• Protocol & Unit Selection: Identify a unit and protocol lacking coverage scenarios. Extract partially covered items from protocol coverage reports.
  • Coverage Analysis & Sequence Retrieval: Browse all coverage reports for the selected protocol across benches. For each item, locate benches where the item is covered. Retrieve traffic sequences using coverage macro traces and monitor logs. Analyze reproducibility of sequences on the target bench considering hardware constraints.
  • Scenario Generation & ML Integration: Use scripting and ML tools to generate new sequences for items consistently under-covered. Train or fine-tune ML models on existing C++ verification code to generate methods that match the environment’s structure. Evaluate the compatibility of ML-generated C++ methods with the existing verification framework (e.g., drivers, monitors, BFMs). Implement and validate these methods in simulation.
  • Execution & Validation: Run generated sequences and verify coverage improvements. Use coverage reports to confirm effectiveness.
  • Statistical Analysis & Optimization: Identify items requiring short vs. long sequences. Evaluate sequence efficiency (number of items hit per sequence). Compile metrics on coverage gains, machine/memory usage, and reproducibility.
  • Final Presentation: Summarize findings, methodology, and results. Present statistical insights and recommendations for future coverage improvements.