Analog Matrix Computing with Resistive Memory: Circuits and Theory

In this talk, we present analog matrix computing (AMC) circuits based on crosspoint resistive memory arrays, together with the underlying theories of the circuits and the AMC concept. Four types of AMC circuits will be introduced, which realize basic matrix operations in one step, including matrix multiplication, matrix inversion, pseudoinverse and eigenvector computations. We describe the main design principles of the AMC circuits, such as local/global or negative/positive feedback configurations, with/without external inputs. Mapping strategies for matrices containing negative values will also be included. The stability issue of the circuits will be described via the transfer function analysis, which also defines time complexities of the circuits.