论文部分内容阅读
椭圆曲线标量乘法的计算量大,对处理器的运算能力要求高,是整个椭圆曲线密码体制实现的关键;标量乘运算依赖于下层的点运算和最底层的有限域乘法、加法、求逆的运算速度,因此设计高效的标量乘和有限域运算硬件结构是提升椭圆曲线标量乘性能的关键.采用指令方式实现Montgomery模逆运算,相比于传统实现减少了近1/3的延迟;提出一种混合点运算的数据流水排布模式,相比于顺序执行点加、倍点运算减少了近一半的延迟,具有较高的并行度和运算部件利用率;设计并实现了一种基于定制指令的椭圆曲线标量乘处理器,它在Virtex6VSX475T FPGA上完成一次标量乘运算需要0.275ms,吞吐量为18180次/秒,相比于相关工作[3-5,8]可以提升约45.9%~97.6%的运算延迟.
Elliptic curve scalar multiplication of the large amount of computing power of the processor demanding, is the key to achieve the entire elliptic curve cryptosystem; scalar multiplication operation depends on the underlying point operations and the bottom of the finite field multiplication, addition, inversion Therefore, the design of efficient hardware architecture of scalar multiplication and finite field operation is the key to improve the performance of scalar multiplication of elliptic curve.Compared with the traditional implementation, the Montgomery modulo inverse operation is realized by using instruction mode, Compared with the sequential execution of point addition and double point operation, the data point allocation reduces the delay by nearly half and has high parallelism and operation component utilization rate. A design and implementation based on custom instructions Of elliptic curve scalar multipliers, it takes 0.275ms to complete a scalar multiplication operation on a Virtex6VSX475T FPGA with a throughput of 18180 bps, which can be improved by about 45.9% ~ 97.6% compared with the related work [3-5,8] Operation delay.