In this letter, we quantify the impact of device limitations on the classification accuracy of an artificial neural network, where the synaptic weights are implemented in a Ferroelectric FET (FeFET) based in-memory processing architecture. We explore a design-space consisting of the resolution of the analog-to-digital converter, number of bits per FeFET cell, and the neural network depth. We show how the system architecture, training models and overparametrization can address some of the device limitations.
The technological exploitation of ferroelectricity in CMOS electron devices offers new design opportunities, but also significant challenges from an integration, optimization and modelling perspective. We here revisit the working principle and the modelling of some novel ferroelectric based devices, with an emphasis on energy efficiency and on applications to new computational paradigms.
We present a study based on numerical simulations and comparative analysis of recent experimental data concerning the operation and design of FeFETs. Our results show that a proper consideration of charge trapping in the ferroelectric-dielectric stack is indispensable to reconcile simulations with experiments, and to attain the desired hysteretic behavior of the current-voltage characteristics. Then we analyze a few design options for polysilicon channel FeFETs and, in particular, we study the influence of the channel thickness and doping concentration on the memory window, and on the ratio be…
A multi-bit digital weight cell for high-performance, inference-only non-GPU-like neuromorphic accelerators is presented. The cell is designed with simplicity of peripheral circuitry in mind. Non-volatile storage of weights which eliminates the need for DRAM access is based on FeFETs and is purely digital. The Multiply-and-Accumulate operation is performed using passive resistors, gated by FeFETs. The resulting weight cell offers a high degree of linearity and a large ON/OFF ratio. The key performance tradeoffs are investigated, and the device requirements are elucidated.
Hardware security has been a key concern in modern information technologies. Especially, as the number of Internet-of-Things (IoT) devices grows rapidly, to protect the device security with low-cost security primitives becomes essential, among which Physical Unclonable Function (PUF) is a widely-used solution. In this paper, we propose the first FeFET-based strong PUF exploiting the cycle-to-cycle (C2C) variation of FeFETs as the entropy source. Based on the experimental measurements, the proposed PUF shows satisfying performance including high uniformity, uniqueness, reconfigurability and rel…
Compute-in-memory (CiM) is a promising approach to alleviating the memory wall problem for domain-specific applications. Compared to current-domain CiM solutions, charge-domain CiM shows the opportunity for higher energy efficiency and resistance to device variations. However, the area occupation and standby leakage power of existing SRAMbased charge-domain CiM (CD-CiM) are high. This paper proposes the first concept and analysis of CD-CiM using nonvolatile memory (NVM) devices. The design implementation and performance evaluation are based on a proposed 2-transistor-1-capacitor (2T1C) CiM mac…
The current work reports an efficient deep neural network (DNN) accelerator where synaptic weight elements are controlled by ferroelectric domain dynamics. An integrated device-to-algorithm framework for benchmarking novel synaptic devices is used. In P(VDF-TrFE) based ferroelectric tunnel junctions, analog conductance states are measured using a custom pulsing protocol and associated custom circuits and array architectures for DNN training is simulated. Our results show precise control of polarization switching dynamics in multi-domain, polycrystalline ferroelectric thin films can produce con…
Field Programmable Gate Array (FPGA) is widely used in acceleration of deep learning applications because of its reconfigurability, flexibility, and fast time-to-market. However, conventional FPGA suffers from the tradeoff between chip area and reconfiguration latency, making efficient FPGA accelerations that require switching between multiple configurations still elusive. In this paper, we perform technology-circuit-architecture co-design to break this tradeoff with no additional area cost and lower power consumption compared with conventional designs while providing dynamic reconfiguration, …
Deep random forest (DRF), which incorporates the core features of deep learning and random forest (RF), exhibits comparable classification accuracy, interpretability, and low memory and computational overhead when compared with deep neural networks (DNNs) in various information processing tasks for edge intelligence. However, the development of efficient hardware to accelerate DRF is lagging behind its DNN counterparts. The key for hardware acceleration of DRF lies in efficiently realizing the branch-split operation at decision nodes when traversing a decision tree. In this work, we propose to…
In this work, we propose a ferroelectric FET(FeFET) time-domain compute-in-memory (TD-CiM) array as a homogeneous processing fabric for binary multiplication-accumulation (MAC) and content addressable memory (CAM). We demonstrate that: i) the XOR(XNOR)/AND logic function can be realized using a single cell composed of 2FeFETs connected in series; ii) a two-phase computation in an inverter chain with each stage featuring the XOR/AND cell to control the associated capacitor loading and the computation results of binary MAC and CAM are reflected in the chain output signal delay, illustrating full…
International College of Semiconductor Technology, National Yang-Ming Chiao Tung University（国立阳明交通大学半导体技术国际学院）
This work proposes a hybrid-precision neural network training framework with an eNVM based computational memory unit executing the weighted sum operation and another SRAM unit, which stores the error in weight update during back propagation and the required number of pulses to update the weights in the hardware. The hybrid training algorithm for MLP based neural network with 28 nm ferroelectric FET (FeFET) as synaptic devices achieves inference accuracy up to 95% in presence of device and cycle variations. The architecture is primarily evaluated using behavioral or macro-model of FeFET devices…
University of Notre Dame（圣母大学）；Fraunhofer IPMS-CNT（弗劳恩霍夫IPMS-CNT）
Nearest neighbor (NN) search is an essential operation in many applications, such as one/few-shot learning and image classification. As such, fast and low-energy hardware support for accurate NN search is highly desirable. Ternary content-addressable memories (TCAMs) have been proposed to accelerate NN search for few-shot learning tasks by implementing $L_\infty$ and Hamming distance metrics, but they cannot achieve software-comparable accuracies. This paper proposes a novel distance function that can be natively evaluated with multi-bit content-addressable memories (MCAMs) based on ferroelect…
A Reconfigurable Time-Domain In-Memory Computing Macro using FeFET-Based CAM with Multilevel Delay Calibration in 28 nm CMOS
Andrew and Erna Viterbi Faculty of Electrical and Computer Engineering, Technion - Israel Institute of Technology（安德鲁和埃尔纳·维特比电气与计算机工程学院，以色列理工学院）；GlobalFoundries Fab1 LLC and Co. KG（格罗方德Fab1有限责任公司及合伙公司）
In a number of machine learning models, an input query is searched across the trained class vectors to find the closest feature class vector in cosine similarity metric. However, performing the cosine similarities between the vectors in Von-Neumann machines involves a large number of multiplications, Euclidean normalizations and division operations, thus incurring heavy hardware energy and latency overheads. Moreover, due to the memory wall problem that presents in the conventional architecture, frequent cosine similarity-based searches (CSSs) over the class vectors requires a lot of data move…
A time-domain analog weighted-sum calculation model is proposed based on an integrate-and-fire-type spiking neuron model. The proposed calculation model is applied to multi-layer feedforward networks, in which weighted summations with positive and negative weights are separately performed in each layer and summation results are then fed into the next layers without their subtraction operation. We also propose very large-scale integrated (VLSI) circuits to implement the proposed model. Unlike the conventional analog voltage or current mode circuits, the time-domain analog circuits use transient…
Comparative Evaluation of Memory Technologies for Synaptic Crossbar Arrays -- Part I: Robustness-driven Device-Circuit Co-Design and System Implications
