Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

China Communications

GPU-based non-binary LDPC decoder with weighted bit-reliability based algorithm

Zhanxian Liu, Rongke Liu, Ling Zhao

Published May 1, 2020• 2 citations
Publisher

Abstract

In this paper, we present a graphics processing unit (GPU)-based implementation of a weighted bit-reliability based (wBRB) decoder for non-binary LDPC (NB-LDPC) codes. To achieve coalesced memory accesses, an efficient data structure for the wBRB algorithm is proposed. Based on the Single-Instruction Multiple-Threads (SIMT) programming model, a novel mapping strategy with high intra-frame parallelism is presented to improve the latency and throughput performance. Moreover, by using Single-Instruction Multiple-Data (SIMD) intrinsics, four 8-bit message elements are packed into a 32-bit unit and simultaneously processed. Experimental results show that the proposed wBRB decoder provides good tradeoff between error performance and throughput for the codes with relatively large column degrees or high rates.

DOI
10.23919/jcc.2020.05.008

Related papers

Bit Reliability-Based Decoders for Non-Binary LDPC Codes

Qin Huang, Shuai Yuan

Message-passing decoders typically perform well for nonbinary low-density parity-check (NB-LDPC) codes with large computational complexity. As another type of simplified decoders, symbol-reliability-based decoders further reduce the computational complexity. However, the previously proposed algorithms suffer severe error performance degradation for NB-LDPC codes with low column weights. In this paper, a weighted bit-reliability based (wBRB) decoder for NB-LDPC codes is developed and implemented with efficient layered partial-parallel structure. It not only balances the tradeoff between complexity and error performance, but also reduces the memory usage significantly. Furthermore, to enhance the performance of the wBRB decoder, a full bit-reliability-based (FBRB) decoder is proposed. The FBRB decoder is derived based on the binary matrix representation of the nonzero entries in the parity-check matrix. Since more bit-reliability values are passed through the edges of the Tanner graph, the FBRB decoder can achieve better error performance and faster convergence rate than the wBRB decoder. Both of the decoders are implemented on a Xilinx Virtex-5 XC5VLX155T FPGA device for a (403,226) code over GF(25). The results shows that they achieve 118.98 and 95.73 Mbps throughput with 15 iterations, respectively.

IEEE Transactions on Communications•Nov 17, 2015•26 citations

Multi-Gbps LDPC Decoder on GPU Devices

Jingxin Dai, Hang Yin, Yansong Lv +2 more

To meet the high throughput requirement of communication systems, the design of high-throughput low-density parity-check (LDPC) decoders has attracted significant attention. This paper proposes a high-throughput GPU-based LDPC decoder, aiming at the large-scale data process scenario, which optimizes the decoder from the perspectives of the decoding parallelism and data scheduling strategy, respectively. For decoding parallelism, the intra-codeword parallelism is fully exploited by combining the characteristics of the flooding-based decoding algorithm and GPU programming model, and the inter-codeword parallelism is improved using the single-instruction multiple-data (SIMD) instructions. For the data scheduling strategy, the utilization of off-chip memory is optimized to satisfy the demands of large-scale data processing. The experimental results demonstrate that the decoder achieves 10 Gbps throughput by incorporating the early termination mechanism on general-purpose GPU (GPGPU) devices and can also achieve a high-throughput and high-power-efficiency performance on low-power embedded GPU (EGPU) devices. Compared with the state-of-the-art work, the proposed decoder had a ×1.787 normalized throughput speedup at the same error correcting performance.

Electronics•Oct 25, 2022•11 citations

Sequential decoding of non-binary LDPC codes on graphics processing units

David L. Romero, Nicholas B. Chang

Non-binary low-density parity-check (LDPC) codes have been shown to attain near capacity error correcting performance in noisy wireless communication channels. It is well known that these codes require a very large number of operations per-bit to decode. This high computational complexity along with a parallel decoder structure makes graphics processing units (GPUs) an attractive platform for acceleration of the decoding algorithm. The seemingly random memory access patterns associated with decoding are generally beneficial to error-correcting performance but present a challenge to designers who want to leverage the computational capabilities of the GPU. In this paper we describe the design of an efficient decoder implementation based on GPUs and a corresponding set of powerful non-binary LDPC codes. Using the belief propagation algorithm with a sequential message updating scheme it is shown that we are able to exploit parallelism inherent in the decoding algorithm while decreasing the number of decoding iterations required for convergence.

