Comparative Analysis Of Integer Factorization Algorithms Using Cpu And Gpu

Abstract

In this work we have evaluated the running time of four integer factorization algorithms, namely, trial division algorithm, Fermat algorithm, Pollard rho and Brent algorithms. Implementation of these algorithms was performed in three ways on c programming language, on c programming language, using GMP 6.0.0 library and on CUDA architecture to run on GPU. Results showed that Fermat algorithm and trivial division algorithm had the fastest running time in parallel implementation on CUDA architecture. The difference of running times between CUDA implementation and GMP implementation was up to 10 times. The difference between c and c implementation was mainly due to difference in these programming languages.

Keywords

• integer factorization
• gpu
• gmp
• trial division algorithm
• fermat algorithm
• pollard rho algorithm
• brent algorithm.
• целочисленная факторизация
• цпу
• gmp
• тривиальный алгоритм факторизации
• алгоритм ферма
• алгоритм полларда ро
• алгоритм брента