ABSTRACT
This chapter describes the principles and methods for the integration of segment-specific virtual components and the complete application-specific processing circuits in spatial and temporal domains. In contrast to conventional programmable computers with fixed architecture, reconfigurable computing systems (RCSs) can change the set and functionality of system components as well as links between components per time. Thus, flexibility in system architecture can provide runtime adaptation to variations of the workload and/or external conditions. The chapter focuses on the main concepts of architecture organizations that already have or may have practical applicability. In general, the RCS architecture should consist of two conceptually different parts: The part oriented on processing of algorithmically intensive segments of a task; and The part oriented on processing of computationally intensive segments of a task. The RCS architecture in its general form looks like a combination of modern general-purpose instruction-based computing systems and a set of programmable logic devices.
