Songyu Sun

Songyu Sun

Auto-generated profile, based on publicly available information

Songyu Sun

Zhejiang University, Hangzhou, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
5 September 2024 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
SSSongyu Sun
Songyu Sun
XDXiao Dong
Xiao Dong
QSQi Sun
Qi Sun
XYXunzhao Yin
Xunzhao Yin
QCQuan Chen
Quan Chen
CZCheng Zhuo
Cheng Zhuo
Emerging domains, such as artificial intelligence, 5G mobile, and automotive, are increasingly reliant on high-speed circuits for efficient processing, in which achieving high operating frequencies and data rates is crucial to enable productive data exchange and rapid responses. High-speed data as well as low noise margin in the high-speed serial links call for efficient models of drivers. In this article, we propose a fast machine-learning-driven macromodel for high-speed drivers, which can efficiently capture the nonlinear characteristics of drivers considering dynamic supply noise with low model complexity. A decoupling-superposition strategy is employed to effectively calculate the impact of power supply noise. Additionally, we introduce a piecewise-segmented method for macromodel solving to further enhance the speed of model utilization. Experimental results demonstrate that compared to HSPICE, the proposed macromodel achieves up to $50\times $ – $1200\times $ speedup while maintaining sufficient accuracy, even for signals with GHz data rate.

22 January 2025 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
XDXiao Dong
Xiao Dong
SSSongyu Sun
Songyu Sun
YJYangfan Jiang
Yangfan Jiang
Jingtong HuJingtong Hu
Jingtong Hu
DGDawei Gao
Dawei Gao
CZCheng Zhuo
Cheng Zhuo
Chiplet technology has recently emerged as a promising solution to improving chip performance through the modularization of complex designs and communication facilitated by high-speed inter-chiplet serial links. However, the increasing on-package routing density and data rates of these links introduce complex signal and power integrity challenges, surpassing those encountered in traditional large monolithic chips. Addressing these complexities with efficient analysis and design tools is crucial for maintaining design robustness. In this article, we propose SPIRAL+: a signal-power integrity co-analysis framework for high-speed inter-chiplet serial links validation. The framework employs machine learning to construct transmitter models and utilizes an impulse response extraction method for modeling the channel and receiver. It then performs signal-power integrity co-analysis through a novel double-edge response based method, leveraging the developed equivalent models. Additionally, an efficient machine learning model is crafted to accurately predict eye diagram metrics. The analysis provides valuable insights for design optimization. Experimental results show that SPIRAL+ achieves eye diagrams with a mean relative error of 0.07-7.47%, while realizing a speedup of 31-326× over traditional commercial tools.

19 May 2025 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
SSSongyu Sun
Songyu Sun
XDXiao Dong
Xiao Dong
YSYanliang Sha
Yanliang Sha
QSQi Sun
Qi Sun
QCQuan Chen
Quan Chen
CZCheng Zhuo
Cheng Zhuo
High-speed serial links are essential for low-latency, high-bandwidth communication in data-intensive systems. Signal integrity (SI) of transmitters (TXs) directly impacts transmission quality of the links, while TXs’ delay also introduces timing mismatches that degrade link integrity. In this article, we propose LiTformer, a Transformer-based model for efficient SI analysis of high-speed link TXs, featuring a nonsequential encoder and a multihead Transformer decoder to incorporate link parameters and capture long-range dependencies. By adopting a non-autoregressive approach, it enables parallel sequence prediction. We also introduce an ANN-based delay model for fast TX delay estimation. Considering link factors, including crosstalk in multiple-link systems, LiTformer enables accurate and fast long-sequence signal prediction at high data rates, achieving efficient SI analysis for TXs. Experimental results show LiTformer achieves $437\times $ – $996\times $ speedup in eye diagram prediction over SPICE, with mean errors of 0.15%–1.57%. It supports 4-bit signals at Gbps data rates for single-ended and differential TXs, including NRZ and PAM4 formats. The delay model predicts TX delay achieving a speedup of four orders of magnitude with errors of 0.86%–2.69%.


19 March 2026 in ACM Transactions on Design Automation of Electronic Systems
Verified
Journal Article
Open access
JHJingchao Hu
Jingchao Hu
YCYufei Chen
Yufei Chen
SSSongyu Sun
Songyu Sun
JSJianfei Song
Jianfei Song
LZLi Zhang
Li Zhang
CZCheng Zhuo
Cheng Zhuo
With escalating power integrity challenges in advanced technologies, acquiring accurate dynamic power supply noise through Dynamic Voltage Drop (DVD) analysis becomes increasingly demanding. As noise margins shrink, the use of Value Change Dump (VCD) files for precise DVD analysis is indispensable but computationally expensive. Furthermore, the substantial storage requirements of VCD files, which record digital waveforms from logical simulations, pose significant challenges. In this article, we propose a machine learning (ML)-assisted VCD processing framework to accelerate DVD analysis and improve data efficiency. Transitions recorded in VCD files are mapped to a Physical Design-Aware Circuit Hierarchy Tree (CHT) for efficient feature extraction. These features are leveraged by an XGBoost-based predictor to identify critical vector time windows within the VCD, significantly reducing simulation complexity. Additionally, Huffman encoding is applied to compress signal names, further optimizing storage utilization. Experimental results show that DVD analysis using our profiled VCD files achieves a speedup of approximately 3.53× with an error margin of only 3.89%.

9 September 2024
Verified
Conference Paper
Open access
YJYangfan Jiang
Yangfan Jiang
JSJianfei Song
Jianfei Song
XYXunzhao Yin
Xunzhao Yin
XDXiao Dong
Xiao Dong
SSSongyu Sun
Songyu Sun
CZCheng Zhuo
Cheng Zhuo
Power grid analysis is essential for integrated circuit design, especially as the scale extends to billions of nodes, making precise calculations across the entire spatial domain prohibitively expensive. Domain decomposition methods (DDM) facilitate block-wise parallel analysis of power distribution networks (PDN) through segmentation, yet they still face bottlenecks in computational load balance during parallel processing. Considering the vastness of PDN structures and the emphasis on verifying areas prone to significant IR drop (i.e., hotspot) rather than those less affected, our research introduces a machine learning (ML)-based method for detecting PDN hotspots. This method focuses on accurately identifying areas within sub-blocks where the hotspot intensity is high. We then propose a novel parallel approximation computing approach that integrates hotspot detection, enabling approximations in regions not marked as hotspots. This strategy optimizes computational resources and effectively accelerates analysis. Our experimental results demonstrate that the hotspot detection model achieves an average accuracy of 94.39%. The approximation strategy for non-hotspot sub-blocks results in 0.56-1.4% mean relative error, while achieving an average computational speedup of 3.21× over the CK-TSO parallel solver and 2.21× over the conventional DDM-based solver.