HJHoichang Jeong

Hoichang Jeong

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hoichang Jeong has a total of 36 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • SERAH-CIM: Sparsity-Aware Effective Row Activation Analog-Digital Hybrid eDRAM CIM With In-Macro Multi-Row-Multi-Task Control

    by Hoichang Jeong, Seungbin Kim, Dongwook Kim, Jueun Jung, Kyuho Jason Lee

    This article presents a sparsity-aware analog-digital hybrid embedded dynamic random access memory (eDRAM) computing-in-memory (CIM) processor for highly energy-efficient deep neural network (DNN) acceleration. Although CIM architectures execute multiplication-and-accumulation (MAC) more efficiently than von Neumann architectures, their practical energy efficiency for DNN acceleration remains limited due to several challenges. First, prior CIMs struggled to exploit massive sparsity because of their highly parallelized structures. Second, CIMs typically adopt computation in either the analog or the digital domain, facing fundamental trade-offs between analog-to-digital converter (ADC) overhead and throughput. Third, system throughput is degraded due to 1) workload imbalance among CIM macros during sparsity-aware computation caused by a random sparsity pattern, which ruins CIM macro utilization and 2)frequent refresh and weight-update operations that have plagued prior eDRAM CIMs. To address these challenges, the proposed eDRAM CIM processor introduces four key features: 1) input activation (IA) grouping convolution, which completely skips zero-weight computations by activating only the effective rows of the CIM macro, increasing the effective computation ratio by $4.59\times $ ; 2) a hybrid-CIM macro integrated with SAR-Flash ADC (SF-ADC) and reversed-MAC near-memory logic (RM-NML) for energy-efficient MAC operations in both the analog and the digital domains, improving macro efficiency by $2.39\times $ ; 3) sparsity-aware proactive scheduling (SPS) to maximize CIM macro utilization, reducing system latency by 10.4%; and 4)in-macro Multi-row-Multi-task (MRMT) control that enables concurrent refresh/update during in-memory computation, resulting in a 22.0% reduction in system latency and a $1.3\times $ increase in system energy efficiency. Fabricated in a 28 nm CMOS process, the proposed processor demonstrates high-energy efficiency across various benchmarks, outperforming previous CIM processors by $1.55\times $ and $10.37\times $ for ResNet-18 and VGGNet-16, respectively.


  • TeSLIC: A 28-nm Ternary Sparsity-Aware Latch-Based Computing-in-Memory Macro

    by Myoung Kim, Hoichang Jeong, Woo-Seok Kim, Yesong Jeong, Junyoung Park, Yong Kyu Lee, Min Woo Ryu, Kyuho Jason Lee, Kyung Rok Kim

  • CINELL: An Energy-Efficient Compute-In/Near-Memory eDRAM Processor for Sparse Transformer-Based Large Language Models

    by Sunhong An, Hoichang Jeong, Seungbin Kim, Kyuho Jason Lee

    Large language models (LLMs) based on transformer architecture have significantly advanced various AI applications. However, the multihead self-attention (MHSA) in a transformer requires extensive computation and memory bandwidth. Techniques such as sparse attention and attention formula reordering have been explored recently to mitigate these challenges. However, existing processors still struggle to process LLMs efficiently. To overcome these limitations, an energy-efficient compute-in/near-memory (CINM) embedded dynamic random access memory (eDRAM) processor CINELL is proposed, incorporating the following three key features: 1) an attention block fusion computation (ABFC), which maximizes data reuse in the attention map, reduces external memory access (EMA) by 85.86%, and improves hardware utilization to 86.1%; 2) a CINM architecture designed to address the imbalance between memory and computation, while the heterogeneous pipeline reduces the system latency by 77.27%; and 3) a compute-in-memory (CIM) array that supports the cross-read operation (CRO) addresses data direction conflicts, achieving a 98.44% reduction in latency, while dual-row computation (DRC) with reduced adder logic (RAL) improves energy efficiency by $1.58\times $ . Designed in 28-nm CMOS technology, the CINELL processor achieves 36.28–58.05 TOPS/W and an F1-score of 92.41% on the SQuAD 1.1v using the BigBird-large model.

