GoalEvolve: From Handcrafted Algorithm Priors to Goal-Driven Evolution of Physical Design Algorithms

2026-08-17

AI Summary

This study addresses the challenges of local gain-induced downstream degradation and lack of result accountability in multi-stage physical design optimization by proposing an objective-driven algorithm evolution framework. The method normalizes requirement gaps to identify bottlenecks and employs a Large Language Model teacher to prune the search space, while parallel student agents perform full-flow verification with evidence tracing to enable accountable, end-to-end QoR-oriented optimization. Evaluated on ASAP7 benchmarks, the approach achieves an average 30.67% improvement in Total Negative Slack (TNS), alongside reductions of 21.18% in leakage power and 9.42% in dynamic power. These results demonstrate significant performance gains over Codex-based baselines and competitiveness with commercial EDA tools, validating the effectiveness of integrating LLM-guided search with rigorous verification for responsible automated physical design.

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.

Problem

• Physical Design Algorithms
• Program Evolution
• Multi-stage Optimization
• Quality of Results
• Stage-local Objectives

Innovation

• Goal-Driven Evolution
• Physical Design Algorithms
• LLM-based Teacher-Student
• Full-Flow QoR Accountability
• Bottleneck Localization