A multi-node dynamic reconfigurable computing system with distributed reconfiguration controller

Abstract

This paper reports the implementation of a multi-node dynamically reconfigurable computing system. The system is based on a scalable dynamic reconfigurable computing node which consists of three resource layers. Scalability of the system is introduced on some of those layers in order to deal with context synchronization aspects. This approach leads to a multi-node reconfigurable computing architecture with a distributed reconfiguration controller, which helps in hiding the reconfiguration and synchronization cost partially.