Asilomar Conference on Signals, Systems and Computers (ACSSC)•Nov 1, 2012•11 citations

A Scalable LDPC Decoder on GPU

Kiran Kumar Abburi

A flexible and scalable approach for LDPC decoding on CUDA based Graphics Processing Unit (GPU) is presented in this paper. Layered decoding is a popular method for LDPC decoding and is known for its fast convergence. However, efficient implementation of the layered decoding algorithm on GPU is challenging due to the limited amount of data-parallelism available in this algorithm. To overcome this problem, a kernel execution configuration that can decode multiple codewords simultaneously on GPU is developed. This paper proposes a compact data packing scheme to reduce the number of global memory accesses and parity-check matrix representation to reduce constant memory latency. Global memory bandwidth efficiency is improved by coalescing simultaneous memory accesses of threads in a half-warp into a single memory transaction. Asynchronous data transfers are used to hide host memory latency by overlapping kernel execution with data transfers between CPU and GPU. The proposed implementation of LDPC decoder on GPU performs two orders of magnitude faster than the LDPC decoder on a CPU and four times faster than the previously reported LDPC decoder on GPU. This implementation achieves a throughput of 160Mbps, which is comparable to dedicated hardware solutions.

International Conference on VLSI Design•Jan 1, 2011•24 citations

Implementation of Decoders for LDPC Block Codes and LDPC Convolutional Codes Based on GPUs

Yue Zhao, Francis C. M. Lau

In this paper, efficient LDPC block-code decoders/simulators which run on graphics processing units (GPUs) are proposed. We also implement the decoder for the LDPC convolutional code (LDPCCC). The LDPCCC is derived from a predesigned quasi-cyclic LDPC block code with good error performance. Compared to the decoder based on the randomly constructed LDPCCC code, the complexity of the proposed LDPCCC decoder is reduced due to the periodicity of the derived LDPCCC and the properties of the quasicyclic structure. In our proposed decoder architecture, Γ (Γ is a multiple of a warp) codewords are decoded together, and hence, the messages of Γ codewords are also processed together. Since all the Γ codewords share the same Tanner graph, messages of the Γ distinct codewords corresponding to the same edge can be grouped into one package and stored linearly. By optimizing the data structures of the messages used in the decoding process, both the read and write processes can be performed in a highly parallel manner by the GPUs. In addition, a thread hierarchy minimizing the divergence of the threads is deployed, and it can maximize the efficiency of the parallel execution. With the use of a large number of cores in the GPU to perform the simple computations simultaneously, our GPU-based LDPC decoder can obtain hundreds of times speedup compared with a serial CPU-based simulator and over 40 times speedup compared with an eight-thread CPU-based simulator.

arXiv:1204.0334•IEEE Transactions on Parallel and Distributed Systems•Oct 28, 2013•34 citations

