This Project Grant award of $450,000 from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070) aims to develop novel approaches to electronic design automation (EDA) for creating high-performance and efficient computer hardware.
The research introduces a strategy that combines formal techniques with learning-based optimization to enable differentiable hardware synthesis, particularly suited for heterogeneous computing. This new approach seeks to address the limitations of existing EDA solutions by facilitating efficient global optimization, with or without training data, while leveraging the computational power of parallel computing devices like GPUs. The project's findings and open-source software tools will be made publicly available to support technology transfers and industry-academia collaborations in this multidisciplinary field. Additionally, the effort will involve educational and workforce development initiatives targeting high school students and underrepresented groups.
Generated 1/28/25, 7:21 AM