用自注意力U-Net提升3D芯片热仿真精度与速度
Self-Attention to Operator Learning-based 3D-IC Thermal Simulation
- 结合自注意力与U-Net结构，捕捉长程依赖和高频局部特征
- 相比传统FEM方法提速842倍，热预测精度达当前最优
- 支持低质量数据微调，减少对高保真数据依赖，适合芯片设计迭代
3D集成电路因功率密度升高，热管理日益严峻。传统基于偏微分方程的求解方法虽精确但计算缓慢，难以满足迭代设计需求。现有机器学习方法如FNO虽加速显著，却存在高频信息丢失及对高保真数据强依赖问题。本文提出Self-Attention U-Net Fourier Neural Operator（SAU-FNO），融合自注意力机制与U-Net结构，有效建模长程依赖关系并保留局部高频特征。通过迁移学习对低保真数据进行微调，大幅降低对高质量训练数据的需求，加快模型训练。实验表明，SAU-FNO在热场预测上达到当前最优精度，相较传统FEM方法实现842倍加速，为先进3D IC热仿真提供高效工具。
原文摘要 · Abstract (English)
Thermal management in 3D ICs is increasingly challenging due to higher power densities. Traditional PDE-solving-based methods, while accurate, are too slow for iterative design. Machine learning approaches like FNO provide faster alternatives but suffer from high-frequency information loss and high-fidelity data dependency. We introduce Self-Attention U-Net Fourier Neural Operator (SAU-FNO), a novel framework combining self-attention and U-Net with FNO to capture long-range dependencies and model local high-frequency features effectively. Transfer learning is employed to fine-tune low-fidelity data, minimizing the need for extensive high-fidelity datasets and speeding up training. Experiments demonstrate that SAU-FNO achieves state-of-the-art thermal prediction accuracy and provides an 842x speedup over traditional FEM methods, making it an efficient tool for advanced 3D IC thermal simulations.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。