EFICAS

Energy Efficient Heterogeneous AI-Framework for Smart Mobile and Embedded Systems

Project description

Essentially all mobile applications are severely power limited, which blocks huge business cases. Increasing functional complexity in mobile and autonomous applications impact the computational load by increasing power demands of embedded platforms. EFICAS introduces a generic and scalable software platform supporting energy-efficient deployment of AI algorithms on the multicore heterogeneous computation technologies. It supports all technology solutions, including localised and distributed computation settings as well as cloud offloading. It addresses resource allocation at runtime and hybrid coherent operation with optimised task allocation at design time.