Finite-Length Scaling of SC-LDPC Codes With a Limited Number of Decoding Iterations

We propose four finite-length scaling laws to predict the frame error rate (FER) performance in waterfall region of spatially-coupled low-density parity-check code ensembles under full belief propagation (BP) decoding with a limit on number iterations and law for sliding window decoding, also limited iterations. The BP provide choice between accuracy computational complexity; good balance them is achieved by that models decoded bits after certain time-integrated Ornstein-Uhlenbeck process. This framework developed further model as race integrated process an absorbing barrier corresponds left boundary window. proposed yield accurate FER predictions semi-structured Olmos Urbanke.