AgenticPD: A Stage-Aware Agentic Framework for Physical Design QoR Optimization
2026-07-06 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors introduce AgenticPD, a new way to improve physical chip design that breaks down the process into stages. Instead of redoing the whole design every time they test a change, their system focuses on individual stages with specialized agents and saves progress to avoid repeating work. A Judge Agent guides the search for better designs by choosing which changes to explore. This method helps them get good results in timing without making power use or chip size worse.
physical designquality-of-resultsEDA flowcheckpointingagent-based optimizationpost-route signofftiming optimizationchip design stagessearch space navigation
Authors
Shuo Ren, Zijin Cheng, Yaohui Han, Libo Shen, Leilei Jin, Wanting Tian, Rongliang Fu, Chao Wang, Bei Yu, Tsung-Yi Ho
Abstract
Physical design quality-of-results~(QoR) optimization is hard and expensive. Choices made at one stage can help or hurt later stages. Each evaluation requires a costly EDA run through the full flow. While existing methods still treat optimization as flat parameter tuning or a LLM-based script generation task, we present AgenticPD, a stage-aware agentic framework for physical design QoR optimization. Instead of re-running the full flow after every trial, AgenticPD is organized around the stage boundaries of the physical design flow, where a Judge Agent navigates the search and stage-specialized agents make local decisions within their own stage using stage-local tools. Additionally, the agent harness in AgenticPD provides structured observations, execution history, and agent context management. As a result, the system can branch from prior intermediate states and reuse checkpoints to continue the optimization procedure, and every candidate is evaluated at the post-route signoff. Across these baselines, AgenticPD achieves strong post-route timing while remaining competitive in power and area.