Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note3 min read

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

Source

arXiv

arxiv.org/abs/2608.16733v1

Read the source

Source

  • Source name: arXiv
  • Source kind: paper
  • Authors / organization: Haixu Liu, Lei Zhou, Yuhao Ren, Yumao Wu, Zhiang Wang
  • Venue: arXiv: cs.AR
  • Published: 2026-08-17
  • Collected: 2026-08-18
  • URL: http://arxiv.org/abs/2608.16733v1
  • PDF: https://arxiv.org/pdf/2608.16733v1
  • DOI: UNKNOWN
  • Registry group: academic_apis_and_preprints
  • Evidence role: canonical_academic_index
  • Language: en
  • Country / region: UNKNOWN

Plain-Language Brief

A new framework called GoalEvolve is presented that aims to improve physical design algorithms for electronic circuits. The framework uses a goal-driven approach to identify and address unmet requirements in the optimization flow, resulting in better final results compared to default OpenROAD. It also shows improvements over commercial tools on certain designs.

What the Document Says

  • The paper presents GoalEvolve, a framework for evolving physical design algorithms.
  • GoalEvolve focuses on multi-objective quality of results (QoR) target regions in tightly coupled, multi-stage optimization flows.
  • It identifies unmet requirements and locates responsible stages using stage-resolved checkpoint evidence and an LLM-based Teacher.
  • Parallel Student agents implement and validate hypotheses through full-flow evaluation.
  • GoalEvolve improves post-route TNS by an average of 30.67%, reduces leakage by 21.18%, and dynamic power by 9.42% compared to default OpenROAD on eight ASAP7 designs.
  • Relative to commercial-tool goals, it closes 62.20% of the normalized power gap on power-dominant designs, surpasses TNS goals on 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%, reduces leakage by 12.38%, and dynamic power by 0.76%.

Key Concepts Mentioned

  • physical design algorithms
  • GoalEvolve framework
  • multi-stage optimization flows
  • quality of results (QoR)
  • unmet requirements
  • stage-resolved checkpoint evidence
  • LLM-based Teacher
  • Student agents
  • post-route TNS
  • leakage
  • dynamic power
  • ASAP7 designs
  • commercial tools
  • timing-dominant designs
  • joint designs
  • Codex goal mode

Claims Made by the Document

  • GoalEvolve improves post-route TNS, reduces leakage, and dynamic power compared to default OpenROAD on ASAP7 designs.
  • Relative to commercial-tool goals, GoalEvolve shows improvements on certain designs.
  • Under matched budgets, GoalEvolve further improves TNS, reduces leakage, and dynamic power compared to Codex goal mode.

Evidence Used

  • Average improvement of 30.67% in post-route TNS, 21.18% reduction in leakage, and 9.42% reduction in dynamic power on eight ASAP7 designs.
  • Closing 62.20% of the normalized power gap on power-dominant designs, surpassing TNS goals on timing-dominant designs, and closing 32.48% of the equal-weight timing-power gap on joint designs compared to commercial-tool goals.
  • Improving TNS by 26.46%, reducing leakage by 12.38%, and dynamic power by 0.76% across all three designs evaluated against Codex goal mode under matched budgets.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

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.