  • HYTEC: Compact and Energy-Efficient Analog-Digital Hybrid CIM With Transpose Ternary eDRAM

    by Hoichang Jeong, Seungbin Kim, Jeongmin Shin, Kyuho Jason Lee

    This article presents a remarkably high-density and energy-efficient analog-digital hybrid computing-in-memory (CIM) processor for ternary neural network (TNN) acceleration, utilizing a transpose ternary embedded DRAM bitcell. The proposed CIM processor significantly improves computational robustness and energy efficiency at both the macro- and system-level through key innovations: 1) current-mode vertical analog multiplication-and-accumulation (MAC) with gate voltage biasing in bitcell, reducing MAC variation by 87% under process, voltage, and temperature (PVT) variation; 2) ternary-bit per cycle (TPC) successive approximation register (SAR) analog-to-digital converter (ADC) with a shared capacitor digital-to-analog converter (CDAC), minimizing ADC area overhead to 15% and improving ADC efficiency by $1.49\times $ ; 3) horizontal digital partial sum (Psum) logic for area- and power-efficient Psum among MACs, reducing area by 39% and power by 57% compared to conventional full adder; and 4) input channel-first tiled-convolution that substantially enhances system energy efficiency by eliminating inter-macro data transactions, decreasing the network-on-chip power overhead to 2%. Fabricated in 28 nm CMOS technology, the proposed CIM processor achieves 1.58 Mb/mm2 of cell density, attaining 478 TOPS/W and 273.48 TOPS/W of macro and system energy efficiencies, respectively, outperforming state-of-the-art CIM processors.


  • DPIM: A 2T1C eDRAM Transformer-in-Memory Chip With Sparsity-Aware Quantization and Heterogeneous Dense–Sparse Core

    by Donghyuk Kim, Jae-Young Kim, Hyunjun Cho, Seungjae Yoo, Sukjin Lee, Sungwoong Yune, Sejeong Yang, Hoichang Jeong, Keonhee Park, Ki-Soo Lee, et al.

    Transformer models have revolutionized artificial intelligence (AI) applications across various domains, but their increasing complexity poses significant challenges in terms of computational and memory demands. While processing-in-memory (PIM) paradigms have been adopted to address these limitations, existing PIM-based transformer accelerators still face hurdles such as: 1) focusing solely on optimizing attention layers; 2) lack of sparsity exploitation for transformers; and 3) limited PIM macro capacity and low cell density, which degrades on-chip data reuse and increases external memory access (EMA). This article presents DPIM, a novel 2T1C eDRAM-based transformer-in-memory chip that addresses these challenges through three key innovations: 1) a sparsity-aware quantization (SAQ) scheme that significantly increases bit-slice sparsity in both activation and weight data, achieving ratios of 83.3% and 88.4%, respectively, with minimal accuracy loss; 2) a heterogeneous PIM core capable of efficiently handling both sparse and dense matrix multiplications (MMs); and 3) a high-density 2T1C eDRAM cell with a density of 1.38 Mb/mm2, enabling large-capacity PIM macros. By integrating these features, DPIM achieves improved computational efficiency and reduced EMA with enhanced on-chip data reuse. The DPIM chip, fabricated using 28-nm CMOS technology, achieves a throughput of 3.03–12.12 TOPS and an energy efficiency of 4.84–19.36 TOPS/W, all measured across INT8 and INT4 operations, respectively. It achieves a throughput density of 0.55 TOPS/mm2 with INT8 operation. With a total macro size of 4608 kb, the chip occupies a die area of 20.25 mm2 and operates at frequencies from 50 to 285 MHz with a supply voltage of 0.85–1.0 V. The DPIM successfully executes BERT-Large on the general language understanding evaluation (GLUE) dataset. Its macro density is 1413 kb/mm2, and the resulting density figure-of-merit (FoM) (macro density $\times $ throughput density) is $1.6\times $ – $115.8\times $ higher than previous works, representing a significant advancement in hardware design for efficient transformer processing.

  • A Multibit ReRAM Computing-in-Memory Processor With Adaptive Decision Level Nonlinear ADC for Ultra-Low-Energy Keyword Spotting in Mobile Devices

    by Dongwook Kim, Hoichang Jeong, Seungbin Kim, Heein Yoon, Kyuho Jason Lee

    This paper presents an ultra-low-energy keyword spotting (KWS) processor based on the charge-mode multi-level resistive random-access memory (ReRAM) bitcell and adaptive analog-to-digital converter (ADC) quantization. Previous ReRAM computing-in-memory (CIM) architectures have suffered several challenges, including excessive computing energy due to direct current branches, computation non-linearity, and throughput degradation resulting from analog-to-digital conversion. The proposed processor addresses these challenges by introducing a charge-mode multi-bit ReRAM bitcell (CRB), which achieves a 61.39% reduction in multiply-and-accumulate (MAC) energy. The CRB also enhances MAC linearity by $1.65\times $ . Additionally, the 5-bit additive powers-of-two ADC achieves a 65.42% reduction in analog-to-digital conversion energy, an 11.70 percentage points decrease in ADC accuracy loss, and a $1.33\times $ increase in conversion throughput. Furthermore, the pipelined layer fusion clusters reduce intermediate data movement energy to 97.2 nJ and enhance KWS system throughput by $1.35\times $ . The proposed processor is designed utilizing 45 nm CMOS technology and compatible ReRAM devices. The processor occupies an area of 0.94 mm2 with a 68.5 KB ReRAM cell. The processor also achieves $0.74~\mu $ J/decision energy consumption, 22.59 TOPS/W energy efficiency, and 92.7% accuracy on the Google Speech Commands Dataset.


