The motivation for our research is the need for computer systems that scale without help from silicon technology. We believe that by rethinking how we design, operate, and program computer systems we can achieve improvements equivalent to several decades of Moore’s Law scaling. To seize this opportunity, we take a cross-cutting approach that revisits the function and interfaces of hardware and software layers. Hence, our work focuses on computer architecture, operating systems, and cloud management technology with the overall goal to improve the scalability and resource usage of current and emerging applications. Follow the links above to project descriptions and visit the publications page for related papers.
In past projects, we worked on transactional memory (TCC), dynamic information flow tracking (Raksha), enterprise systems power analysis (PowerNet), vector architectures, low-power processors (BLISS), etc.