19 May 2023 in IEEE Journal of Solid-State Circuits
A highly energy-efficient neuromorphic computing-in-memory (Neuro-CIM) processor is proposed for ultralow-power deep learning applications. Neuro-CIM can support spiking neural network (SNN) to eliminate the power and area overhead of previous CIM processor. The sign extended bits gating reduces the bitline (BL) voltage switching rate due to negative small-magnitude weights allowing 38% power reduction at 8-b weight condition and 25% at 4-b weight condition. In addition, Neuro-CIM replaces high-precision analog-to-digital converter (ADC) with 1-b comparator by exploiting the characteristic of the SNN, and thus, power and area efficiencies are significantly enhanced. Furthermore, the early stopping scheme terminates unnecessary neuronal operations, reducing power consumption by 31%. In addition, the analog and digital networks are integrated for high reconfigurability and energy efficiency. The analog network with voltage folding circuit enables accurate analog-domain aggregation by increasing the dynamic range without compromising the voltage resolution. The digital network-in-memory supports input–output channel extension for high reconfigurability and input data reuse scheme for reducing input memory (IMEM) access. Neuro-CIM is fabricated in 28-nm CMOS technology and occupies the 2.9-mm $^{2}$ die area. It achieves the state-of-the-art energy consumption per classification of 0.72 $\mu$ J and 92.1% accuracy for CIFAR-10 with 4-b input and 4-b weight and 372.2 $\mu$ J and 65.8% accuracy for ImageNet with 6-b input and 8-b weight at 200 MHz, and 1.1-V conditions. Moreover, the proposed CIM processor achieves 310.37 tera operations per second/watt (TOPS/W) and 90.7% accuracy with 4-b input and 1-b weight for Canadian Institute for Advanced Research, 10 classes (CIFAR-10) classification.
10 October 2023 in IEEE Journal of Solid-State Circuits
This article presents DynaPlasia, a reconfigurable eDRAM-based in-memory computing (IMC) processor with a novel triple-mode cell. It enables higher system-level performance and efficiency in a resource-limited environment. DynaPlasia proposes five key features that can enhance the energy efficiency and area efficiency of the IMC accelerator: 1) dynamic reconfigurable core architecture (DRECA), which dynamically reconfigures the effective IMC macro size according to DNN workloads; 2) the triple-mode cell is reconfigured as IMC PE, unit DAC, and unit cell to optimize system resource utilization adaptively; 3) hierarchical in-memory ADC (Hi-ADC) reduces ADC area overhead with tripe-mode cell and saves power with the hierarchical operation; 4) signed-input signed-weight (SISW) IMC reduces logic switching activity with signed magnitude representation; and 5) leakage-tolerant computing (LTC) to enable high-density cell design. The DynaPlasia is fabricated in 28-nm CMOS technology and occupies a 20.25-mm 2 die area with 9.6-Mb cell capacity. Its measured system peak throughput is 19.5 TOPS (4 b–5 b), and system peak efficiency is 56.0 TOPS/W (4 b–5 b). Even with the ResNet50 @ ImageNet dataset, system efficiency is 29.5 TOPS/W which is 2.5 $\times$ higher than the previous state-of-the-art.
16 March 2022 in IEEE Transactions on Circuits and Systems II: Express Briefs
Computing-in-memory (CIM) shows high energy-efficiency through the analog DNN computation inside the memory macros. However, as the DNN size increases, the energy-efficiency of CIM is reduced by external memory access (EMA). One of the promising solutions is eDRAM based CIM to increase memory capacity with a high density cell. Although the eDRAM-CIM has a higher density than the SRAM-CIM, it suffers from both poor robustness and a low signal-to-noise ratio (SNR). In this brief, the energy-efficient eDRAM-CIM macro is proposed while improving computational robustness and SNR with three key features: 1) High SNR voltage-based accumulation with segmented BL architecture (SBLA), resulting in 17.1 dB higher SNR, 2) canceling PVT/leakage-induced error with common-mode error canceling (CMEC) circuit, resulting in 51.4% PVT variation reduction and 51.4% refresh power reduction, 3) a ReLU-based zero-gating ADC (ZG-ADC), resulting in ADC power reduction up to 58.1%. According to these new features, the proposed eDRAM-CIM macro achieves 81.5-to-115.0 TOPS/W energy-efficiency with 209-to-295 $\mu \text{W}$ power consumption when 4b $\times $ 4b MAC operation is performed with 250 MHz core frequency. The proposed macro also achieves 91.52% accuracy at the CIFAR-10 object classification dataset (ResNet-20) without accuracy drop even with PVT variation.
