HZHaonan Zhu

Haonan Zhu

*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


Haonan Zhu has a total of 29 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Ouroboros: Wafer-Scale SRAM CIM with Token-Grained Pipelining for Large Language Model Inference

    by Yiqi Liu, Yudong Pan, Mengdi Wang, Shixin Zhao, Haonan Zhu, Yinhe Han, Lei Zhang, Ying Wang


  • A Programmable Compute-Near-Memory Architecture for 130 fps Visual Odometry on Embedded Systems

    by Gangliang Lin, Haonan Zhu, Yuquan He, Mengdi Wang, Yujie Wang, Kai Zhang, Yinhe Han, Cong Shi, Huawei Li, Xiaowei Li, et al.

  • A Potential Enabler for High-Performance In-Memory Multi-Bit Arithmetic Schemes With Unipolar Switching SOT-MRAM

    by Haonan Zhu, Bi Wu, Tianyang Yu, Ke Chen, Chenggang Yan, Weiqiang Liu

    Due to the physical separation of data processing and storage, the conventional Von Neumann architecture exists excessive data migration overhead to curtail the progress of data-intensive applications. In this way, the Computing-in-Memory (CiM) architecture is proposed. Due to the boolean property of the memory cell, the current CiM mainly focuses on single-bit logic design. For the multi-bit arithmetic design, a prevalent patchwork approach is employed using single-bit logic, leaving the design with insufficient parallelism. This paper proposes a high-performance in-memory multi-bit addition (M-Add) and multiplication (M-Mul) scheme based on unipolar switching SOT-MRAM. For the M-Add scheme, transmission logic-based circuit design is proposed to realize single-step inter-column XOR operations, which is logically fits perfectly the g operator of parallel prefix algorithm. Further, the oBK algorithm is presented to maximize the g operator occupancy. For the M-Mul scheme, mapping the Booth decoder to the control signal of the proposed modified flip-flop queue, only two steps are required to realize the decoding of three encoded signals in parallel. The simulation results indicate the proposed design reduces the latency of N-bit Add (N-bit Mul) by an average of 82.6% (31.5%) compared to state-of-the-art CiM designs. Further, a CNN application based on proposed operations achieves 1.23 TOPS/w on the CIFAR-10 dataset, with an average of 47.57% increase over other CiM designs.


  • High-Performance STT-MRAM-Based Computing-in-Memory Scheme Utilizing Data Read Feature

    by Bi Wu, Kai Liu, Tianyang Yu, Haonan Zhu, Ke Chen, Chenggang Yan, Erya Deng, Weiqiang Liu

    With the development of Artificial Intelligence (AI) and Binary neural networks (BNN), the computing efficiency of the computing system is expected to be much better, however, enormous amounts of data processing have caused an intolerable ‘memory wall’ and ‘power wall’ challenge for traditional Von Neumann architectures. Therefore, more advanced Computing-in-memory (CiM) architectures are proposed. The emerging non-volatile memory Spin Transfer Torque Magnetic Random Access Memory (STT-MRAM), with its fast access speed, near-zero leakage power consumption and high density is one of the most competitive carriers for CiM architectures.Due to the inherent read-write asymmetry of STT-MRAM, the read overhead is multiples smaller than the exaggerated write overhead. To better exploit the potential of MRAM, we focus on implementing an CiM based on the reading process.This work introduces the principle of CiM and proposes four basic logic operations (XNOR, XOR, AND and OR) based on STT-MRAM. Furthermore, a delay-based XNOR-popcount (XNOR-PC) operation for BNN based on XNOR operation is proposed. Incorporating the reading characteristics of STT-MRAM and slight modifications to the peripheral circuitry, these operations achieve significant optimisation in terms of performance and energy consumption. Simulation results show the proposed scheme can reduce the latency of XOR, AND, and OR operations at least by 99.3%, 82.2% and 80.2% compared with the existing design. The proposed XNOR-PC operation achieves at least a 59.4% reduction in power consumption and has a high degree of reliability. Also, Monte Carlo simulations prove the feasibility and robustness of the schemes.


