Stochasticity in the Switching of Nanodisks for Probabilistic Computing

Abstract

Stochasticity in magnetic nanodevices is an essential characteristic for harnessing those devices to computing based on population coding or the building blocks of probabilistic computing, pbits. A magnetic tunneling junction (MTJ) consisting of a patterned magnetic element is considered as a promising computing unit in the concept of artificial neurons and p-bits. The comprehensive understanding of the stochasticity in the switching of patterned magnetic elements is crucial for realizing MTJ based probabilistic computing technology. In the present work, the stochastic behavior in switching process of a perpendicularly magnetized Co/Pt disk within an array was directly observed utilizing full-field soft X-ray microscopy. Within 50 repeated hysteretic cycles, the stochastic magnetization switching of individual Co/Pt disks within disk arrays is identified. We found that the stochasticity in the magnetization switching of disks considerably depends on the disk size. The stochasticity initially decreases as the disk radius gets bigger from 125 nm to 375 nm (Region I), then increases with further enlarging disk size to 625 nm (Region II). The variance of thermal fluctuation relevant to the disk size and the multi-level switching within a disk are severely involved in the observed size dependent stochasticity. This work provides the way for controlling the stochasticity in the switching of nanopatterned elements, which is a key aspect for MTJ-based probabilistic computing.

Three-dimensional magnetic nanotextures with high-order vorticity in soft magnetic wireframes

“…To this end, we show that a pyramid-shaped wireframe can support six antivortices without vortices, while a cube-shaped structure allows to stabilize eight surface antivortices without vortices. Because of the topological stability of solitons of similar kind, even upon interaction they are stable against annihilation, which suggests their relevance for applications in low-power computation schemes utilizing paradigms of reservoir 3 , 4 and probabilistic 21 , 22 computing.

Fig.

…”

Efficient Probabilistic Computing with Stochastic Perovskite Nickelates

“…P-circuits have also been demonstrated to accelerate hard optimization problems − and quantum Monte Carlo algorithms. , Probabilistic bits have previously been implemented in existing CMOS technology. However, these implementations suffer from costly (pseudo) random number generators requiring thousands of transistors − and poor quality of randomness, increasing energy and area costs for stochastic circuits. Even when the most advanced technology nodes are used, the ultimate scalability of standard digital CMOS-based p-bits is limited.…”