LAO-NCS: Laser Assisted Spin Torque Nano Oscillator-Based Neuromorphic Computing System

Abstract

Dealing with big data, especially the videos and images, is the biggest challenge of existing Von-Neumann machines while the human brain, benefiting from its massive parallel structure, is capable of processing the images and videos in a fraction of second. The most promising solution, which has been recently researched widely, is brain-inspired computers, so-called neuromorphic computing systems (NCS). The NCS overcomes the limitation of the word-at-a-time thinking of conventional computers benefiting from massive parallelism for data processing, similar to the brain. Recently, spintronic-based NCSs have shown the potential of implementation of low-power highdensity NCSs, where neurons are implemented using magnetic tunnel junctions (MTJs) or spin torque nano-oscillators (STNOs) and memristors are used to mimic synaptic functionality. Although using STNOs as neuron requires lower energy in comparison to the MTJs, still there is a huge gap between the power consumption of spintronic-based NCSs and the brain due to high bias current needed for starting the oscillation with a detectable output power. In this manuscript, we propose a spintronic-based NCS (196 × 10) proof-of-concept where the power consumption of the NCS is reduced by assisting the STNO oscillation through a microwatt nanosecond laser pulse. The experimental results show the power consumption of the STNOs in the designed NCS is reduced by 55.3% by heating up the STNOs to 100 • C. Moreover, the average power consumption of spintronic layer (STNOs and memristor array) is decreased by 54.9% at 100 • C compared with room temperature. The total power consumption of the proposed laser assisted STNO-based NCS (LAO-NCS) at 100 • C is improved by 40% in comparison to a typical STNO-based NCS at room temperature. Finally, the energy consumption of the LAO-NCA at 100 • C is expected to reduce by 86% compared with a typical STNO-based NCS at the room temperature.

Effect of magnetic anisotropy relaxation on laser-induced magnetization precession in thin galfenol films

“…Such time-dependent shifts of precession frequency controlled by laser-induced anisotropy changes may be further exploited in optically reconfigurable magnonic devices based on tuning the magnetostatic wave dispersion [40]. The possibility to tune the eigenfrequency of magnetic dynamics may find its application in the realization of laser-assisted spintorque nano-oscillators [48].…”

Prospect of Spin-Orbitronic Devices and Their Applications

“…Electrically controlled spin-orbitronic devices with a clear critical magnetization switching current are therefore capable of mimicking such artificial neurons in principle ( Diep et al., 2014 ; Sengupta et al., 2015b ; Kurenkov et al., 2019 ), but the magnetization states of such nonvolatile neurons may have to be initialized after each fire. The volatile spin-torque nano-oscillator (STNO) ( Torrejon et al., 2017 ; Farkhani et al., 2019 ; Liang et al., 2020 ), a specific type of STT-MTJ that oscillates spontaneously with a low d.c. current-tunable precession amplitude and finite magnetization relaxation, has recently been experimentally demonstrated as an efficient artificial neuron for speech recognition owing to its stability, low noise, and high nonlinearity of frequency and amplitude. As the spin-orbitronic version of STNO, spin hall nano-oscillators (SHNOs) had been also developed to implement as neurons ( Sato et al., 2019 ; Zahedinejad et al., 2020 ); however, as depicted in Figure 1 E, the relationship between the magnetic damping and the SOT-induced in-plane damping-like torque in a PMA-FM is not as symmetrical as the case of STT shown in Figure 1 B; hence, challenges and new opportunities coexist in proposing such an artificial neuron based on the PMA SHNO ( Fulara et al., 2019 ) in the absence of an in-plane magnetic field.…”

Multistate Magnetic Domain Wall Devices for Neuromorphic Computing

“…It is reported that there is no need to switch back and forth the free layer magnetization and thus the recovery step during information processing is faster (%600 ps) compared to a few nanoseconds in the case of switching. [49] Other than MTJ devices, a spin Hall nano-oscillator based on NiFe nanostructures was proposed. [50] The mutual synchronization of a large number of nano-oscillators showed a large signal quality factor compared to MTJ-based devices.…”