张晓磊,戴紫彬,刘燕江,等.面向分组密码的可重构联合移位运算单元设计与实现[J]. 微电子学与计算机,2023,40(7):118-128. doi: 10.19304/J.ISSN1000-7180.2022.0615
引用本文: 张晓磊,戴紫彬,刘燕江,等.面向分组密码的可重构联合移位运算单元设计与实现[J]. 微电子学与计算机,2023,40(7):118-128. doi: 10.19304/J.ISSN1000-7180.2022.0615
ZHANG X L,DAI Z B,LIU Y J,et al. Design and implementation of reconfigurable joint shift operation unit for block ciphers[J]. Microelectronics & Computer,2023,40(7):118-128. doi: 10.19304/J.ISSN1000-7180.2022.0615
Citation: ZHANG X L,DAI Z B,LIU Y J,et al. Design and implementation of reconfigurable joint shift operation unit for block ciphers[J]. Microelectronics & Computer,2023,40(7):118-128. doi: 10.19304/J.ISSN1000-7180.2022.0615

面向分组密码的可重构联合移位运算单元设计与实现

Design and implementation of reconfigurable joint shift operation unit for block ciphers

  • 摘要: 移位运算是分组密码中的重要构件,针对可重构密码处理器中移位运算单元灵活性低的问题,通过分析分组密码移位操作特点,提出了面向分组密码的联合移位运算. 通过交换操作数、配置移位位数的方式,实现不同位宽不同方向的移位运算. 在此基础上,分别采用半定制和全定制IC设计方法分别设计了基于标准单元库、基于CMOS传输门和基于NMOS单管传输门的可重构联合移位电路. 实验表明,与基于标准数字单元库的联合移位电路(S-RJSM)相比,基于CMOS传输门实现的联合移位电路(C-RJSM)和基于NMOS单管传输门实现的联合移位电路(N-RJSM)延时分别减少了85.30%、50.69%,功耗-延时积分别减少了97.27%、90.43%. 与其他CGRA移位运算单元相比处理速度更快,可有效提升CGRA计算性能.

     

    Abstract: Shift operation is an important component of block cipher. Aiming at the problem of low flexibility of shift operation unit in reconfigurable cipher processor, by analyzing the characteristics of shift operation of block cipher, the joint shift operation for block cipher was proposed. By exchanging operands and configuring shift bits, shift operations with different bit widths and different directions were realized. On this basis, the semi-custom and full-custom IC design methods are used to design the reconfigurable joint shift circuits based on standard cell library, CMOS transfer gate and NMOS single transistor transfer gate, respectively. The experimental results show that compared with the joint shift circuit based on standard digital cell library (S-RJSM), the delay of joint shift circuit based on CMOS transfer gate (C-RJSM) and NMOS single transistor transfer gate (N-RJSM) is reduced by 85.30% and 50.69%, respectively. The power-delay product is reduced by 97.27% and 90.43% respectively. Compared with other shift operation units in CGRA, it also has obvious advantages. Compared with other CGRA shift computing units, it has faster processing speed and can effectively improve the CGRA computing performance.

     

/

返回文章
返回