Efficient Reversible Logic Synthesis via Isomorphic Subgraph Matching

Reversible logic is being studied extensively due to its applications in the design of nanoscale circuits with ultra-low power dissipation and future technologies such as quantum computing. Most methods that synthesize reversible logic circuits are restricted to small functions and are thus, not scalable. Previous methods based on Binary Decision Diagrams (BDD) are scalable but introduce a large number of additional circuit lines. In this paper, we present a technique that maps subgraphs of the BDD to structures with known reversible-circuit templates by posing this as a subgraph isomorphism problem. Experimental results show that this approach reduces the number of additional circuit lines introduced by previous approaches while retaining their scalability. We further show that for small functions, there is also an improvement in the quantum cost.