Two Dimensional Dynamic Multigrained Reconfigurable Hardware

Abstract

Partial dynamic reconfigurable (PDR) systems designed with state-of-the-art tool chains, like the Early Access Partial Reconfiguration (EAPR) Flow from Xilinx, does not exploit the full flexibility and all features which a state of the art FPGA chip offers. For example the utilized chip area and the position of a region which can be reconfigured dynamically is traditionally specified during design-time. Thereby the shape and the size of the reconfigurable area are set by the size of the largest module to be reconfigured. The consequence is that if a smaller module is placed on that region, chip area stays unused as so called black silicon. This drawback is only one example for the limitation of development tools of reconfigurable hardware architectures. In this book section, a new approach for exploiting the capability of reconfigurable hardware architectures is presented. It allows exploiting the reconfigurable architectures more efficient than other solutions introduced before. This is achieved through a novel concept of using micro blocks for the communication infrastructure as well as for the functional elements on the FPGA. The granularity and the online versus offline tradeoff for the usage of the micro blocks for building up more complex structures on the FPGA will be presented in this chapter.