An ideal nanoscale memristor crossbar array can naturally carry out vector-matrix multiplication-a computationally expensive task for many important applications-in a single constant time step [2]. By ...