向下全互连可重构阵列设计与实现

吴皓月; 邓军勇; 山蕊; 贺飞龙; 刘新闯

Chinese

您当前的位置：

首页 >

文章列表页 >

向下全互连可重构阵列设计与实现

更新时间：2026-09-04

- 向下全互连可重构阵列设计与实现
- Design and implementation of a fully interconnected reconfigurable array
- 微电子学与计算机 2019年36卷第10期页码：59-64
- 作者机构：
  
  1. 西安邮电大学电子工程学院,陕西,西安,710121
  2. 西安邮电大学计算机学院,陕西,西安,710121
- 作者简介：
- 基金信息：
  
  国家自然科学基金资助项目 61802304国家自然科学基金资助项目 61772417国家自然科学基金资助项目 61834005国家自然科学基金资助项目 61602377国家自然科学基金资助项目 61634004陕西省国际科技合作计划项目 2018KW-006陕西省科技统筹创新工程项目 2016KTZDGY02-04-02陕西省重点研发计划 2017GY-060
- DOI：
  中图分类号：
- 纸质出版：2019
- 稿件说明：
移动端阅览
吴皓月, 邓军勇, 山蕊, 贺飞龙, 刘新闯. 向下全互连可重构阵列设计与实现[J]. 微电子学与计算机, 2019, 36(10): 59-64.

WU Hao-yue, DENG Jun-yong, Shan Rui, et al. Design and implementation of a fully interconnected reconfigurable array[J]. 2019, 36(10): 59-64.
吴皓月, 邓军勇, 山蕊, 贺飞龙, 刘新闯. 向下全互连可重构阵列设计与实现[J]. 微电子学与计算机, 2019, 36(10): 59-64. DOI：

WU Hao-yue, DENG Jun-yong, Shan Rui, et al. Design and implementation of a fully interconnected reconfigurable array[J]. 2019, 36(10): 59-64. DOI：

摘要

可重构电路的灵活性是电路中的绝对优势

但其电路互连资源有限、PE (Processing Element)资源利用率低.向下全互连可重构阵列结构实现了阵列中每个PE与下一行PE的全连接

提高了PE资源利用率、整体数据处理速度.本设计使用modelsim、quartus分别进行了功能仿真、综合下载

结果表明所设计的可重构数据流处理单元能够进行多种配置的可重构.通过验证FFT算法在向下全互连的可重构阵列上的映射

结果显示可重构数据流处理单元成功实现了预期功能

并且满足其时序等要求.综合结果显示

4*4的可重构阵列

包括其外围电路

共占用19429(58%)门逻辑资源

支持76.07 MHZ的工作频率; 单个的可重构数据流处理器占用2565(8%)门逻辑资源

支持最高108.96 MHZ的工作频率.

Abstract

The flexibility of the reconfigurable circuit is an absolute advantage in the circuit

but its circuit interconnection resources are limited

and the utilization of PE (Processing Element) resources is low. The downward fully interconnected reconfigurable array structure realizes full connection between each PE and the next row of PEs in the array

improving PE resource utilization and overall data processing speed. This design uses modelsim and quartus to perform functional simulation and comprehensive downloading. The results show that the reconfigurable data stream processing unit can be reconfigurable in various configurations. By verifying the mapping of the FFT algorithm on the fully interconnected reconfigurable array

the results show that the reconfigurable data stream processing unit successfully implements the expected functionality and meets its timing requirements. The combined results show that 4*4 reconfigurable arrays

including its peripheral circuits

occupy a total of 19

429 (58%) gate logic resources

supporting a working frequency of 76.07 MHz; a single reconfigurable data stream processor occupies 2565 (8%) The gate logic resource supports a working frequency of up to 108.96 MHz.

关键词

Keywords

references

KIM J, BALFOUR J, DALLY W. Flattened butterfly topology for on-chip networks[C]//Proceedings of the 40th Annual IEEE/ACM International Symposium on Microarchitecture. IEEE Computer Society, 2007: 172-182.

XU T C, LEPPANEN V, LOLJEBERG P, et al. PDNOC: Partially diagonal network‐on‐chip for high efficiency multicore systems[J]. Concurrency and Computation: Practice and Experience, 2015, 27(4): 1054-1067. DOI: 1054

STITT G, VAHID F. Thread warping: a framework for dynamic synthesis of thread accelerators[C]//Proceedings of the 5th IEEE/ACM international conference on Hardware/software codesign and system synthesis. ACM, 2007: 93-98.

CAI X, HU W, MA T, et al. A hybrid scheduling algorithm for reconfigurable processor architecture[C]//2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA). IEEE, 2018: 745-749.

SABBAGHI-NADOOSHAN R, PATOOGHY A. Analytical performance modeling of de Bruijn inspired mesh-based network-on-chips[J]. Microprocessors and Microsystems, 2015, 39(1): 27-36. DOI: 27

WANG G, STITT G, LAM H, et al. A framework for core-level modeling and design of reconfigurable computing algorithms[C]//Proceedings of the Third International Workshop on High-Performance Reconfigurable Computing Technology and Applications. ACM, 2009: 29-38.

DAIL M, GIBSON R M, AMIRA A, et al. An efficient MIMO-OFDM radix-2 single-path delay feedback FFT implementation on FPGA[C]//2015 NASA/ESA Conference on Adaptive Hardware and Systems (AHS). IEEE, 2015: 1-7.

MANUEL B R, KONGUVEL E, KANNAN M. An area efficient high speed optimized FFT algorithm[C]//2017 Fourth International Conference on Signal Processing, Communication and Networking (ICSCN). IEEE, 2017: 1-5.

YIN S, OUYANG P, YANG J, et al. An ultra-high energy-efficient reconfigurable processor for deep neural networks with binary/ternary weights in 28nm CMOS[C]//2018 IEEE Symposium on VLSI Circuits. IEEE, 2018: 37-38.

MOOKHERJEE S, DEBRUNNER L, DEBRUNNER V. A low power radix-2 FFT accelerator for FPGA[C]//2015 49th Asilomar Conference on Signals, Systems and Computers. IEEE, 2015: 447-451.

DALES M. Managing a reconfigurable processor in a general purpose workstation environment[C]//2003 Design, Automation and Test in Europe Conference and Exhibition. IEEE, 2003: 980-985.

ROUHANI B D, MIRHOSEINI A, KOUSHANFAR F. Rise: An automated framework for real-time intelligent video surveillance on fpga[J]. ACM Transactions on Embedded Computing Systems (TECS), 2017, 16(5s): 158. DOI: 158

HSIUNG P A, LIN C S, LIAO C F. Perfecto: A systemc-based design-space exploration framework for dynamically reconfigurable architectures[J]. ACM Transactions on Reconfigurable Technology and Systems (TRETS), 2008, 1(3): 17. DOI: 17

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

一种基于共享缓存的星载数据交换架构设计

5G定位算法的FPGA加速设计研究

YOLO目标检测后处理算法的优化和硬件加速

基于FPGA的H.264可变块运动估计算法的优化及实现

基于TRNG与PUF集成结构的安全加密认证系统