Abstract

Many embedded signal-processing applications require the computing performance of multiprocessors and have varying power constraints imposed by their operating environment. These performance and power constraints can vary dynamically as situations and environments change. High-performance applications often require multiprocessor hardware to achieve sufficient processing performance. However, the need for high performance does not preclude the relevance of power management, either to extend battery life or to mitigate heat dissipation problems. The Power Aware Multiprocessor Architecture (PAMA) project has developed a power-aware multiprocessor architecture and has investigated the application of power management techniques to a space-based remote sensing application. The PAMA project built three generations of prototypes and demonstrated significant power and energy savings …

Date

January 1, 1970

Authors

Stephen Crago