YWYi Wang

Yi Wang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Yi Wang has a total of 65 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • A Counter-Based Addition Circuit Design for Stochastic Computing

    by Shaowei Wang, Zhengxian Wu, Yaohua Xu, Yuanyuan Zhang, Yi Wang, Yongqiang Zhang

    Stochastic computing (SC) encodes real values via probabilistic bitstreams, enabling complex arithmetic operations to be realized by simple logic gates. However, the requirement of longer bitstreams to ensure computing accuracy leads to higher latency, partially offsetting the low-complexity advantage of SC. To address this, this work utilizes a dynamic truncation method for stochastic bitstreams, and designs an energy-efficient counter-based addition circuit (CBAC) through effective bit recognition and correlation. Further, a tree-structured cascading architecture is then used to perform multi-input addition computing. Experimental results demonstrate that the proposed CBAC outperforms the state-of-the-art designs. For instance, a 16-input configuration achieves at least 75.9% reduction in mean square error (MSE) and a more than 43.1% reduction in area. When applied to polynomial computation and Gaussian filtering, the proposed architecture exhibits superior accuracy and efficiency, delivering MSE reductions of at least 10.7% and area reductions exceeding 6.8%.


  • A High-Performance Low-Power Double-Node Upset Resilient Latch for Harsh Radiation Environments

    by Na Bai, Yusheng Xia, Yaohua Xu, Yi Wang, Shaowei Wang, Xiaoqing Wen

    With the advancement of semiconductor technology, circuits have become increasingly susceptible to errors induced by radiation. Traditional approaches to enhancing the resilience of circuits against single-node upsets (SNUs) are insufficient to meet the robustness standards of modern designs. This article proposes a high-performance, low-power latch, named high-performance low-power double-node upset resilient latch (HLDRL), which is designed to exhibit exceptional resilience against double-node upsets (DNUs). Its design has six intricately interconnected C-elements (CEs) and two three-input CEs, for error interception, ensuring robust performance even in the case of DNUs. The Technology Computer Aided Design (TCAD) tool is used to validate the effectiveness of the HLDRL. Besides, comprehensive simulations are conducted utilizing the advanced SMIC 55-nm process technology. These simulation results show that our proposed HLDRL latch can autonomously recover from any DNU and thereby ensure the integrity of the system. Moreover, compared with existing DNU-resilient latches, the proposed HLDRL latch exhibits substantial improvements in terms of multiple metrics. On average, the proposed latch achieves an impressive 29.39% dynamic power saving, a remarkable 40.04% increase in speed, a notable 4.81% reduction in area, and a substantial 53.96% decrease in the power-delay–area product (PDAP). In the post-layout simulation, the proposed latch achieves a 30.39% dynamic power saving, a 36.47% increase in speed, and an impressive 52.38% decrease in PDAP. Furthermore, the proposed latch demonstrates enhanced resilience against variations in process, supply voltage, and temperature (PVT).


  • Resistive Defect Analysis and Fault Modeling of DICE Memory in Commercial 40-nm CMOS Technology

    by Yi Wang, Jiahao Yin, Yaohua Xu, Na Bai, Chunmei Hu

    The complex architecture of double interlocked storage cell (DICE) static random access memories (SRAM), which consists of 12 Metal Oxide Semiconductor Field Effect Transistors (MOSFETs), presents challenges in manufacturing test. Manufacturing defects can cause faults in DICE SRAMs, and undetected faults may lead to test escapes, resulting in early in-field failures. These issues can prevent DICE SRAMs from meeting the applications requiring high quality in sectors such as aerospace. This paper proposes a new type of SEU fault for DICE SRAM, analyzing how physical defects influence its SEU resilience, which forms a new fault space together with parameter faults and functional faults.This paper presents a detailed analysis and fault modeling of resistive defects that may occur in DICE SRAMs and proposes a novel test algorithm to enhance fault coverage and reduce test escapes during manufacturing test. First, we define and classify the fault space, including functional faults, parameter faults, and single event upset (SEU) faults, and outline the methodology to validate faults within this space. Next, we inject resistive defects into the SPICE netlist of DICE SRAMs, conduct SPICE simulation, and inspect its corresponding behavior. Furthermore, through the analysis of the fault modeling results of DICE SRAMs and evaluating the test coverage and limitations of existing test solutions, we propose a new test algorithm. This results in a substantial increase in dynamic fault coverage from 20.57% to 28.37% and an overall improvement in fault coverage from 33.88% to 39.89% when compared to the March C+ algorithm, while reducing the possibility of test escape for parameter faults (including HSNM and RSNM faults) and SEU faults. In summary, the findings of the paper are effective in detecting the realistic faults in DICE SRAMs, thereby bolstering the effectiveness of DICE SRAMs test for critical applications.

