Automated Synthesis of Computing Nanoscale Crossbars using Formal Methods

Since the fabrication of nanoscale memristors by HP Labs in 2008, there has been a renewed interest in the use of crossbars of nanoscale memristors as digital storage and neuromorphic computing devices. However, the same success has not been replicated in the use of crossbars for performing general-purpose computations that can support the existing software infrastructure originally designed for Von Neumann architectures. One of the key challenges facing this technology is the existence of sneak paths. While it has been shown that sneak paths can be used to perform Boolean computations in crossbars, the human mind is not particularly suited to reason about the exponential complexity of sneak paths in crossbars. Hence, the size of the crossbar designs proposed in the literature has been large for practical applications. In this paper, we demonstrate how formal methods can be used to automatically synthesize compact crossbar designs that can be used to evaluate Boolean formula by using the sneak paths phenomena as a design primitive. We show that our automated synthesis technique can be used to generate a state-of-the-art nano-crossbar design for a 1-bit full adder.