arXiv：cs.AI· Shu Gu, Matthew Youngman, Yuqin Zhao, Cristian Sestito, Ali Siddique, Ramesh Krishnamurthy, Danial Chitnis, Themis Prodromakis·· 16 小时前
LATCH：面向布线后 FPGA 优化的 LLM 辅助时序收敛框架
LATCH: LLM-Assisted Timing Closure Harness for Post-Route FPGA Optimisation
AI 导读
研究提出 LATCH 双层框架，将 LLM 编排与确定性 FPGA 实现和评估分离。在 16 个设计和 7 个 LLM 上比较三种智能体控制粒度，功能通过率分别为 98.2%、71.4% 和 17.9%，直接 EDA 操作在通过时 Fmax 中位提升 3.4%，略高于配方式的 3.2%。
正文
Abstract:Post-route FPGA timing closure remains expert-dependent and computationally expensive. We present LATCH (LLM-Assisted Timing Closure Harness), a two-layer framework that separates large language model (LLM) orchestration from deterministic FPGA implementation and evaluation. Across sixteen designs and seven LLMs, a single-iteration study compares three agent control granularities - expert-written recipes, direct EDA operations and LLM-constructed procedures - to ask how much control an agent needs per attempt. Functional-pass rates are 98.2%, 71.4% and 17.9%, respectively; yet direct operations give slightly larger 3.4% median Fmax gains when passing, versus 3.2% for recipes. This exposes a trade-off between execution reliability and timing improvement. Informed by this trade-off, a multi-iteration study fixes the action tier at expert recipes to ask how repeated attempts are best organised, comparing a fitted deterministic heuristic with sequential LLM reasoning, branching exploration and knowledge base-guided exploration within 12 checkpoints per run. Branching raises mean maximum-frequency improvement from 10.5% (sequential) to 16.0%, and the knowledge base to 20.3%, improving 15 of 16 designs. Characterisation across 2,421 recipe attempts profiled their gain and validation rates, showing hard-block re-placement gives the largest cumulative mean gain at 41.4MHz.
| Comments: | 14 pages, 11 figures, 4 tables. Code: this https URL |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.16625 [cs.AI] |
| (or arXiv:2610.16625v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.16625 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Matthew Youngman [view email]
[v1]
Mon, 12 Oct 2026 11:44:40 UTC (14,436 KB)
来源：arXiv：cs.AI · arxiv.org