Risk Propagation Based Vector Profiling for High Coverage Dynamic IR-Drop Analysis

• Published in: 2023 IEEE/ACM International Conference on Computer Aided Design (ICCAD) pp. 1 - 8
• Main Authors: Wen, Yihan, Li, Juan, Wang, Xiaoyi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 28-10-2023
• Subjects: Chip scale packaging; Design automation; Dynamic IR-drop analysis; Estimation; IEEE Regions; Scalability; Silicon; Vector profiling; Vector-based IR-drop analysis
• ISSN: 1558-2434
• DOI: 10.1109/ICCAD57390.2023.10323636

Vector-based dynamic IR-drop analysis is a crucial aspect for enhancing yield in chip fabrication since it provides accurate IR-drop simulation with real waveform. To evaluate waveforms with a large duration from numerous working scenarios, vector profiling is widely used to increase scalability. In real cases, only a few windows selected by vector profiling are assessed by dynamic IR-drop analysis, rather than the whole waveform. Therefore, the coverage of vector profiling methods becomes a major concern, especially in EUV process node. The IR-drop locality effect on multi-pattern layers makes traditional vector profiling methods less robust. The real worst-case waveform window which may lead to silicon failure is frequently missed, which ultimately impacts the coverage of profiling. This paper proposes a novel risk propagation-based vector profiling method that achieves better estimation of IR-drop risk by considering the locality through examining not only the self-power-induced IR-drop but also the drop propagated from surrounding regions. The experimental results have shown that the proposed vector profiling achieved 4.3 times greater probability of covering the worst IR-drop window compared to traditional profiling. The proposed profiling also discovered additional IR-drop risky regions which were missed by traditional profiling.