Multicore Chip Design and Architecture (MCDA)

As Moore’s law scaling runs its course, researchers in industry and academia must explore new means by which to ensure continued technological advances in computing.  CMOS scaling is increasingly limited by the realities imposed by physics, making architectural innovations the most likely means by which to achieve increased computational performance. Multicore-based systems that modularly integrate multiple, heterogeneous processor cores on a single chip promise computational performance enhancements for both high and low end computing platforms (e.g., from petaflop machines to embedded systems). 

Novel research is needed on design, fabrication, architecture and programmability of homogeneous as well as heterogeneous multicore systems that will address virtually all hardware and software aspects of computing system design. NSF and the Semiconductor Research Corporation (SRC) recognize this need, and have agreed to embark on this new collaborative research program to address compelling research challenges in multicore-based systems that are of paramount importance to industry, academia and society at large.