Heterogeneous Reconfigurable Systems-on-Chip (HRSoC) contain as their name suggests, heterogeneous processing elements in a single chip. Namely, several processors, hardware accelerators as well as communication networks between all these components. In order to leverage the programming complexity of this kind of platform, applications are described with software threads, running on processors, and hardware threads, running on FPGA partitions. Combining techniques such as dynamic and partial reconfiguration and partial readback with the knowledge of the bitstream structure offer the ability to target several partitions using a unique configuration file. Such a feature permits to save critical memory resources. In this article, we propose to tackle the issue of designing fully independent partitions, and especially to avoid the routing conflicts which can occur when using the standard Xilinx FPGA design flow. To achieve the relocation process successfully, we propose a new design flow dedicated to the module relocation, using the standard tools and based on the Isolation Design Flow (IDF), a special flow provided by Xilinx for secure FPGA applications.

This paper addresses the problem of image processing algorithms implementation onto dynamically and reconfigurable architectures. Today, these Systems-on-Chip (SoC), offer the possibility to implement several heterogeneous processing elements in a single chip. It means several processors, few hardware accelerators as well as communication mediums between all these components. Applications for this kind of platform are described with software threads, running on processors, and specific hardware accelerators, running on hardware partitions. This paper focuses on the complex problem of communication management between software and hardware actors for dataflow oriented processing, and proposes solutions to leverage this issue.

No abstract