Understanding the trade-offs of device, circuit and application in ReRAM-based neuromorphic computing systems

Resistive memory (ReRAM) features nonvolatile storage, high resistance, dense structure, and analogy to the matrix-vector multiplication operation. These characteristics demonstrate the great potential of ReRAM in the development of neuromorphic computing systems. In this paper, we show the importance of the comprehensive understanding across the device, circuit, and application levels in ReRAM-based neuromorphic system, through the discussion of three major problems-weight mapping, reliability, and system integration.