A Predictive Approach for Conditional Execution of Memristive Material Implication Stateful Logic Operations

Memristive devices are promising building blocks for designing beyond von-Neumann computing applications. In this scenario, memristive material implication is one of the most notable techniques for implementing stateful logic circuits. One of the major drawbacks of this approach is the need to perform lengthy sequences of material implications to synthesize a given Boolean function. However, we show that many of these operations are ineffective and can be dismissed. Based on this assumption, this paper proposes a new implementation method to reduce the length of the sequences. Our proposal uses an auxiliary circuit to predict the necessity to perform each material implication, whose execution is conditioned to its effectiveness. We present the designed circuit and demonstrate its operation through electrical simulations. This work also discusses the results obtained from the analysis of several benchmark functions, in which we observed a reduction of 50– 90% in the time spent performing this kind of computing.