Analyzing the GoldMine Tool

GoldMine is an automatic assertion generation tool which uses data mining and static analysis techniques on the register-transfer level (RTL) design. The assertions generated from the RTL design are processed through the feedback mechanism after passing the formal verifier to improve their quality. GoldMine minimizes the human effort in this process by ranking the generated assertions based on the evaluation of their coverage. The purpose of this research is to evaluate the quality of the highly-ranked assertions generated by GoldMine on the data intensive Calc1 design. The objective of this research is to analyze the various data mining algorithms used in GoldMine, and to propose new functionalities that will make the tool more preferable for industrial use. The results of this research show that GoldMine is able to create complex and sophisticated assertions on combinational and sequential modules.U of I OnlyUndergraduate senior thesis not recommended for publication by adviso