Nonvolatile Spintronic Memory Cells for Neural Networks
This work addresses energy efficiency for neural network hardware, but it appears incremental as it builds on existing spintronic and neural computing concepts.
The authors tackled the problem of energy consumption in neural networks by proposing a new spintronic nonvolatile memory cell and dual-circuit architecture, resulting in a performance of about 100 pJ per image processed, outperforming charge-based implementations.
A new spintronic nonvolatile memory cell analogous to 1T DRAM with non-destructive read is proposed. The cells can be used as neural computing units. A dual-circuit neural network architecture is proposed to leverage these devices against the complex operations involved in convolutional networks. Simulations based on HSPICE and Matlab were performed to study the performance of this architecture when classifying images as well as the effect of varying the size and stability of the nanomagnets. The spintronic cells outperform a purely charge-based implementation of the same network, consuming about 100 pJ total per image processed.