We investigate the technologies that support scalable high-performance computing, including hardware, software, and theory.

Our goal is to make scalable computing simpler, faster, and more effective. The group is currently engaged in developing tools to simplify parallel computing, such as the Cilk multithreaded programming platform and its related tools for programmer productivity. In addition, we are developing parallel applications, focusing in particular on irregular problems such as graph applications. The group also studies adaptive computing, race detection, and cache-oblivious algorithms.