Researchers have developed BugGen 2.0, an improved system for generating realistic bugs in Register-Transfer Level (RTL) code, which is crucial for testing hardware designs. This new version refines the process by using deterministic algorithms for structural analysis and target selection, while employing Large Language Models (LLMs) specifically for localized semantic rewriting. BugGen 2.0 significantly reduces computational costs, using 6.8x to 15.4x fewer tokens compared to its predecessor, and also decreases simulation overhead by an estimated 33.9% to 63.6% by intelligently selecting which bug variants require full simulation. AI
IMPACT This research could lead to more efficient and cost-effective testing of hardware designs by improving the scalability and accuracy of AI-driven bug synthesis.
RANK_REASON The cluster contains a research paper detailing a new method for generating synthetic bugs in hardware design code using LLMs. [lever_c_demoted from research: ic=1 ai=1.0]
- Advanced Encryption Standard
- BugGen 1.0
- BugGen 2.0
- FabScalar: Automating Superscalar Core Design
- Lampyridae
- Large Language Models
- OpenTitan
- RTL
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →