Random Bitstream Generation Using Voltage-Controlled Magnetic Anisotropy and Spin Orbit Torque Magnetic Tunnel Junctions

Probabilistic computing using random number generators (RNGs) can leverage the inherent stochasticity of nanodevices for system-level benefits. Device candidates this application need to produce highly “coinflips” while also having tunable biasing coin. The magnetic tunnel junction (MTJ) has been studied as an RNG due its thermally-driven magnetization dynamics, often spin transfer torque (STT) current amplitude control switching MTJ free layer magnetization, here called stochastic write method. There are additional knobs MTJ-RNG, including voltage-controlled anisotropy (VCMA) and orbit (SOT), there is more systematically study compare these methods. We build analytical model characterize VCMA SOT generate bit streams. results show that both methods high quality, uniformly distributed bitstreams. Biasing bitstreams either STT or applied field shows a sigmoidal distribution vs. bias SOT, compared less write. energy consumption per sample calculated be 0.1 pJ 1 (stochastic write), 20 (VCMA), revealing potential benefit showing may require higher damping materials. generated then two tasks: generating arbitrary probability MTJ-RNGs neurons perform simulated annealing, where ability effectively minimize system with small delay low energy. These flexibility true elucidate design parameters optimizing device operation applications.