目的是实现基于cuda的加密算法，批量加密提升速度
目前已经能够在GPU当中正常执行大整数的加、减、乘、除、取余、比较、高次幂取余、位移等操作
并且实现了进制转换和字符串初始化以及按位随机生成大整数的方法
目前的主要经历是在优化各个算法的执行速度
结合GMP实现了大质数的生成方法，并可以初始化RSA的加密公钥私钥对
random cost:0.933009
AoS add total_cost:3.2006s
SoA add total_cost:2.20597s
AoS subtract total_cost:3.1308s
SoA subtract total_cost:2.31791s
AoS multiply total_cost:10.6855s
SoA multiply total_cost:7.7833s
random cost:4.94042
AoS divide total_cost:11.4038s
SoA divide total_cost:7.67827s
AoS mod total_cost:11.7484s
SoA mod total_cost:9.35145s
random cost:4.05884
AoS power-mod total_cost:191.225s
SoA power-mod total_cost:155.377s
random cost:3.58363
AoS encrypt_total:1.56357s
AoS decrypt_total:158.047s
AoS total_cost:159.613s
SoA encrypt_total:1.12171s
SoA decrypt_total:143.849s
SoA total_cost:144.973s
add cost time:13.173848017636292s
subtract cost time:8.430511683475782s
multiply cost time:167.16211960807226s
div cost time:278.7772124865172s
mod cost time:266.27461345036033s
power_mod cost time:2973.2182057661344s
power_mod cost time:49.02678179423386s
power_mod cost time:2750.914678764818s