  • MLiM: High-Performance Magnetic Logic in-Memory Scheme With Unipolar Switching SOT-MRAM

    by Bi Wu, Haonan Zhu, Ke Chen, Chenggang Yan, Weiqiang Liu

    Conventional computing architectures based on the von Neumann structure are suffering from the severe ‘memory wall’ issue due to the isolation and speed mismatch between memory and processor. As a promising solution, the concept of logic in-memory (LiM) has been proposed to effectively reduce the overhead of data migration and has been extensively studied in various memory technologies such as SRAM, DRAM, MRAM, ReRAM, etc. Among them, SOT-MRAM combines the advantages of non-volatility, low static power consumption, ultra-fast read/write speed, and high density, has emerged as one of the most promising candidates for low-power LiM implementations. In this paper, four in-memory logic operations, AND, OR, MAJ and full-addition (FA), are proposed based on the Unipolar Switching (US) SOT-MRAM devices. Incorporating the emerging switching behavior of SOT-MRAM, these operations can be performed with the basic memory access operations (read/write) with negligible modifying peripheral circuits. Meanwhile, by optimizing the operation steps, the performance degradation caused by the instability of SOT-MRAM device can be minimized in the proposed LiM architecture. Detailed simulation results show that the proposed design can reduce the latency (energy) of AND, OR operations at least by 71.2%, 74.4% (30.0%, 35.4%) compared with the existing SRAM and STT-MRAM designs. For MAJ and FA operations, the performance is improved by at least 34.7% and 44.8% compared to the existing design. The robustness of our design is demonstrated by the 100% pass of the 1000 samples Monte Carlo simulations for the sufficient switching current margin and the effectiveness of basic operations.


  • An Energy-Efficient Computing-in-Memory (CiM) Scheme Using Field-Free Spin-Orbit Torque (SOT) Magnetic RAMs

    by Bi Wu, Haonan Zhu, Dayane Reis, Zhaohao Wang, Ying Wang, Ke Chen, Weiqiang Liu, Fabrizio Lombardi, Xiaobo Sharon Hu

    The separation of memory and computing units in the von Neumann architecture leads to undesirable energy consumption due to data movement and insufficient memory bandwidth. Energy-efficient in-memory computing platforms have the potential to address such issues. Due to its non-volatility and advantageous features over CMOS (such as low power, near-zero leakage current and high integration density), spin-based devices have been advocated for in-memory computing. This paper proposes a field-free Spin Orbit Torque (FF-SOT) MRAM based computing-in-memory (CiM) scheme that realizes XNOR/XOR logic and a cascading adder. This novel FF-SOT-CiM design does not require expensive peripheral circuits for computation while using the same memory cell design as a SOT-MRAM. Furthermore, FF-SOT-CiM does not require additional write cycles to save the result of its computations in the memory. The design offers higher write speed and; lower operating energy compared to CiM schemes based on other technologies; it also alleviates the source degeneration effect by leveraging an advanced switching mechanism. Extensive simulation results show that the proposed FF-SOT-CiM achieves up to 3.1x (2.6x) latency (energy) reduction compared to SRAM-based CiM, with negligible hardware overhead when performing in-memory XOR. ADD operations; the proposed FF-SOT-CiM can be to 5.0X and 1.5X faster and 3.4X and 1.1X more energy efficient than existing STT-based and FeFET-based schemes, respectively.


  • Open Access

    High-performance STT-MRAM Logic-in-Memory Scheme Utilizing Data Read Features

    by Kai Liu, Bi Wu, Haonan Zhu, Weiqiang Liu

    In the Big Data era, enormous amounts of data processing have caused an intolerable 'memory wall' challenge for traditional Von Neumann architectures. Therefore, more advanced Logic-in-memory (LiM) computing architectures are proposed with integrated computing and memory units that reduce data migration. The emerging non-volatile memory STT-MRAM, with its fast access speed, near-zero leakage power consumption and high density is one of the most competitive carriers for LiM architectures. This work introduces the principle of LiM and proposes four basic logic operations (XNOR, XOR, AND and OR) based on STT-MRAM. Incorporating the reading characteristics of STT-MRAM and slight modifications to the peripheral circuitry, these operations achieve significant optimisation in terms of latency and energy consumption. From the experimental results, the proposed scheme can reduce the latency of XOR, AND and OR operations at least by 99.3%, 82.2% and 80.2% compared with the existing design. Also, 500 Monte Carlo samples prove the feasibility and robustness of the proposed scheme.

  • Open Access

    Capacity-oriented High-performance NV-TCAM Leveraging Hybrid MRAM Scheme

    by Didi Zhang, Bi Wu, Haonan Zhu, Weiqiang Liu

    As a special type of memory, Ternary Content Addressable Memory (TCAM) has been widely employed in network routers and various applications that require high speed table lookup. However, the large cell area and high static power consumption of traditional CMOS-based TCAMs have always constrained its advance toward higher capacity. Therefore, many non-volatile memory devices, such as ReRAM, FeFET, and MRAM, are emerging in TCAM designs to address these issues. However, SOT-MRAM devices with ultra-high write speed, low power consumption and small cell area are not effectively utilized in TCAM designs due to their low tunnel magnetoresistance (TMR) ratio, etc. In this work, first, combining the ultra-high performance of SOT-MRAM and the reliability of STT-MRAM, a 4T-3MTJ SOT/STT hybrid NV-TCAM structure with ultra-high integration density is proposed. However, to eliminate the lower match line utilization caused by the low TMR of SOT-MRAM, another highly reliable 7T-3MTJ NV-TCAM structure is proposed. Simulation results show that the proposed two TCAM structures can achieve search speed of 0.4ns and 0.15ns with only 4 and 7 transistors, respectively. While maintaining high performance, their write energy is only 0.133pJ/bit, a maximum reduction of 91.6% over existing TCAM designs. Meanwhile, the array-level simulations prove the great potential of the proposed design in terms of integration density and performance.


  • Common-Mode Voltage Cancellation for Reducing the Common-Mode Noise in DC–DC Converters

    by Lihong Xie, Xinbo Ruan, Haonan Zhu, Yu-Kang Lo

    Common-mode (CM) noise cancellation method is an attractive approach to reduce the original CM noise in power converters, which helps to shrink the size of electromagnetic interference (EMI) filter. Traditional CM noise cancellation method cancels the CM noise current by adding a branch consisted of a compensation voltage source and a compensation capacitor, which generates a cancellation current and nullifies the noise current. As the duality, the voltage cancellation method is also feasible for reducing CM noise, which has not been investigated in DC-DC converters. In this paper, a common-mode voltage cancellation (CMVC) method is proposed, which cancels the CM voltage of the converter by inserting compensation voltage source into the input power cord. The requirement for achieving CMVC is that the cancellation voltage and the parasitic capacitances in the circuit are matched. The applications of CMVC method in basic non-isolated and isolated dc-dc converters are presented, and the considerations for practical applications of this method are discussed. Finally, a buck converter and a half-bridge LLC resonant converter prototype are built, and the proposed method is verified by experiment, respectively.


  • Hybrid Passive Cancelation Method for Reducing Common-Mode Noise in Isolated Power Converters

    by Lihong Xie, Xinbo Ruan, Haonan Zhu, Yu-Kang Lo

    Shielding technique is an effective method for reducing the common-mode (CM) electromagnetic interference (EMI) in isolated power converter. However, the inter-winding capacitance between the shielding layer and its adjacent winding provides path for CM noise current. This CM noise current can suppressed by making the averaged electric potential (AEP) of the shielding layer equal to that of its adjacent winding. In this paper, a balancing winding is introduced to be connected with the shielding layer, for adjusting the AEP of the shielding layer to be equal to that of its adjacent winding and thus cancelling the CM noise current. Furthermore, the hybrid passive cancellation (HPC) method is proposed, in which the balancing winding is reused as the compensation winding to generate the compensation voltage for canceling the noise current flowing through the parasitic capacitance between the high dv/dt node in the primary circuit and protective earth. The proposed shielding technique with balancing winding and the HPC method are verified by the experimental results.