This National Science Foundation (NSF) Computer and Information Science and Engineering (CISE) grant (CFDA 47.070) awarded to George Washington University (GWU) provides $424,291 over 4 years to research the development of a novel dual-purpose photonic fabric that can enable both power-efficient on-chip data communications and high-performance neural network acceleration for future heterogeneous chiplet-based computing architectures. The key objectives are to: (1) provide high-bandwidth and energy-efficient data transfer between processor cores and accelerators during high network load, and (2) offload select accelerator functions to the photonic network during low load periods to maximize hardware utilization and accelerate computations. This research aims to improve the power-efficiency, performance, and scalability of on-chip communication fabrics for next-generation heterogeneous computing systems through the innovative integration of silicon photonics, architecture, and machine learning techniques.
Mod # | Description | Reason For Modification | Federal Obligation (Click to sort descending) | Date (Click to sort ascending) |
|---|---|---|---|---|
| Not listed | $424.3k | 7/26/23 |