NUS Spintronic Processors Deliver 3.2-Fold Speedup and 58% Energy Savings for Optimization Tasks
Researchers at the National University of Singapore have developed probabilistic spintronic processors built from magnetic tunnel junctions. These processors achieved a 3.2-fold speedup and 58.3 percent energy savings compared with conventional CPUs on quadratic assignment problems. The hardware also outperformed commercial D-Wave quantum annealers by consistently returning feasible, high-quality solutions as problem size increased. Cluster parallel updates and simulated quantum annealing techniques delivered further gains of up to 10-fold acceleration and 20-fold better solution quality.
July 2, 2026