A MATE-GDBF Algorithm for Irregular Punctured LDPC Codes and Its Decoder Implementation

This paper proposes a MATE-GDBF decoding algorithm and a corresponding decoder architecture for punctured LDPC codes. The proposed algorithm employs majority-vote and tabu-enhancement approaches to improve decoding performance and convergence speed. Additionally, the decoder architecture includes a new flipping access scheme designed to reduce power consumption. The proposed LDPC (9344, 64, 8384) decoder is implemented in TSMC 90nm CMOS process and achieves a 6.39 Gbps throughput at a clock rate of 400 MHz. The decoder's core area is 0.67 mm
²
, and its power consumption is 79.6 mW. Overall, the proposed algorithm and decoder architecture offer a promising solution for improving the performance and efficiency of punctured LDPC codes.