应用科学学报 ›› 2023, Vol. 41 ›› Issue (6): 1031-1045.doi: 10.3969/j.issn.0255-8297.2023.06.010

• 计算机科学与应用 • 上一篇    下一篇

哈希算法异构可重构高能效计算系统研究

郑博文1, 聂一1, 柴志雷2,3   

  1. 1. 江南大学 物联网工程学院, 江苏 无锡 214122;
    2. 江南大学 人工智能与计算机学院, 江苏 无锡 214122;
    3. 江苏省模式识别与计算智能工程实验室, 江苏 无锡 214122
  • 收稿日期:2021-12-13 出版日期:2023-11-30 发布日期:2023-11-30
  • 通信作者: 柴志雷,教授,研究方向为嵌入式系统、软硬件协同设计等。E-mail:zlchai@jiangnan.edu.cn E-mail:zlchai@jiangnan.edu.cn
  • 基金资助:
    国家自然科学基金(No.61972180)资助

Research on Hash Algorithm Heterogeneous Reconfigurable High Energy Efficiency Computing System

ZHENG Bowen1, NIE Yi1, CHAI Zhilei2,3   

  1. 1. School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, Jiangsu, China;
    2. School of Artificial Intelligence and Computer Science, Jiangnan University, Wuxi 214122, Jiangsu, China;
    3. Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence, Wuxi 214122, Jiangsu, China
  • Received:2021-12-13 Online:2023-11-30 Published:2023-11-30

摘要: 针对应用场景中不同哈希算法乃至多哈希算法组合的高速计算需求,纯软件方式难以满足性能需求,基于FPGA或ASIC的硬件方式又面临灵活性不足的问题,设计了一种异构且加速端硬件可重构的哈希算法高能效计算系统。计算系统由算法硬件加速模块、数据传输模块、多线程管理模块实现,并且通过硬件的动态可重构设计提升了计算能效。实验结果表明,在Intel Stratix10 FPGA异构计算平台上,针对加解密计算,选择MD5、SHA-1、SHA-256、SHA-512和RIPEMD-160算法作为测试对象,所实现的系统相比Intel Core I7-10700CPU,最高可获得18.7倍的性能提升和34倍的能效提升,相比NVIDIA GTX 1650 SUPER GPU,最高可获得2倍的性能提升和5.6倍的能效提升。

关键词: 异构计算, 哈希算法, SHA-256, 硬件加速, 现场可编程逻辑门阵列

Abstract: To meet the high-speed computing requirements of different hash algorithms and the combination of different hash algorithms in various application scenarios, a highefficiency computing system for Hash algorithm with heterogeneous and reconfigurable acceleration end hardware is presented in this paper. The computing system consists of an algorithm hardware acceleration module, a data transmission module, and a multithread management module. The computing energy efficiency is improved through the dynamically reconfigurable hardware design. Experimental results on the Intel Stratix10FPGA heterogeneous computing platform demonstrate significant performance and energy efficiency improvements. Compared with the Intel Core I7-10700 CPU, the system achieves up to 18.7 times performance improvement and 34 times energy efficiency improvement.Compared with the NVIDIA GTX 1650 SUPER GPU, the system achieves up to 2 times performance improvement and 5.6 times energy efficiency improvement.

Key words: heterogeneous computing, hash algorithm, SHA-256, hardware acceleration, field programmable gate array(FPGA)

中图分类号: