AIGROW: A Feedback-Driven Test Generation Framework for Hardware Model Checkers

This research abstract introduces an effective and efficient approach to automatically generate high-quality hardware model checker benchmarks. The key contribution of this work is to model the input format of hardware model checkers using a tree-based structure named ARTree and build an effective feedback-driven test generation framework based on ARTree named AIGROW. The evaluation shows that AIGROW generates very small but high-quality benchmarks for coverage-oriented and performance-oriented testing and outperforms the existing generation-based testing tools.