多核CPU和GPU系统上椭圆曲线点乘并行计算研究-计算机系统结构专业论文.docxVIP

II II 验结果表明：利用GPU可以显著加速椭圆曲线点乘的计算，与多核CPU并 行计算椭圆曲线点乘算法相比，多核CPU和GPU协同计算椭圆曲线点乘并 行算法所需运行的时间明显缩短。 关键字：椭圆曲线密码 点乘计算 并行算法 多核CPU计算 GPU计算 PAGE PAGE IV PARALLEL COMPUTING FOR SCALAR MULTIPLICATION IN ELLIPTIC CURVE CRYPTOGRAPHY ON MULTI-CORE CPU AND GPU SYETEM ABSTRACT Elliptic Curve Cryptography is recognized as the most promising third generation of public key cryptography. Elliptic curve point multiplication computing elliptic curve cryptography system is to achieve critical operations. In this paper, the problem of multiplication to calculate time-consuming to analyze for elliptic curve point, combined with the parallel computing ideas, design and implementation of the multi-core CPU and GPU systems parallel computing elliptic curve point multiplication algorithm research. When the number of computing cores running on multi-core computers, and so does the number of parallel threads simultaneously, elliptic curve point multiplication to calculate the time required will vary accordingly. This paper presents task partitioning strategy, elliptic curve point multiplication to calculate a reasonable division of tasks assigned to each computing core processing on multi-core computer systems, allowing each core to perform multi-threaded parallel computing dot design an elliptic curve point efficient multi-core computer systems multiplication algorithm for calculating thread-level parallelism. Experimental results show that: with the elliptic curve point multiplication to calculate the serial algorithm, the elliptic curve point multiplication to calculate the proposed multi-core CPU parallel algorithm significantly improves the speed, the acceleration effect is obvious. Consider the powerful GPU computing capabilities for high performance elliptic curve point multiplication calculation-intensive tasks such calculations practical application, this paper studies the CPU and GPU co-processing task scheduling method elliptic curve point multiplication calculations, a