Electronic engineers and computer scientists are aiming to design and implement more energy-efficient and dependable ‘many-core’ processor embedded systems, with the launch of a new creative research project.
The five-year programme is being carried out by research team PriME (Power-efficient, Reliable, Many-core Embedded systems), involving the Universities of Southampton, Imperial College, Manchester and Newcastle, and is funded by £5.6m from the Engineering and Physical Sciences Research Council (EPSRC).
“While many-core processing is viewed as a way to improve the performance of computing systems, the energy consumption and reliability of these systems with hundreds or thousands of cores has yet to be fully understood,” says Professor Bashir Al-Hashimi of University of Southampton, and Director of PriME.
Many-core chip architecture is programmable hardware that is more specialised than for example multi-core architecture, which is more general-purpose. Many-core chips are considered to be of superior performance but on a narrower range of problems.
The researchers are hoping to achieve a step-change in design methods and cross-layer system optimisation, by reigning in energy consumption increases and making many-core systems more reliable.
Technology outcomes from the research are to be identified and translated for quick commercial exploitation, facilitated by the close collaboration of researchers and industry players from five companies, including ARM Ltd and Imagination of the UK.
According to Al-Hashimi, designing high-performance many-core embedded systems will help to create value and wealth in the future for the UK, therefore “suitable industrial and knowledge-transfer collaborators have been carefully identified that are best placed to influence and exploit the research.”
The UK electronics industry is the fifth largest in the world, worth some £23 billion a year. Further, it is considered a world-leader within embedded systems, due to the presence of companies such as ARM Ltd, Imagination, Microsoft Research, Freescale and Altera.
Many-core processing has applications in embedded, mobile, general-purpose and high-performance computing