AI 中文总结
针对布线树优化问题，本研究开发智能体框架发现SALT的结构局限，提出FABO算法，在129万个ICCAD15网络上实现导线长度降低，FABO-FAST可高效优化多数网络。
AI 中文摘要
布线树的拓扑结构决定了多引脚网络如何分支并共享物理导线，直接影响导线长度、拥塞度、电容和延迟。我们研究早期布线阶段的核心问题：在约束每个汇点的根到汇点路径拉伸度的同时最小化导线长度。SALT是该问题的当前最优构造算法。我们探究语言模型引导的搜索是否能发现优于SALT的构造算法。为使该搜索可靠，我们开发了结合并行探索与独立校验的智能体框架。将该框架应用于SALT，我们发现了其结构局限：SALT每次仅修复一个汇点路径，因此从未联合决定共享根侧导线的路径应在何处拆分。这种汇点局部选择可能导致路径过早拆分并产生冗余导线。该发现催生了流感知断点优化算法（FABO），其在保留每个汇点拉伸预算的同时，联合优化共享导线的根到汇点路径的断点。在129万个ICCAD15网络及SALT的20点拉伸容差调度下，FABO在所有设置下均降低了FLUTE归一化导线长度，整体峰值降低0.83%，引脚数不少于30的网络峰值降低2.66%。FABO-FAST的运行时间为SALT的1.3倍，可识别并优化FABO能大幅降低导线长度的多数网络。代码可在指定URL获取。
英文摘要
The topology of a routing tree determines how a multi-pin net branches and shares physical wire, directly affecting wirelength, congestion, capacitance, and delay. We study a central early-stage routing problem: minimizing wirelength while bounding the root-to-sink path stretch for every sink. SALT is the state-of-the-art constructive algorithm for this problem. We ask whether language-model-guided search can discover a constructive algorithm that improves on SALT. To make this search reliable, we develop an agent framework that combines parallel exploration with independent checking. Applied to SALT, the framework discovers a structural limitation: SALT repairs one sink path at a time and therefore never jointly decides where paths sharing root-side wire should split. This sink-local choice can split the paths too early and duplicate wire. This discovery leads to Flow-Aware Breakpoint Optimization (FABO), which jointly optimizes breakpoints across root-to-sink paths that share wire while preserving every sink's stretch budget. Across 1.29 million ICCAD15 nets and SALT's 20-point stretch-tolerance schedule, FABO reduces average FLUTE-normalized wirelength at every setting, with peak same reductions of 0.83% overall and 2.66% for nets with at least 30 pins. With 1.3x SALT's runtime, FABO-FAST identifies and optimizes most nets for which FABO provides a substantial wirelength reduction. Code is available at https://github.com/DevinShang/routing-FABO.
Comments13 pages
