Superparamagnetic Tunnel Junctions for Reliable True Randomness
Authors:
Dooyong Koh,
Qiuyuan Wang,
Brooke C. McGoldrick,
Luqiao Liu,
Marc A. Baldo
Abstract:
Stochastic devices have the potential to disrupt computing, revolutionizing low-power machine learning acceleration, probabilistic computing, and hardware security. As implemented, however, superparamagnetic tunnel junctions (sMTJs) face significant challenges including the need for external magnetic fields, and poor reliability and scalability. Here, we present experimental demonstration of three…
▽ More
Stochastic devices have the potential to disrupt computing, revolutionizing low-power machine learning acceleration, probabilistic computing, and hardware security. As implemented, however, superparamagnetic tunnel junctions (sMTJs) face significant challenges including the need for external magnetic fields, and poor reliability and scalability. Here, we present experimental demonstration of three-terminal sMTJs as scalable and reliable sources of true randomness under a field-free regime. By leveraging dual-current controllability and incorporating feedback systems, we substantially enhance the stability and reliability of sMTJ-based systems under varying conditions, even in the field-free regime. Our findings demonstrate the generation of cryptographic-quality random bitstreams and the practical use of sMTJs as efficient and reliable random number generators, successfully integrated into advanced computing algorithms like generative artificial intelligence. Field-free, truly random sMTJs promise to address critical challenges in cryptography, edge computing, and beyond, significantly advancing the field of random number generation.
△ Less
Submitted 11 July, 2024;
originally announced July 2024.
Magnetic domain wall based synaptic and activation function generator for neuromorphic accelerators
Authors:
Saima A Siddiqui,
Sumit Dutta,
Astera Tang,
Luqiao Liu,
Caroline A Ross,
Marc A Baldo
Abstract:
Magnetic domain walls are information tokens in both logic and memory devices, and hold particular interest in applications such as neuromorphic accelerators that combine logic in memory. Here, we show that devices based on the electrical manipulation of magnetic domain walls are capable of implementing linear, as well as programmable nonlinear, functions. Unlike other approaches, domain-wall-base…
▽ More
Magnetic domain walls are information tokens in both logic and memory devices, and hold particular interest in applications such as neuromorphic accelerators that combine logic in memory. Here, we show that devices based on the electrical manipulation of magnetic domain walls are capable of implementing linear, as well as programmable nonlinear, functions. Unlike other approaches, domain-wall-based devices are ideal for application to both synaptic weight generators and thresholding in deep neural networks. Prototype micrometer-size devices operate with 8 ns current pulses and the energy consumption required for weight modulation is < 16 pJ. Both speed and energy consumption compare favorably to other synaptic nonvolatile devices, with the expected energy dissipation for scaled 20 nm devices close to that of biological neurons.
△ Less
Submitted 7 September, 2019;
originally announced September 2019.
