A modified belief-propagation decoder for the parallel decoding of product codes

摘要

In this dissertation a modification to the belief-propagation algorithm is presented.udThe algorithm modifies the belief-propagation algorithm to allow for the paralleluddecoding of product codes. The algorithm leverages the fact that each componentudcode in the product code can be independently decoded because the codewords areudencoded by independent and identically distributed (i.i.d.) processes. The algorithmudmaximises the parellelisation by decoding all the component codes in each dimensionudin parallel. In order to facilitate this process we developed new additional stagesudwhich are added to the belief-propagation algorithm: the codeword reliability estimation,udthe belief-aggregation and the exit test stages. The parallel product codeuddecoder o ers a 0.2 dB worsening of the decoding BER performance when comparedudto the best serial decoder. However, the parallel belief-propagation decoderudo ers a 7.26 time speedup on an eight-core processor, which is 0.91 of the theoreticaludmaximum of eight for an eight-core processor.