Tanner图
线程(计算)
计算机科学
汉明码
低密度奇偶校验码
算法
消息传递
图形
解码方法
离散数学
组合数学
数学
并行计算
错误层
理论计算机科学
区块代码
操作系统
标识
DOI:10.1109/dspa57594.2023.10113442
摘要
In order to count the number of trapping sets (TS), we changed Cole's Importance Sampling (IS) approach, which led to a message-passing decoder problem. Several concepts have been combined to suggest improvements to Cole's IS: utilizing short cycles and simultaneous TS impulse tree decomposition, un-wrapping of message passing iterations Tanner Graph/Forney's Normal Graph symmetry's Graph Authomorphism is a complex yet simple concept. Superior Velasquez-Subramani and Karimi-Banihashemi TS enumerating techniques were supported. The proposed technique under PEG (1008, 504) Mackay code for single thread implementation is 43 times quicker than the original Cole's method and 5027 times (71463 times, multi-thread) faster than the Velasquez-Subramani LP method. For the TS enumerating problem at (2640, 1320), Margulis's code is 37958 times quicker than the suggested Velasquez-Subramani LP approach for single thread implementation, 82 times faster than Karimi-Banihashemi, and 134 times faster than Cole's original solution. On the basis of the creation of QC-LDPC codes, we demonstrate the effects of improved EMD spectrum and increased hamming distance on the TS spectrum and BER/FER error-floor level.
科研通智能强力驱动
Strongly Powered by AbleSci AI