Energy-Efficient Flexibility for Processors
Embedded computer systems have become an inseparable part of our lives in the form of smart phones and watches, household appliances, toys, and cars. Processors play a central part in these systems, as they can be programmed to perform complex tasks. Internet-of-Things (IoT) consist of these systems connected via different networks and the internet.
“IoT is expected to cause a large increase in the number of embedded systems in the near future, increasing the proportion that they consume of global energy production. Typically, these devices are battery powered, making low energy consumption crucial. At the same time, the rate of improvement in the energy-efficiency of systems has declined,” says Multanen.
Although fixed-function accelerators (FFAs) result in the best energy-efficiency, flexibility and reusability of devices can be improved by including programmability.
“Programmability allows updating the system’s functionalities. For example, if new tasks to be performed are identified during device usage, they can be added via programmability."
However, programmability comes with a cost. Processor execution is controlled by computer programs, which consist of instructions. These are stored in memories that can constitute the majority of a system’s energy consumption. In his doctoral dissertation Multanen developed new techniques to improve the energy-efficiency of processor instruction delivery mechanisms.
“Software control can be used to improve the energy-efficiency of delivering instructions to processing elements. Instruction compression and utilizing unconventional emerging memory technologies for instructions offer interesting possibilities for energy savings,” says Multanen.
The doctoral dissertation of M.Sc. (Tech.) Joonas Multanen in the field of Computer Engineering titled Energy-Efficient Instruction Streams for Embedded Processors will be publicly examined at the Faculty of Information Technology and Communication Sciences of Tampere University at 12 o'clock on Friday 26 November, 2021. The venue is Tietotalo building TB109, Korkeakoulunkatu 1. Associate Professor Magnus Själander from Norwegian University of Science and Technology will be the opponent while Associate Professor Pekka Jääskeläinen from the Faculty of Information Technology and Communication Sciences of Tampere University will act as the custos.
The event can also be followed via remote connection (Zoom).
The dissertation is available online at http://urn.fi/URN:ISBN:978-952-03-2193-2
Photo: Heidi Korvela