The National Science Foundation Office of Advanced Cyberinfrastructure awarded Oakland University $180,000 on October 1, 2025, to develop Scientific GPU Compression Cyberinfrastructure (SGCC), a user-friendly, GPU-accelerated data reduction system for scientific data analysis on heterogeneous supercomputing systems. The award is made under the Computer and Information Science and Engineering program (CFDA 47.070).
SGCC addresses gaps in existing GPU-based scientific data compression frameworks by creating an end-to-end cyberinfrastructure that combines portability across multiple GPU architectures with user-friendly interfaces. The project ports, extends, and optimizes multiple existing capabilities including the CUSZ family of error-bounded lossy compressors, GPU-based lossless encoders, QCAT (a CPU-based compression quality assessment toolkit), the Kokkos ecosystem for multi-backend performance portability, LibPressIO (a unified programming interface for scientific compressors), and HDF5. The infrastructure targets exascale data production challenges by improving data analysis efficiency on GPU-equipped supercomputing systems and accelerating scientific discovery. The project integrates education and training components through enhanced computing-related curricula in heterogeneous scientific computing, data management, and visualization for graduate students.
Work is performed in Rochester, Michigan. The period of performance runs from October 1, 2025, through September 30, 2028. This is a Project Grant, a form of assistance funding.