Return to Funding
2024: Energy-efficient Communication for AI Using Optical Network Fabrics
The rapid expansion of artificial intelligence is consuming huge amounts of energy. One major area of inefficiency is communication bottlenecks between graphics processing units (GPUs). Cornell researchers will design light-based, optical network fabrics that enable faster communication between GPUs and bypass bottlenecks caused by slower electrical connections.
Investigators: Rachee Singh, Computer Science; Emaad Manzoor, Marketing and Management Communication