With the emergence of the chip-level multiprocessor (CMP) comes the promise of integrating enormous computing power in a single chip, thereby enabling parallel computing in all types of platforms including handheld computers and desktop machines. Nonetheless, there are significant barriers and challenges to exploiting the power of CMPs. Very simply we must rethink every aspect of these systems—the underlying architecture(s) of the individual processing elements, the on-chip and off-chip memory architectures, the on-chip and off-chip communications structures and mechanisms, the operating system, the software architecture of applications, and the programming language paradigms and support required.

The Consortium for Adaptive CMP Compilation and Optimization (CACAO) is a group of researchers from Pennsylvania State University, University of Pittsburgh, and University of Virginia that is investigating all aspects of exploiting the power offered by CMPs. We invite you to visit our website and review our ongoing and past projects.