GoalEvolve: From Handcrafted Algorithm Priors to Goal-Driven Evolution of Physical Design Algorithms
2026-08-17 • Hardware Architecture
Hardware ArchitectureArtificial Intelligence
AI summaryⓘ
The authors propose GoalEvolve, a new method to improve physical design algorithms by focusing on how well the entire design process meets final goals, rather than just improving individual steps. Their system identifies which design stages are causing problems and uses AI agents to test and refine changes based on those issues. When tested on several chip designs, GoalEvolve significantly improved timing and power performance compared to existing tools. This approach helps guide design improvements by directly linking changes to overall quality instead of local fixes.
physical designmulti-stage optimizationquality of results (QoR)timing closurepower optimizationLLM (Large Language Model)OpenROADASAP7post-route analysisoptimization debt
Authors
Haixu Liu, Lei Zhou, Yuhao Ren, Yumao Wu, Zhiang Wang
Abstract
Physical design algorithms operate within tightly coupled, multi-stage optimization flows, where stage-local gains may vanish or induce downstream degradation. Existing program-evolution frameworks often rely on stage-local objectives or undifferentiated multi-metric feedback, which neither guarantee better final results nor identify which unmet requirement should guide the next iteration. We present GoalEvolve, a goal-driven framework that makes physical design algorithm evolution accountable for the final quality of results (QoR) of the complete flow. Given a multi-objective QoR target region, GoalEvolve converts unmet requirements into normalized target gaps, identifies the dominant bottleneck, and uses stage-resolved checkpoint evidence to locate the responsible stage. An LLM-based Teacher then narrows the search to a relevant algorithmic decision and source region, while parallel Student agents implement and validate hypotheses through full-flow evaluation. Local effects, optimization debt, and downstream retention are retained as mechanism evidence for subsequent evolution. Across eight ASAP7 designs, GoalEvolve improves post-route TNS by 30.67% on average and reduces leakage and dynamic power by 21.18% and 9.42% versus default OpenROAD. Relative to commercial-tool goals, it closes 62.20% of the normalized power gap on power-dominant designs, surpasses the TNS goals on both timing-dominant designs, and closes 32.48% of the equal-weight timing-power gap on joint designs. Across all three designs evaluated against Codex goal mode under matched budgets, GoalEvolve further improves TNS by 26.46% while reducing leakage and dynamic power by 12.38% and 0.76%, respectively.