  • A DICE Flip-Flop Design by Resetting Redundancy Hardening for Single Event Upset Tolerance

    by Yaohua Xu, Zeteng Liu, Yi Wang, Na Bai, Ye Liu

    The standard cell library forms the foundation of chip design. Conducting rational and efficient radiation-hardening design and performance verification for the library is essential to ensure the safe on-orbit operation of aerospace components. This paper investigates the resettable dual interlocked storage cell (DICE) flip-flop (DICE-FF) from a radiation-hardened cell library and proposes a flip-flop (DICEREFF) with reset-redundant control circuitry to improve its radiation tolerance. The study examines the effects of Single Event Upset (SEU) both before and after the redundancy hardening of the reset control, identifying sensitive layout regions and critical circuit nodes. Simulations show that the redundancy-hardened DICE-REFF demonstrates strong SEU tolerance, with a Linear Energy Transfer (LET) threshold of 37 MeV/mg/cm2. Additionally, the circuit area increased by only 5% compared to the non-hardened design, with no loss in performance. Additionally, a test chip was designed, and the SEU resistance of the flip-flop was validated through both fault injection and irradiation experiment. The study also explores the impact of different implementation strategies on fault rates.


  • A Low Cost Triple-Node-Upset Self-Recoverable Latch Design for Nanoscale CMOS

    by Na Bai, Pengshuai Zhang, Yaohua Xu, Yunhui Guo, Hang Li, Yi Wang, Xiaoqing Wen

    Due to the constant reduction of semiconductor feature sizes, the charge sharing effect is getting worse. The occurrence probability of triple-node upset (TNU), which has a significant impact on the circuit’s reliability, is rising in nanoscale digital circuits. This paper proposes a low overhead and TNU self-recovery latch (TNU-SR) based on the 28 nm CMOS technology to efficiently tolerate TNU and reduce the power dissipation and latency of the latch. The latch design proposed in this article adopts stacked transistors to reduce sensitive nodes and area overhead. The TNU-SR latch uses clock gating and high-speed transfer path technology, resulting in low power dissipation and low delay. For simulation results, compared with existing latches, it can be known that the TNU-SR latch performs well in terms of delay, power consumption, and area, so it is a good choice for radiation-hardened design for space applications. In addition, compared with the latest TNU self-recovery latch (HLTNURL), the TNU-SR latch proposed in this paper reduces the Power-Area-Delay Product (PADP) metric by 111.13%.


  • Design and Application of Mean and Square Root Circuits for Stochastic Computing

    by Shaowei Wang, Kai Shi, Yaohua Xu, Yi Wang, Yongqiang Zhang

    Stochastic computing (SC) is an unconventional computing paradigm that represents values using probabilities. This representation enables simple logic gates to perform complex arithmetic operations. This work proposes two low hardware-cost stochastic mean circuits for even and odd inputs, respectively, along with a high-accuracy stochastic square root circuit. The circuits are designed by considering correlation technique and achieve excellent performance. Experimental results demonstrate that the proposed mean circuits surpass previous counterparts in computing accuracy and hardware cost. For instance, the proposed 9-input mean circuit can achieve at least an 86.7% reduction in mean square error (MSE) and a 28.6% reduction in area. For the square root circuit, the proposed design achieves a reduction in MSE of at least 19.6%. The proposed circuits are further demonstrated with the Niblack binarization algorithm, which shows superior performance of accuracy.


  • Low-iteration hybrid computing CORDIC architecture

    by Na Bai, Ruizheng Qu, Yaohua Xu, Yi Wang, Xiaojie Chen, Li Li


  • Open Access

    A 6.77-to-7.67 GHz Dual-Core VCO with Varactors Noise Compensation Technique

    by Na Bai, Feifan Zhu, Yi Wang, Yao Hua Xu, Tixing Gan, Fei Yang

    In this paper, A 7.2 GHz dual-core H-shaped-inductor based voltage-controlled oscillator (VCO) with single-mode oscillation is presented. The VCO utilizes a novel 6-port H-shaped structure formed by coupling two single-turn inductors, which effectively suppresses unwanted oscillation modes and enhances the quality factor (Q) of the desired mode. Furthermore, a noise compensation technique is proposed, which employs varactors to mitigate thermal noise induced by bias resistors, thereby improving the VCO’s overall performance. Implemented in a 110 nm CMOS process, the VCO consumes 39.6 mW and achieves a phase noise of -128.9 dBc/Hz at a 1 MHz frequency offset from a 7.2 GHz carrier frequency, with a corresponding figure of merit (FoM) of 190.1 dBc/Hz.