Back to AI News
arxiv

Verify the Verifier: Specification-Grounded LLM Cross-Checking for Bugs That Escape UVM Scoreboards

The paper introduces Verify the Verifier, a framework whose LLM-based Anomaly Cross-Checker examines traces from passing UVM regressions against the specification to find scoreboard escapes. In an evaluation inserting 564 bugs into four OpenTitan IPs, 68 were identified as candidate scoreboard escapes, and the framework confirmed 41 as escapes using assertions that fail on the buggy design and pass on the original. The work targets specification violations that survive mature testbenches because the scoreboard only checks agreement with a reference model.

AI Brief

It presents a concrete LLM-based verification method with quantified OpenTitan results for catching specification violations that conventional UVM scoreboards miss.

arXiv:2611.26588v1 Announce Type: cross Abstract: A passing UVM regression shows the design agrees with the reference model on what the scoreboard checks, not that it meets the specification. Bugs that pass every test yet violate the specification, termed scoreboard escapes, survive mature testbenches. We present Verify the Verifier, to our knowledge the first framework whose LLM-based Anomaly Cross-Checker examines a passing regression's traces against the specification. Of 564 bugs inserted into four OpenTitan IPs, 68 are candidate scoreboard escapes: non-equivalent, exercised by the tests, yet passing. The framework confirmed 41 as scoreboard escapes with assertions failing on the buggy design and proven on the original.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!