Abstract
Two versions of a new parallel symbol-flipping decoding algorithm for non-binary low-density parity-check (NB-LDPC) codes are proposed. Simulation results show that the soft-decision parallel symbol-flipping decoding outperforms quite a number of existing reliability-based message-passing algorithms. It provides an effective trade-off between error performance anddecoding complexity compared with the q-ary sum-product algorithm. The algorithm can be simplified to hard-decision parallel symbol-flipping decoding for applications in communication or storage systems where either soft information is not available or a simple decoder is needed. They are particularly effective for decoding NB-LDPC codes whose parity-check matrices have large row/column weights.