评估 AI 智能体做工业级芯片布局：框架与实证研究
Evaluating AI Agents for Industrial-Level Chip Floorplanning: A Framework and Empirical Study
AI 导读
研究者提出一套评估框架，将 AI 智能体提交的芯片布局连接到指定的下游实现流程，检查几何有效性并测量拥塞、时序等物理质量。在一个 99 个宏单元占据约 71% 固定 L 形布局区域的工业设计上，专用系统 Nova-SI 在 0-100 综合质量评分中取得 57.97 的均值，最强的通用智能体 GPT-6 Astra 为 31.17，两者差距得到原始拥塞和时序测量支持。
正文 · 原文
Abstract:Floorplanning is an early step in turning a logical circuit into a physical chip: it positions large circuit modules and constrains the space available for subsequent logic placement and wiring. These decisions affect congestion and timing, but their physical consequences emerge through downstream implementation. For AI agents, the task requires translating engineering constraints into spatial decisions whose quality cannot be judged by successful tool execution alone. We present an evaluation framework that connects submitted layouts to a prescribed downstream implementation flow, checks geometric validity, measures physical quality, and reports resource use. On one industrial design whose 99 macros occupy approximately 71\% of a fixed L-shaped layout region, we compare our specialized system Nova-SI with general-purpose engineering agents and agents restricted to file access, obtaining complete physical measurements for 39 layout outputs. Under our 0--100 composite quality score, Nova-SI achieves a mean of 57.97 without online feedback from the prescribed power-grid and placement evaluation, compared with 31.17 for the strongest measured general-purpose configuration, GPT-6 Astra. The difference is supported by raw congestion and timing measurements and persists across scoring-sensitivity checks. Layout-level analysis further shows that similar global utilization can conceal substantially different local congestion. Together, the protocol and case study assess agents through the physical consequences of their delivered layouts and provide evidence about system performance on this industrial instance.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.12048 [cs.AI] |
| (or arXiv:2611.12048v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.12048 arXiv-issued DOI via DataCite |
Submission history
From: Linyang Li [view email]
[v1]
Mon, 9 Nov 2026 08:25:58 UTC (5,276 KB)
[v2]
Thu, 12 Nov 2026 02:42:30 UTC (963 KB)
来源：arXiv cs.AI · arxiv.org