A 21.66 Gbps Nonbinary LDPC Decoder for High-Speed Communications

Jing Tian, Jun Ting Lin, Zhongfeng Wang

Compared to binary Low-Density Parity-Check (LDPC) codes, nonbinary LDPC (NB-LDPC) codes have better error correction performance under short-to-moderate block lengths or high-order modulations. Traditional min-sum-based soft decoding algorithms for NB-LDPC codes suffer from large computational complexity, which leads to inefficient hardware implementations. The Multiple-symbol-reliability weighted Bit-Reliability-Based (MwBRB) hard decoding algorithm achieves a good tradeoff between error correction performance and decoding complexity. However, efficient hardware implementations based on the MwBRB algorithm have not been investigated. In this brief, an improved layered MwBRB algorithm is first proposed, which results in faster convergence rate than the MwBRB algorithm. Then, an ultra-high-throughput low-complexity decoder architecture with an efficient partially parallel processing schedule is also presented. Finally, the proposed architecture is coded with RTL and synthesized under the TSMC 90-nm CMOS technology. The synthesis results demonstrate that the proposed decoder for a (837, 726) quasi-cyclic NB-LDPC code over GF(25) achieves a throughput of 21.66 Gbps and an area efficiency of 4.77 Gbps/M-gates under the TSMC 90-nm CMOS technology. The proposed decoder reaches a throughput more than 20 Gbps for the first time among the prior NB-LDPC decoders, and the area efficiency is far beyond the state-of-the-art designs.

