A thresholding algorithm for improved Split-Row decoding of LDPC codes

The recently proposed split-row decoding algorithm provides significant improvements in the throughput, hardware efficiency and energy efficiency when compared to existing soft decision decoding algorithms at the cost of some error performance loss. In this paper we propose split-row threshold which outperforms the split-row algorithm while maintaining the same level of complexity. Simulation results show that the algorithm provides 0.15-0.2 dB coding gain over split-row decoding and is within 0.15-0.2 dB of SPA and MinSum normalized.