  • C 2 IM-NN: A Low-Power 3D Point Clouds Matching Processor With 1D-CNN Prediction and CAM-Based In-Memory k-NN Searching

    by Jeongmin Shin, Hoichang Jeong, Seungbin Kim, Keonhee Park, Sangho Lee, Kyuho Jason Lee

    This paper presents a content addressable memory (CAM)-based computing-in-memory (C $^{2}$ IM) system designed for energy-efficient k-nearest neighbor (k-NN) searching in 3D point clouds. For autonomous driving applications, an essential process for perceiving the mobile robot’s movements in 3D space is k-NN searching. Especially with the limited hardware resources of mobile processors, the 3D point cloud is too large to upload onto the chip, leading to $O(N^{2})$ of external memory accesses and distance calculations. The proposed C $^{2}$ IM processor enhances energy efficiency and reduces power consumption through three key features: 1) Dilated 1D-CNN prediction enables voxel-based partitioning, reducing the external memory accesses from $O(N^{2})$ to $O(N)$ ; 2) Vertex clustering reorganizes groups of points into evenly distributed clusters based on the underlying data distribution and reduces the number of points of comparisons by 49.8%; and 3) In-memory k-NN searching with CAM achieves high system energy efficiency while minimizing data transactions between memory and computation logic. Designed with 28 nm CMOS technology, the proposed C $^{2}$ IM achieves up to 23.08 $\times$ energy efficiency, and 48.4% reduction in memory footprint compared to previous ASIC accelerators, and a 99.51% reduction in power consumption compared to state-of-the-art processor implemented in FPGA with high-bandwidth memory.


  • A 701.7 TOPS/W Compute-in-Memory Processor With Time-Domain Computing for Spiking Neural Network

    by Keonhee Park, Hoichang Jeong, Seungbin Kim, Jeongmin Shin, Minseo Kim, Kyuho Jason Lee

    Artificial neural networks have led to a higher computational burden, complicating inference tasks on low-power edge devices. Spiking neural network (SNN), which leverages sparse spikes for computation and data transmission, is an effective energy-efficient computing technique. However, the length of spike sequences in SNN varies significantly depending on the input coding method, among which rate coding still results in substantial data movement. A highly energy-efficient SNN accelerator with a time-domain CIM processor is proposed with three key features: 1) time-domain bitcell array for high linearity with lower energy, reducing 58.6% power consumption compared to inverter-chain architecture, 2) time-domain multi-bit accumulate for assisting multi-bit weights without analog-to-digital converter, achieving 47.2% energy reduction of domain-conversion energy, 3) analog precision reconstruction unit for supporting phase coding. The proposed TS-CIM is designed in 65 nm CMOS technology and achieves 701.7 TOPS/W energy efficiency, marking a 1.58 $\times$ enhancement compared to the state-of-the-art SNN CIM.


  • A Ternary Neural Network Computing-in-Memory Processor With 16T1C Bitcell Architecture

    by Hoichang Jeong, Seungbin Kim, Keonhee Park, Jueun Jung, Kyuho Jason Lee

    A highly energy-efficient Computing-in-Memory (CIM) processor for Ternary Neural Network (TNN) acceleration is proposed in this paper. Previous CIM processors for multi-bit precision neural networks showed low energy efficiency and throughput. Lightweight binary neural networks were accelerated with CIM processors for high energy efficiency but showed poor inference accuracy. In addition, most previous works suffered from poor linearity of analog computing and energy-consuming analog-to-digital conversion. To resolve the issues, we propose a Ternary-CIM (T-CIM) processor with 16T1C ternary bitcell for good linearity with the compact area and a charge-based partial sum adder circuit to remove analog-to-digital conversion that consumes a large portion of the system energy. Furthermore, flexible data mapping enables execution of the whole convolution layers with smaller bitcell memory capacity. Designed with 65 nm CMOS technology, the proposed T-CIM achieves 1,316 GOPS of peak performance and 823 TOPS/W of energy efficiency.