Continuous development and improvement of chip multi-processors (CMPs) is expected to allow the integration of a large number of processor cores on a multicore chip in the near future. Thus, it is extremely important to explore efficient designs for future CMPs to enable such integration to result in scalable performance. This exploratory project investigates heterogeneous CMP design alternatives that leverage from previous research on single core processors to achieve good utilization of core resources, high per core performance, good multicore scalability, and better core yield over CMP designs composed of homogeneous cores.
