首页> 外文期刊>Circuits and Systems II: Express Briefs, IEEE Transactions on >Design of a High-Throughput QC-LDPC Decoder With TDMP Scheduling
【24h】

Design of a High-Throughput QC-LDPC Decoder With TDMP Scheduling

机译:具有TDMP调度的高通量QC-LDPC解码器的设计

获取原文
获取原文并翻译 | 示例

摘要

Low-density parity-check (LDPC) codes with turbo-decoding message-passing (TDMP) scheduling can obtain good performance and high convergence rates. In addition, the min–sum (MS) algorithm can reduce the complexity. The hybrid normalized MS algorithm with TDMP scheduling is presented to achieve good performance and to lower the complexity. For a quasi-cyclic LDPC (QC-LDPC) code with a long code length, parallel degree optimization and an offset iterative sequence rule are proposed. With the proposed techniques, the data correlation problem and memory access conflicts during TDMP scheduling can be resolved so that the iteration can smoothly proceed through the reasonable division of each block row. Fabricated in the 90-nm 1-Poly 9-Metal (1P9M) CMOS process, a multimode 96 000-bit irregular QC-LDPC decoder is implemented. It attains throughputs of 1.7–3.0 Gb/s and dissipates an average power of 502 mW at an operation frequency of 100 MHz and at 10 iterations. The decoder chip area is 13.32 , with a core area of 9.73 .
机译:具有Turbo解码消息传递(TDMP)调度功能的低密度奇偶校验(LDPC)码可以获得良好的性能和较高的收敛速度。另外,最小和(MS)算法可以降低复杂度。提出了具有TDMP调度的混合归一化MS算法,以实现良好的性能并降低复杂度。对于具有长码长的准循环LDPC(QC-LDPC)码,提出了并行度优化和偏移迭代序列规则。利用所提出的技术,可以解决TDMP调度期间的数据相关性问题和存储器访问冲突,从而可以通过合理地划分每个块行来使迭代顺利进行。采用90-nm 1-Poly 9-Metal(1P9M)CMOS工艺制造,实现了多模96 000位不规则QC-LDPC解码器。在100 MHz的工作频率和10次迭代时,它的吞吐量达到1.7-3.0 Gb / s,平均功耗为502 mW。解码器芯片区域为13.32,核心区域为9.73。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号