...
首页> 外文期刊>Journal of computational and theoretical nanoscience >Design of Coarse Grained Architecture with Gated Clock Technique for Low Power Applications
【24h】

Design of Coarse Grained Architecture with Gated Clock Technique for Low Power Applications

机译:低功耗应用粗粒架构粗粒架构设计

获取原文
获取原文并翻译 | 示例
           

摘要

In recent times, reconfiguration system is widely used due to their combination of efficiency and flexibility. Hardware based reconfiguration offers to use the same hardware for the different purpose of the optimized hardware utilization. Coarse grained (CGRA) and fine grained (FGRA)are the two types of reconfigurable architectures used in the present time. In the CGRA, the hardware functionality is specified at the word level. Likewise, the FGRA, hardware functionality is specified at a bit level. Coarse grained reconfigurable architecture (CGRA) needs many processingelements (PEs) and configuration memory for reconfiguration operation. CGRA reconfigure the processing element from one process to another process dynamically without any performance degradation. The CGRA system is designed with 16 processing elements; it’s arranged in a 4×4 matrixstructure. Each processing elements are configured by more than one operation. The operation can be selected and changed by using the selection bits. According to the selection bit, the operation of the processing element is changed; likewise, reconfiguration bit is used to reconfigure theprocessing elements. Although the reconfiguration architecture provides higher performance and flexibility, it consumes a significant amount of power. Reducing power in the CGRA and FGRA is the very critical issue; avoid this kind of situation by proposed the clock gating and pipelining reusetechniques. The proposed techniques avoid the unwanted usage of the clock signal during the operation of processing elements. Interconnections between the processing elements are essential in the CGRA structure. The proposed architecture was simulated by Modelsim 6.3c, synthesized and analyzedby Xilinx ISE simulator and Quartus with cyclone II device.
机译:最近,由于它们的效率和灵活性的组合,重新配置系统被广泛使用。基于硬件的重新配置优惠要使用相同的硬件,以实现优化的硬件利用率的不同目的。粗粒(CGRA)和细粒细粒(FGRA)是本时间使用的两种可重构架构。在CGRA中,硬件功能在字级别指定。同样,FGRA,硬件功能在比特级别指定。粗粒粒度可重配置架构(CGRA)需要许多处理单(PES)和配置内存以进行重新配置操作。 CGRA在动态地将处理元素从一个过程重新配置到另一个过程,没有任何性能下降。 CGRA系统设计有16个加工元件;它布置在4×4矩阵结构中。每个处理元件由多个操作配置。可以通过使用选择位来选择和改变操作。根据选择位,改变了处理元件的操作;同样,重新配置位用于重新配置Hocessing元素。虽然重新配置架构提供更高的性能和灵活性,但它会消耗大量功率。降低CGRA和FGRA的力量是非常关键的问题;通过提出时钟门控和流水线的REUSETECHNIQUES来避免这种情况。所提出的技术避免了在处理元件的操作期间避免了时钟信号的不需要的使用。处理元件之间的互连在CGRA结构中是必不可少的。所提出的架构是由ModelSim 6.3C,合成和分析的Xilinx ISE Simulator和Quartus模拟的模拟架构。

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