System Services for Reconfigurable Hardware Acceleration in Mobile Devices

摘要: FPGAs have been deployed to provide custom hardware acceleration for applications such as networking, vision, and machine learning in embedded devices. A mobile device serves increasingly diverse computing under constrained resources, hence, making FPGA-based more attractive achieve high performance, yet, energy efficiency. However, due lack of systematic interface between application software framework FPGA design flow, developers avoid using accelerators developing Apps applications. In this paper, we introduce a that integrates the Android operating system with programmable accelerator flow. We present an service platform manages mapping compute-intensive kernels on FPGA. The proposed enables multiple access accelerators, simultaneously. OS porting Xilinx Zynq SoC, equipped dual-core ARM Cortex-A9 processors logic. experimental results show performance gain case studies canny edge detection algorithm, digit recognition neural network, which are used by Apps.