首页> 外文会议>High Performance Computing on the Information Superhighway, 1997. HPC Asia '97 >Comparisons of fast 2D-DCT algorithms for parallel programmabledigital signal processors
【24h】

Comparisons of fast 2D-DCT algorithms for parallel programmabledigital signal processors

机译:用于并行可编程的快速2D-DCT算法的比较数字信号处理器

获取原文

摘要

Recent advances in technology make it possible to integratemultiple processors into a single chip to build high performanceparallel programmable digital signal processors (PPDSPs). Theseprocessors are expected to replace many dedicated digital signalprocessors to implement important image/signal processing algorithmssuch as discrete cosine transform (DCT). The paper addresses the issueof how to compare fast 2D-DCT algorithms when they are implemented on aPPDSP. Previously, the efficiency of these algorithms is compared basedon the number of operations. This comparison is reasonable when thesealgorithms are implemented on a dedicated DSP. However, this comparisonmay not be suitable for general-purpose PPDSPs. The paper proposes touse three parameters, the number of data accesses, the number ofcommunications, and the distance of communications, as new criterion forperformance comparison of DCT algorithms. An algorithm-level techniqueis developed to estimate these parameters for DCT algorithms. Thecomparison results based on the parameters show that the algorithmproposed by Cho and Lee (1991) might be the best choice for a PPDSPunless it requires large overhead for communication between remoteprocessors. In this case, the conventional row-column method with a fast1D-DCT algorithm might be the most efficient
机译:最近的技术进步使得可以整合 多个处理器进入单个芯片以构建高性能 并行可编程数字信号处理器(PPDSPS)。这些 预计处理器将替换许多专用数字信号 处理器实现重要图像/信号处理算法 如离散余弦变换(DCT)。本文解决了这个问题 如何比较快速的2D-DCT算法在其上实现时 PPDSP。以前,基于这些算法的效率 关于运营数量。这比较是合理的 算法在专用DSP上实现。但是,这种比较 可能不适合通用PPDSPS。本文提出了 使用三个参数,数据访问数量,数量 通信和通信距离,作为新标准 DCT算法的性能比较。一种算法级技术 开发用于估计DCT算法的这些参数。这 基于参数的比较结果显示该算法 Cho和Lee(1991)提议可能是PPDSP的最佳选择 除非它需要大的开销以进行遥控器之间的通信 处理器。在这种情况下,传统的行列方法快速 1D-DCT算法可能是最有效的

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