Energy Optimization for Many-Core Platforms

Manycore architectures are the most recent shift in multiprocessor design. This new design paradigm replaces large monolithic processors by thousands of simple processing elements on a single chip. With such a large number of cores, it promises significant throughput advantage over traditional multicore. Also, it enables better localized control of power consumption and thermal gradients. This is achieved by selective reduction of the core's supply voltage and operating frequency. Eventually, cores with the same voltage are grouped together, forming voltage islands, to reduce the design complexity. Building such voltage islands has to consider also speed and power differences among cores due to Process, Voltage, and Temperature (PVT) variations. The impact of such variations poses major obstacles as technology scales down. In this book, the author presents energy optimization techniques to reduce power consumption in such massively parallel architectures considering PVT. The wide gap between battery technology and power consumption makes energy reduction a vital step towards the wide-spread use of manycore architectures in the current era of mobile computing.

Autorentext

Sohaib Majzoub finished his PhD from the University of British Columbia, Vancouver, Canada. His research was a joint work between the System-on-Chip and Image Processing labs. He is currently assistant professor at the American University in Dubai, UAE. His main interest is performance and power optimization for Multiprocessor System-on-Chip.