Verrific
This demo shows how Verrific can accelerate your verification outcomes: using reinforcement learning (RL), it closes coverage faster by learning the search space and discovers cousin bugs by learning instruction patterns that trigger each hazard.
Overview
- Coverage: On the OpenTitan Ibex core, RL finds hard-to-reach bins 2x faster than constrained random generation.
- Bugs: RL find more bugs than traditional methods and can target specific bug types to find cousin bugs.
Try it
You can explore the experiment setups yourself - send us an email at [email protected] and we'll get in touch.