关键词: SM4算法, 流水线结构, S盒设计, SIMD, 轮密钥优化

Abstract: SM4 algorithm is a Chinese commercial cryptography standard and is widely used in the field of information security. Since SM4 implementation requires multiple iterations and loops, so SM4 algorithm is suitable for hardware implementation, this paper proposes a ROMbased parallel Sbox design, for SM4 ECB mode round key update pipeline structure, and the pipeline structure is further optimized based on SIMD（single instruction multidata） ideas, reducing the task overhead of round key generation, realizing simultaneous operation of singleround key generation with multiencryption and decryption data flow. Compared with the LUT pipeline structure and BRAM pipeline structure, the ROM pipeline structure proposed in this paper can reduce the use of hardware resources, and the throughput (input and output) can be increased by 1.33 times and 1.46 times, respectively; compared with the BRAM+ register pipeline structure, a lot of register resources are saved, but the throughput is not significantly reduced. When the 4 road encryption and decryption structure is generated for the 1 round key, the resource consumption is only 2.5 times of the ROM pipeline structure, but the efficiency is increased by 4 times.

Key words: SM4 algorithm, pipeline structure, SBox design, SIMD, round key optimization