LAUSR

Low-Density Parity-Check (LDPC) code is a type of forward error-correction code with excellent performance, and has been widely used in many modern communication standards. The second-generation satellite broadcasting standard (DVB-S2) and its extension (DVB-S2X) adopt a special Irregular Repeated Accumulate (IRA) LDPC code as inner coding scheme. However, due to the large block size, most of the architectures proposed so far use Random Access Memory (RAM) to store and update the encoding results, and the delay caused by address-controlled read and write operations and barrel shift during computation inevitably limits the upper bound of encoder throughput. In this paper, by extracting the periodicity of the parity-check matrix, we introduce a fast encoding algorithm that can efficiently process the multiplication of the information sequence and a large-dimensional sparse matrix, and propose an encoder architecture with low encoding delay and high throughput. The proposed architecture has been implemented and tested on a Xilinx Kintex-7 FPGA, and the result show that the encoder architecture can achieve the highest throughput of 47.5 Gbps at a clock frequency of 280 MHz.