An Integrated Logic Function Decomposition Flow for XOR-Majority Graphs

Logic function decomposition is critical to logic synthesis. In this paper, we propose an integrated logic function decomposition flow, which consists of disjoint support decomposition, exact synthesis, Shannon decomposition, and the recently proposed majority-of-three (MAJ) decomposition. Through this process, the logic function can be decomposed into any directed acyclic graph (DAG). An experimental evaluation shows that our algorithm leads to better XOR-majority graphs (XMGs) compared to state-of-the-art algorithms. Our experiments show that the proposed method achieves an average of 5% and 6% reduction in the lookup tables (LUTs) size/depth product applied to the EPFL arithmetic and random control benchmarks after LUT mapping, respectively.