17 November 2023 in IEEE Transactions on Circuits and Systems II: Express Briefs
Computing-in-memory (CIM) has emerged as an energy-efficient hardware solution for machine learning and AI. While static random access memory (SRAM)-based CIM has been prevalent, growing attention is directed towards leveraging dynamic random access memory (DRAM) and non-volatile memory (NVM) with its unique characteristics such as high-density and non-volatility. This paper reviews the evolving trends in DRAM and NVM-based CIM, which have faced unique challenges that arise from SRAM despite their advantages. For instance, the DRAM cell’s density comes with leakage and refresh issues, impacting efficiency and computing accuracy. NVM-CIM faces computing accuracy challenges of resistance-based computation with low signal margins and non-linear characteristics. This tutorial discusses the current status and future directions in DRAM-CIM and NVM-CIM research, which address the abovementioned challenge.
16 February 2024 in IEEE Journal of Solid-State Circuits
This article presents Scaling-computing-in-memory (CIM), an energy-efficient embedded dynamic random access memory (eDRAM)-based in-memory-computing (IMC) accelerator with a dynamic-scaling readout for signal-to-quantization-noise ratio (SQNR) boosting and analog-to-digital converter (ADC) overhead reduction. It greatly saves the ADC cost by reducing the required number of ADC-bit and ADC operations by codesigning the algorithm and hardware. Scaling-CIM proposes three key features: 1) dynamic scaling ADC (DSA) boosts SQNR of multibit operation even with low-bit ADC; 2) adaptive analog bit-parallel (AABP) accumulation reduces the redundant ADC operation; and 3) layer-wise adaptive bit-truncation (LABT) search further enhances efficiency on benchmarks. The Scaling-CIM is fabricated in 28-nm CMOS technology and occupies a 2.03-mm2 die area with an 800-kb eDRAM cell. It achieves 39.7-TOPS/W (8–9 b) energy efficiency on the RestNet-18 benchmark and $1.96\times $ higher efficiency figure of merit (FoM) than the previous IMC-based accelerator.
15 November 2023 in IEEE Journal of Solid-State Circuits
In this article, we propose a complementary deep-neural-network (C-DNN) processor by combining convolutional neural network (CNN) and spiking neural network (SNN) to take advantage of them. The C-DNN processor can support both complementary inference and training with heterogeneous CNN and SNN core architecture. In addition, the C-DNN processor is the first DNN accelerator application-specific integrated circuit (ASIC) that can support CNN–SNN workload division by using their magnitude–energy tradeoff. The C-DNN processor integrates the CNN–SNN workload allocator and attention module to find a more energy-efficient network domain for each workload in DNN. They enable the C-DNN processor to operate at the energy optimal point. Moreover, the SNN processing element (PE) array with distributed L1 cache can reduce the redundant memory access for SNN processing, resulting in a 42.2%–49.1% reduction. For high energy-efficient DNN training, the C-DNN processor integrates the global counter and local delta-weight (LDW) unit to eliminate power-consuming counters for a forward delta-weight generation. Furthermore, the forward delta-weight-based sparsity generation (FDWSG) is proposed to reduce the number of operations for training by 31%–79% The C-DNN processor achieves an energy efficiency of 85.8 and 79.9 TOPS/W for inference with CIFAR-10 and CIFAR-100, respectively (VGG-16). Moreover, the C-DNN processor achieves ImageNet classification with state-of-the-art energy efficiency of 24.5 TOPS/W (ResNet-50). For training, the C-DNN processor achieves the state-of-the-art energy efficiency of 84.5 and 17.2 TOPS/W for CIFAR-10 and ImageNet, respectively. Furthermore, it achieves 77.1% accuracy for ImageNet training with ResNet-50.