arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2312.12784cs.LG
收藏

基于图神经网络的快速单元库表征用于设计技术协同优化

Fast Cell Library Characterization for Design Technology Co-Optimization Based on Graph Neural Networks

Shanghai Jiao Tong University(上海交通大学)

Tianliang Ma, Guangxi Fan, Zhihui Deng, Xuguang Sun, Kainlu Low, Leilai Shao

AI总结:

本文提出基于图神经网络的机器学习模型,实现快速准确的单元库表征,在512个未见角点及百万级数据上预测误差≤0.95%,加速100倍,并进一步提出驱动强度插值方法提升中小规模设计PPA约1-3%。

摘要:

设计技术协同优化(DTCO)在实现先进半导体工艺开发的最优功耗、性能和面积(PPA)方面起着关键作用。单元库表征是DTCO流程中不可或缺的一环,但传统方法耗时且成本高昂。为克服这些挑战,我们提出了一种基于图神经网络(GNN)的机器学习模型,用于快速且准确地完成单元库表征。我们的模型融合了单元结构,并在多种工艺-电压-温度(PVT)角点和工艺参数下展现出高预测精度。使用512个未见过的工艺角点和超过一百万个测试数据点进行验证,结果表明,对于33种类型的单元,该模型在延迟、功耗和输入引脚电容上的预测均准确,平均绝对百分比误差(MAPE)≤0.95%,相比SPICE仿真实现了100倍的加速。此外,我们利用基于GNN的模型在未见过的角点上获得的预测,研究了最差负裕量(WNS)、泄漏功耗和动态功耗等系统级指标。与黄金参考值相比,我们的模型实现了精确预测,WNS的绝对误差≤3.0 ps,泄漏功耗的百分比误差≤0.60%,动态功耗的百分比误差≤0.99%。借助所开发的模型,我们进一步提出了一种细粒度驱动强度插值方法,以增强中小规模设计的PPA,实现了约1-3%的改进。

英文摘要:

Design technology co-optimization (DTCO) plays a critical role in achieving optimal power, performance, and area (PPA) for advanced semiconductor process development. Cell library characterization is essential in DTCO flow, but traditional methods are time-consuming and costly. To overcome these challenges, we propose a graph neural network (GNN)-based machine learning model for rapid and accurate cell library characterization. Our model incorporates cell structures and demonstrates high prediction accuracy across various process-voltage-temperature (PVT) corners and technology parameters. Validation with 512 unseen technology corners and over one million test data points shows accurate predictions of delay, power, and input pin capacitance for 33 types of cells, with a mean absolute percentage error (MAPE) $\le$ 0.95% and a speed-up of 100X compared with SPICE simulations. Additionally, we investigate system-level metrics such as worst negative slack (WNS), leakage power, and dynamic power using predictions obtained from the GNN-based model on unseen corners. Our model achieves precise predictions, with absolute error $\le$3.0 ps for WNS, percentage errors $\le$0.60% for leakage power, and $\le$0.99% for dynamic power, when compared to golden reference. With the developed model, we further proposed a fine-grained drive strength interpolation methodology to enhance PPA for small-to-medium-scale designs, resulting in an approximate 1-3% improvement.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