduce interference and improve overall throughput for scientific computing applications such as molecular dynamics, graph neural networks, and particle-in-cell simulations. The project is expected to run from October 1, 2025 to September 30, 2028 and will validate the framework through representative workloads on petascale and exascale systems.