Growth of factor-graph width under sequential unrolling

Characterize how the induced width of the gate-level factor graph for a sequential locked design grows with the depth of its fixed-depth unrolling.

Background

The paper’s exact counting machinery is developed for combinational, acyclic locked netlists. It observes that a sequential locked design unrolled to a fixed depth becomes a combinational factor graph with state variables shared between successive frames, allowing the same counting framework to be applied.

However, the computational cost of sum–product inference depends exponentially on induced factor-graph width. The dependence of that width on unrolling depth is not established, leaving open the scalability of the approach to sequential locking.