Quality Evaluation of Digital Soft IP Core for FPGA System

Field Programmable Gate Array (FPGA) system is widely used in deep learning application and cloud system for acceleration. Quality and reliability of IP block is essential to the successful development of today's complex hardware acceleration design. In this paper, we discuss the important issues of quality and reliability of digital soft IP, and propose a qualification measurement system that can evaluate IP quality. The metric system consists of eight categories to characterize IP quality. Based on this indicator system, we propose an IP qualification frameworks. As a case study, the proposed framework has been used to qualify three open source IP cores. The experimental results show that the proposed metric system and qualification framework can not only quantify the quality of IP, but also speed up the certification process and improve the accuracy of the certification results.