A Variation-Aware and Energy-Efficient Spintronic True Random Number Generator

This paper proposes a spintronic true random number generator (TRNG) optimized for low power, reliability, and sensitivity to variations. It uses the random behavior of magnetic tunnel junctions in a low critical current switching regime for generating numbers. An internal post-processing unit enhances randomness without major hardware costs. It's proven robust against process, voltage, and temperature changes via Monte-Carlo simulations, meeting the National Institute of Standards and Technology randomness criteria even with these variations. The proposed TRNG also shows a 38% energy efficiency improvement over its state-of-the-art counterparts.