A Data Non-destructive IMPLY-based Memristive Semi-parallel Full-Adder for Computing-in-memory Systems

This paper presents a data non-destructive IMPLY-based memristive semi-parallel full-adder (FA) for Computing-in-memory system, which solves the issue that the conventional semi-parallel FA cannot protect the input operands. The proposed full adder not only reduce the operating time by executing more steps in parallel, and also prevent the input operands from being destructed by adding an extra memristor with only one additional step. Simulation results show that the proposed full adder can realize 18% reduction on operation time against the conventional design, while protect the input operand from destruction, which is an efficient design for data reuse in neural network acceleration.