Research in Energy-efficient Embedded Computing

Our research on Parallel Embedded Systems focuses on streaming computations on heterogeneous many-core embedded systems. The design and implementation of such systems needs a coordinated effort to co-develop both the hardware architectures and the application development environment. Our research objective is to develop a programming framework that will improve programmer productivity through the use of domain specific languages while still allowing program portability between a multitude of heterogeneous architectures using many-core virtual machine concepts and architecture description languages. With regard to hardware we focus on how parallel heterogeneous architectures should be organized, designed, and evaluated, as well as how they should interface with the application development tools and frameworks.