论文部分内容阅读
钱搜索是与非型闪速(NAND flash)存储器中BCH译码器的重要组成部分,并行钱搜索延迟较小并可高速运行,但过高的复杂度制约了其的应用。为了降低并行钱搜索的复杂度,提出一种并行钱搜索的改进和优化方法。首先对传统并行钱搜索以及有关文献进行了分析和研究;然后对公共子表达式的搜索范围进行了扩展,并合并了相关计算;最后对迭代匹配算法进行了改进,提出一种基于二维搜索的改进迭代匹配算法。实验结果表明,本文方法取得了较好的优化效果,有效地降低了并行钱搜索的复杂度;在对BCH(2 047,1 926,11)的32bit并行钱搜索优化后,与传统并行钱搜索以及有关文献的两种并行钱搜索相比,本文方法的优化率分别达到了85.4%、38.7%和29.2%,并可以更好地适应不同并行度和不同纠错能力的并行钱搜索结构。
Money search is an important part of the BCH decoder in NAND flash memory. The parallel money search delay is small and can be run at high speed, but the high complexity restricts its application. In order to reduce the complexity of parallel money search, a method of improving and optimizing parallel money search is proposed. Firstly, the traditional parallel money search and related literature are analyzed and researched. Then, the search range of public sub-expressions is expanded and the related calculation is merged. Finally, the iterative matching algorithm is improved, and a new algorithm based on two-dimensional search Improved iterative matching algorithm. The experimental results show that the proposed method achieves better optimization results and reduces the complexity of parallel money search effectively. After optimizing 32bit parallel money search of BCH (2 047, 1 926, 11), the traditional parallel money search Compared with the two kinds of parallel money search in literature, the proposed method achieves 85.4%, 38.7% and 29.2% respectively, and can better adapt to the parallel money search structure with different degree of parallelism and different error correction ability.