High On/Off Ratio Spintronic Multi‐Level Memory Unit for Deep Neural Network

Abstract

Spintronic devices are considered as one of the most promising technologies for non‐volatile memory and computing. However, two crucial drawbacks, that is, lack of intrinsic multi‐level operation and low on/off ratio, greatly hinder their further application for advanced computing concepts, such as deep neural network (DNN) accelerator. In this paper, a spintronic multi‐level memory unit with high on/off ratio is proposed by integrating several series‐connected magnetic tunnel junctions (MTJs) with perpendicular magnetic anisotropy (PMA) and a Schottky diode in parallel. Due to the rectification effect on the PMA MTJ, an on/off ratio over 100, two orders of magnitude higher than intrinsic values, is obtained under proper proportion of alternating current and direct current. Multiple resistance states are stably achieved and can be reconfigured by spin transfer torque effect. A computing‐in‐memory architecture based DNN accelerator for image classification with the experimental parameters of this proposal to evidence its application potential is also evaluated. This work can satisfy the rigorous requirements of DNN for memory unit and promote the development of high‐accuracy and robust artificial intelligence applications.

Demonstration of Synaptic Behavior in a Heavy-Metal-Ferromagnetic-Metal-Oxide-Heterostructure-Based Spintronic Device for On-Chip Learning in Crossbar-Array-Based Neural Networks

“…Song et al have reported simulation-based on-chip learning in ANNs using experimentally obtained synaptic characteristic in a skyrmionic device . But the remaining experimental spintronic studies only show device-level data (without any system-level simulation based on the device data), ,, or show training of an SNN or some other biologically motivated network ,, (which is different from an ANN and does not use LTD and LTP of a synapse the way an ANN does), or use binary spintronic synapses/magnetic tunnel junctions (MTJs) , as opposed to the multibit synapses we focus on here (one device stores 5 bits in our case).…”

Recent advances in two-dimensional intrinsic ferromagnetic materials Fe3X(X=Ge and Ga)Te2 and their heterostructures for spintronics

“…166 Spintronic devices have entered the scientific and industrial communities owing to their promising scope in non-volatile memory and computing applications. 167 In spin memory devices, binary information is stored in the magnetic state of an FM component, in which "0" and "1" states correspond to opposite orientations of the magnetic moment. 167,168 In 2021, Zhang et al 169 implemented a multi-level spin memory device with 8-states based on Fe 3 GeTe 2 and demonstrated a switching current density of 4 × 10 5 A cm −2 and corresponding switching power dissipation of 2.5 × 10 14 W m −3 by varying writing current from 0.0 to 2.0 mA at a fixed magnetic field of −0.06 T or −0.04 T (Fig.…”

On-chip learning of a domain-wall-synapse-crossbar-array-based convolutional neural network

“…Lequeux et al experimentally demonstrated current-based control of multiple conductance states (measured through TMR) of such a tunnel-junction device [68]. Zhang et al, very recently, experimentally demonstrated current-controlled multiple conductance states in a chain of MTJ-s (as opposed to a single domain-wall device) and showed that the chain can act as a synapse unit for neural-network crossbars [23].…”

“…On-chip learning is known to provide several advantages for edge devices in terms of data security and the like [6,7]. This makes the domain-wall synapse device an important device in the context of the ongoing research on spintronics-based neuromorphic computing [19][20][21][22][23][24][25][26].…”