Stephan, Andrew2022-09-262022-09-262020-07https://hdl.handle.net/11299/241706University of Minnesota Ph.D. dissertation.July 2020. Major: Electrical Engineering. Advisor: Steven Koester. 1 computer file (PDF); ix, 142 pages.Neuromorphic and spintronic computing are two different technologies that have progressed significantly in recent years. The former has seen advancement in algorithms and hardware implementations that solve previously intractable problems in data processing and classification, while the latter has seen the advent of an entirely new class of physical computing primitives. In particular, spin-torque-switched magnetic tunnel junctions and magnetelectric-ferromagnetic heterostructures are promising candidates for neuromimetic devices. Using physics-based model simulations of SHE-MTJ and ME-FM circuits, I predict the performance of various spintronic neural network formulations on standard benchmarking tasks and show that circuits like these are likely to be the neuromorphic hardware of the future.enDevice DesignNeuromorphicsSimulationSpintronicsThesis or Dissertation