In this case, the emergent property sought to establish is system- level fault tolerance, the inspiration from biology are social insects (ant colonies), and the hardware system is a many-core computing architecture where application tasks and data need to be allocated transferred and organised. The model of processing elements com- municating amongst each other via a network on chip (NoC) provides a conceptual link with many scalable biological models. Based on this, a self-optimising and adaptive, yet fundamentally scalable, design approach for many-core systems based on the emer- gent behaviours of social-insect colonies are developed. Experiments aim to capture the relevant decision processes made by each member of the colony to exhibit such high-level behaviours and embed these decision engines within the routers of the many-core system. Results with the bespoke 128-core Centurion platform suggest that there is potential for the social insect model as a distributed, embedded intelligence within a many-core system and with the relevant knobs and monitors, such as clock frequency and temperature, to close the loop for emergent autonomous adaptation and fault tolerance.
|Published - 17 Mar 2019
|Second International TRANSIT workshop on Cross-disciplinary Research: Emergence - University of York, York, United Kingdom
Duration: 27 Mar 2019 → 28 Mar 2019
|Second International TRANSIT workshop on Cross-disciplinary Research: Emergence
|27/03/19 → 28/03/19
- Many core
- Social Insect Inspired Systems