Authors
Amir M. Rahmani; Pasi Liljeberg; Jose L. Ayala; Hannu Tenhunen; Alexander V. Veidenbaum
Journal Paper
https://doi.org/10.1016/j.jpdc.2016.04.013
Publisher URL
https://www.sciencedirect.com/
Publication date
March 2016
Recent trends in the microprocessor industry have important ramifications for the design of the next generation of high-performance as well as embedded parallel and network-based systems. By increasing number of cores, it is possible to improve the performance while keeping the power consumption at the bay. This trend has reached the deployment stage in parallel and network-based systems ranging from small ultramobile devices to large telecommunication servers. It is expected that the number of cores in these systems increases dramatically in the near future. For such systems, energy efficiency is one of the primary design constraints. The cessation of Dennard scaling and the dark silicon phenomenon have limited recent improvements in transistor speed and energy efficiency, resulting in slowed improvements in multi-core and many-core systems. Consequently, architectural innovation has become crucial to achieve performance and efficiency gains. New technologies that combine different types of cores or similar cores with different computation capabilities can result in a better match between the workload and the execution hardware improving overall system energy efficiency. In addition, multi-core and many-core systems need to be able to reconfigure themselves adaptively by monitoring their own condition and the surrounding environment in order to adapt themselves to different scenarios and performance-power requirements. Runtime monitoring becomes crucial in the near future parallel and distributed multicore systems due to increase in thermal issues as well as due to the need for various adaptive managements. This special issue addresses all aspects of energy-efficient computing in parallel and distributed multi-core and many-core systems.