Pedram Khalili: Computing with voltage-controlled magnetic tunnel junctions (Invited)

2023 IEEE AtC-AtG Magnetics Conference Session 9
Pedram Khalili, Northwestern University, USA
For more information, please visit the website below.
www.atc-atg.or...
0:00 Speaker Introduction
0:02:01 Introduction
0:04:00 Outline
0:05:50 Energy Challenges in Electronics
0:07:50 Magnetic Random-Access Memory
0:11:04 Voltage-Controlled Magnetic Anisotropy (VCMA)
0:12:32 VCMA Switching
0:14:33 Device Structure
0:15:13 J is larger than T regime. Evolution with B
0:15:45 Film-level TMR results
0:16:35 Voltage-controlled switching of MTJ (70 nm)
0:17:30 Voltage-controlled switching of MTJ (50 nm)
0:17:35 Voltage-controlled switching of MTJ (30 nm)
0:18:13 VCMA coefficients measurements
0:20:33 Summary of high-VCMA MRAM results
0:21:10 Machine Learning (ML) at the edge
0:21:51 Stochastic Computing
0:23:17 True tunable random number generators (TRNGs)
0:24:55 Implementation of Artificial Neural Networks (ANNs)
0:25:16 Inference on MNIST dataset and results
0:28:40 Ultrafast TRNG using VC-MRAM bits
0:31:57 Probabilistic Computing: Implementation using VC-MRAM
0:34:34 P-bit Design based on Conventional CMOS
0:35:33 Implementation of p-bits with 50% VCMA-TRNGs
0:35:59 VCMA-based Invertible Logic Gates & Multipliers/Factorizers
0:38:37 VCMA-based Reconfigurable PUF for Secure Chips
0:38:49 Summary
0:39:25 Q&A