An effective fault localization approach for Verilog based on enhanced contexts

Zhuo Zhang, Ya Li, Lei Xia, Jianxin Xue, Jiang Wu, Xiaoguang Mao

Abstract: The HDL research community recognizes the pivotal role of bug localization in the debugging workflow, offering substantial relief to developers. This paper introduces ContextHD, a robust bug localization method tailored for Verilog, leveraging enhanced contexts. ContextHD integrates static slicing techniques with advanced suspiciousness measurement methods to generate a prioritized list of statements based on their suspicious values, arranged in descending order. Our experimental evaluation involved comparing ContextHD with seven established state-of-the-art bug localization techniques. The results unequivocally demonstrate that ContextHD outperforms these methods significantly. Future endeavors include refining ContextHD to further enhance its accuracy.