An Output-Controllable Stochastic MTJ Spin Based Ising Machine for Solving Combinatorial Optimization Problems

Combinatorial optimization problems (COPs) have long been challenging for traditional computers. We have design an Ising annealing machine to solve COPs based on an output-controllable magnetic tunnel junctions (MTJ) based spins (MTJ-based spins) and an efficient annealing algorithm called QFactor. Our design has the following advantages: (1) enabling operation at room temperature and offering excellent scalability and manufacturability, (2) MTJ-based spins demonstrating robust thermal stability, true randomness and high controllability, (3) annealing much faster than traditional annealing methods.