IEEE Transactions on Circuits & Systems II Express Briefs•May 19, 2017•10 citations
In this paper, we present a graphics processing unit (GPU)-based implementation of a weighted bit-reliability based (wBRB) decoder for non-binary LDPC (NB-LDPC) codes. To achieve coalesced memory accesses, an efficient data structure for the wBRB algorithm is proposed. Based on the Single-Instruction Multiple-Threads (SIMT) programming model, a novel mapping strategy with high intra-frame parallelism is presented to improve the latency and throughput performance. Moreover, by using Single-Instruction Multiple-Data (SIMD) intrinsics, four 8-bit message elements are packed into a 32-bit unit and simultaneously processed. Experimental results show that the proposed wBRB decoder provides good tradeoff between error performance and throughput for the codes with relatively large column degrees or high rates.
Message-passing decoders typically perform well for nonbinary low-density parity-check (NB-LDPC) codes with large computational complexity. As another type of simplified decoders, symbol-reliability-based decoders further reduce the computational complexity. However, the previously proposed algorithms suffer severe error performance degradation for NB-LDPC codes with low column weights. In this paper, a weighted bit-reliability based (wBRB) decoder for NB-LDPC codes is developed and implemented with efficient layered partial-parallel structure. It not only balances the tradeoff between complexity and error performance, but also reduces the memory usage significantly. Furthermore, to enhance the performance of the wBRB decoder, a full bit-reliability-based (FBRB) decoder is proposed. The FBRB decoder is derived based on the binary matrix representation of the nonzero entries in the parity-check matrix. Since more bit-reliability values are passed through the edges of the Tanner graph, the FBRB decoder can achieve better error performance and faster convergence rate than the wBRB decoder. Both of the decoders are implemented on a Xilinx Virtex-5 XC5VLX155T FPGA device for a (403,226) code over GF(25). The results shows that they achieve 118.98 and 95.73 Mbps throughput with 15 iterations, respectively.
To meet the high throughput requirement of communication systems, the design of high-throughput low-density parity-check (LDPC) decoders has attracted significant attention. This paper proposes a high-throughput GPU-based LDPC decoder, aiming at the large-scale data process scenario, which optimizes the decoder from the perspectives of the decoding parallelism and data scheduling strategy, respectively. For decoding parallelism, the intra-codeword parallelism is fully exploited by combining the characteristics of the flooding-based decoding algorithm and GPU programming model, and the inter-codeword parallelism is improved using the single-instruction multiple-data (SIMD) instructions. For the data scheduling strategy, the utilization of off-chip memory is optimized to satisfy the demands of large-scale data processing. The experimental results demonstrate that the decoder achieves 10 Gbps throughput by incorporating the early termination mechanism on general-purpose GPU (GPGPU) devices and can also achieve a high-throughput and high-power-efficiency performance on low-power embedded GPU (EGPU) devices. Compared with the state-of-the-art work, the proposed decoder had a ×1.787 normalized throughput speedup at the same error correcting performance.
Non-binary low-density parity-check (LDPC) codes have been shown to attain near capacity error correcting performance in noisy wireless communication channels. It is well known that these codes require a very large number of operations per-bit to decode. This high computational complexity along with a parallel decoder structure makes graphics processing units (GPUs) an attractive platform for acceleration of the decoding algorithm. The seemingly random memory access patterns associated with decoding are generally beneficial to error-correcting performance but present a challenge to designers who want to leverage the computational capabilities of the GPU. In this paper we describe the design of an efficient decoder implementation based on GPUs and a corresponding set of powerful non-binary LDPC codes. Using the belief propagation algorithm with a sequential message updating scheme it is shown that we are able to exploit parallelism inherent in the decoding algorithm while decreasing the number of decoding iterations required for convergence.
A flexible and scalable approach for LDPC decoding on CUDA based Graphics Processing Unit (GPU) is presented in this paper. Layered decoding is a popular method for LDPC decoding and is known for its fast convergence. However, efficient implementation of the layered decoding algorithm on GPU is challenging due to the limited amount of data-parallelism available in this algorithm. To overcome this problem, a kernel execution configuration that can decode multiple codewords simultaneously on GPU is developed. This paper proposes a compact data packing scheme to reduce the number of global memory accesses and parity-check matrix representation to reduce constant memory latency. Global memory bandwidth efficiency is improved by coalescing simultaneous memory accesses of threads in a half-warp into a single memory transaction. Asynchronous data transfers are used to hide host memory latency by overlapping kernel execution with data transfers between CPU and GPU. The proposed implementation of LDPC decoder on GPU performs two orders of magnitude faster than the LDPC decoder on a CPU and four times faster than the previously reported LDPC decoder on GPU. This implementation achieves a throughput of 160Mbps, which is comparable to dedicated hardware solutions.
In this paper, efficient LDPC block-code decoders/simulators which run on graphics processing units (GPUs) are proposed. We also implement the decoder for the LDPC convolutional code (LDPCCC). The LDPCCC is derived from a predesigned quasi-cyclic LDPC block code with good error performance. Compared to the decoder based on the randomly constructed LDPCCC code, the complexity of the proposed LDPCCC decoder is reduced due to the periodicity of the derived LDPCCC and the properties of the quasicyclic structure. In our proposed decoder architecture, Γ (Γ is a multiple of a warp) codewords are decoded together, and hence, the messages of Γ codewords are also processed together. Since all the Γ codewords share the same Tanner graph, messages of the Γ distinct codewords corresponding to the same edge can be grouped into one package and stored linearly. By optimizing the data structures of the messages used in the decoding process, both the read and write processes can be performed in a highly parallel manner by the GPUs. In addition, a thread hierarchy minimizing the divergence of the threads is deployed, and it can maximize the efficiency of the parallel execution. With the use of a large number of cores in the GPU to perform the simple computations simultaneously, our GPU-based LDPC decoder can obtain hundreds of times speedup compared with a serial CPU-based simulator and over 40 times speedup compared with an eight-thread CPU-based simulator.
Compared to binary Low-Density Parity-Check (LDPC) codes, nonbinary LDPC (NB-LDPC) codes have better error correction performance under short-to-moderate block lengths or high-order modulations. Traditional min-sum-based soft decoding algorithms for NB-LDPC codes suffer from large computational complexity, which leads to inefficient hardware implementations. The Multiple-symbol-reliability weighted Bit-Reliability-Based (MwBRB) hard decoding algorithm achieves a good tradeoff between error correction performance and decoding complexity. However, efficient hardware implementations based on the MwBRB algorithm have not been investigated. In this brief, an improved layered MwBRB algorithm is first proposed, which results in faster convergence rate than the MwBRB algorithm. Then, an ultra-high-throughput low-complexity decoder architecture with an efficient partially parallel processing schedule is also presented. Finally, the proposed architecture is coded with RTL and synthesized under the TSMC 90-nm CMOS technology. The synthesis results demonstrate that the proposed decoder for a (837, 726) quasi-cyclic NB-LDPC code over GF(25) achieves a throughput of 21.66 Gbps and an area efficiency of 4.77 Gbps/M-gates under the TSMC 90-nm CMOS technology. The proposed decoder reaches a throughput more than 20 Gbps for the first time among the prior NB-LDPC decoders, and the area efficiency is far beyond the state-of-the-art designs.