Researchers have developed DRC-Aid, an agentic framework that automates the correction of design rule violations (DRVs) in chip layouts. The system uses a Large Language Model (LLM) to select geometric edits from a predefined menu, guided by a deterministic rule engine and a memory bank to prevent redundant exploration. Evaluated on FreePDK45 layouts, DRC-Aid successfully repaired DRVs in approximately 92.5% of cases, achieving a near-complete violation reduction. This LLM-driven approach significantly outperformed random and heuristic methods in repair accuracy. AI
IMPACT Automates a critical step in chip design, potentially accelerating development cycles and improving yield.
RANK_REASON Academic paper detailing a novel framework for chip design automation. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
