An Efficient High-Rate Non-Binary LDPC Decoder Architecture With Early Termination

Abstract

This paper presents a modified Trellis Min-Max (T-MM) algorithm together with the associated architecture for non-binary (NB) low-density parity-check (LDPC) decoders. The proposed T-MM algorithm is able to reduce the memory requirements for the check-node messages through an efficient compression method and enhance the error-rate performance using the appropriate decompression. A method of updating the a posteriori log-likelihood ratio in the delta domain is used to simplify the computational and storage complexity. In order to enhance the decoding throughput, a low-complexity early termination (ET) scheme is devised by using the hard decisions of the variable-to-check messages, where, although a minor overhead is introduced, there is no visible degradation in error rate. As a proof of concept, a row-parallel layered decoder for the 32-ary (837, 726) LDPC code is implemented using a 90-nm CMOS process. The proposed decoder achieves a throughput of 1.64 Gb/s at 526.32 MHz based on eight iterations and has an area of 6.86 mm 2. When the ET scheme is enabled, the decoder achieves a maximum throughput of 4.68 Gb/s with a frame error rate of 3.25 × 10 −6 at E b /N 0 = 4.5 dB. The proposed NB-LDPC decoder achieves the highest throughput and hardware efficiency compared to the state-of-the-art decoders, even when the ET scheme is not enabled. INDEX TERMS Non-binary low-density parity-check (NB-LDPC) codes, trellis min-max (T-MM) algorithm, layered decoding, early termination (ET), high-throughput decoder, very large scale integration (VLSI) architecture. YEONG-LUH UENG (M'05-SM'15) received the Ph.D. degree in communication engineering from

An Effective Multi-Mode Iris Authentication System on a Microprocessor-FPGA Heterogeneous Platform With QC-LDPC Codes

“…Implementing LDPC codes, especially QC-LDPC codes onto FPGAs or application-specific integrated circuits (ASICs), are very common today. Comprehensive studies [11]- [13] have been conducted to reduce the complexity and improve the accuracy and throughput of the QC-LDPC decoders, such as applying appropriate calculations [14]- [17] and different structures of LDPC codes [18]- [20]. All the aforementioned studies adopt FPGAs or ASICs as their testing platforms in order to parallel and pipeline their decoding algorithm to the best extent.…”

An Efficient NB-LDPC Decoder Architecture for Space Telecommand Links

“…In comparison with other alternatives, such as BP with Fast Fourier Transform (BP-FFT) [7] or extended min-sum (EMS) [8], it has the least decoding complexity with very small loss of performance. Trellis versions of the algorithms [9]- [11] allow a higher degree of parallelism, targeting throughputs far above the ones used in telecommand links. Space telecommand rates vary from hundreds of kbps to 2 Mbps, which is the upper limit recommended by the CCSDS [12].…”

An Efficient Short High-Order Non-Binary LDPC Decoder Architecture Using a Message-Adaptation EMS Algorithm

“…It can be expected that a full-syndrome check at the end of each layered processing will achieve a lower average number of iterations compared to a full-syndrome check at the end of each iteration. In order to reduce the complexity of the ET, the authors of [19] proposed calculating s m rather than s at the end of processing layer m. If s m = 0, layer m is valid. Since the LLR messages might be subsequently updated, this means that any previously valid check node might become invalid during the current layered processing task.…”

“…Since the LLR messages might be subsequently updated, this means that any previously valid check node might become invalid during the current layered processing task. Consequently, the authors of [19] proposed terminating the decoding process only if τ ET consecutive layers are valid. The threshold τ ET should be sufficiently large so as to avoid any degradation in error rate.…”

“…Table 2 shows the hardware performance for the proposed decoder, which can achieve a maximum throughput of 191.14 Mbps when the ET scheme is omitted, and a throughput of 322.92 Mbps when the ET scheme at SNR = 3 dB is included. Also contained in Table 2 is the hardware performance for the baseline design, where the truncated EMS algorithm and the backward memory scan method presented in [25] are realized, together with the ET scheme proposed in [19]. Fig.…”