Projects

Click on a project to learn more about its people, publications, and funding.

Energy-Minimal Programmable Architectures for Intelligence Beyond the Edge

This project is building a new class of ultra-low-power systems to enable sophisticated on-device computation (e.g., machine learning) in the Internet of Things (IoT). (Collaboration with Brandon Lucia.)

Near-Data Dataflow Architectures

This project is revisiting classic dataflow concepts, reconfigurable computing, and near-data computing to develop a new class of “general-purpose accelerators” that will finally overcome dark silicon and the memory wall.

Polymorphic Cache Hierarchy

This project is developing a “polymorphic” memory hierarchy that can be re-programmed by applications to significantly improve performance and efficiency, e.g., by moving computation to execute within the cache hierarchy.

Smart Caching at Datacenter Scale

Datacenter applications must serve petabytes of data at high throughput and minimum cost. This project merges caching theory with systems to build smart, efficient, and low-cost caching solutions for the datacenter.

Software-Defined Cache Hierarchy for Multicore Processors

This project dynamically re-organizes the large, distributed shared cache in modern multicore processors to keep applications’ data as close as possible, letting applications meet their goals with minimal data movement.