Search

Stochastic computing (SC) is a promising technology that can be used for low-cost hardware designs. However, SC suffers from its long latency. Although parallel processing can efficiently shorten the latency, duplicated stochastic number generators (SNGs) are necessary, which cause substantial hardware overhead. This paper proposes a scalable SNG based on the spin-Hall-effect (SHE), which is capable of generating multiple independent stochastic streams simultaneously. The design takes advantages of the efficient charge-to-spin conversion from the Spin-Hall material and the intrinsic stochasticity of nanomagnets. Compared to previous spintronic SNGs, the SHE-SNG can reduce the area by $1.6\times -7.8\times $ and the power by $4.9\times -13\times $ while increasing the degree of parallelism from 1 to 16. Compared to CMOS-based SNGs, the proposed SNG obtained $24\times -120\times $ and $53\times $ reduction in terms of area and power, respectively. Finally, three benchmarks were implemented, and the results indicate that SC implementations with the proposed SHE-SNG can achieve $1.2\times -29\times $ reduction of hardware resources compared to implementations with previous CMOS- and spintronic-based designs while scaling the degree of parallelism from 1 to 64. [ABSTRACT FROM AUTHOR]

An ac small signal model for lateral spin valves (LSVs) is developed, and is applied to two different types of devices: spin transmitters and magnetic field sensors. The small signal models explicitly decouple the charge- and spin-dependent components, which describe the intrinsic RC delay and the ac spin accumulation dynamics, separately. The analysis is applied to graphene LSVs as an example to illustrate the optimization strategies for LSV devices. By using the charge-spin circuit model, the scattering parameters and the reflection coefficients at the output terminals are quantitatively calculated as a function of frequency. This model can be used as a guide for future LSV design and to benchmark performance across different material systems. [ABSTRACT FROM AUTHOR]

Graphene is an ideal material for spintronic devices due to its low spin-orbit coupling and high mobility. One of the most important potential applications of graphene spintronics is for use in neuromorphic computing systems, where the tunable spin resistance of graphene can be used to apply analog weighting factors. A key capability needed to achieve spin-based neuromorphic computing systems is to achieve distinct regions of control, where injected and detected spin currents can be tuned independently. Here, we demonstrate the ability to achieve such independent control using a graphene spin valve geometry where the injector and detector regions are modulated by two separate bottom gate electrodes. The spin transport parameters and their dependence on each gate voltage are extracted from Hanle precession measurements. From this analysis, local spin transport parameters and their dependence on the local gate voltage are found, which provide a basis for a spatially-resolved spin resistance network that simulates the device. The data and model are used to calculate the spin currents flowing into, through, and out of the graphene channel. We show that the spin current flowing through the graphene channel can be modulated by 30% using one gate and that the spin current absorbed by the detector can be modulated by 50% using the other gate. This result demonstrates that spin currents can be controlled by locally tuning the spin resistance of graphene. The integration of chemical vapor deposition (CVD) grown graphene with local gates allows for the implementation of large-scale integrated spin-based circuits. [ABSTRACT FROM AUTHOR]