ABSTRACT

In this paper, a digital neural networks (DNN) implementation approach based on complex complementary pass-transistor logic (C2PL) is presented. Some types of 3–2 compressor designs in C2PL are discussed and a number of experiments are conducted to optimize their ‘performance. Based on the 3–2 compressors, two typical building blocks in DNN implementation, i.e., 4–2 and 7–3 compressor, are developed. Example of the application is given to illustrate the effectiveness of the approach.