Performance factors of a CUDA GPU parallel program: A case study on a PDF password cracking brute-force algorithm

Brute-force algorithm needs large amount of computational resources. CUDA is one of computing platforms which are suitable to support this algorithm. In this paper, we discussed about 5 factors of which may be affecting a GPU based parallel program performance indicator. We had constructed custom and testbed algorithms to evaluate those factors. Testbed algorithms were constructed based on a previous thesis work regarding PDF password cracking. The final algorithm was constructed from significantly affecting factors. All parallel algorithms were implemented on Tesla C2075. Speedup result of final algorithm implementations are 2.92 for 2 bytes alphanumeric passwords and 4.77 for 6 bytes numeric passwords.

[1]  Jie Cheng,et al.  Programming Massively Parallel Processors. A Hands-on Approach , 2010, Scalable Comput. Pract. Exp..

[2]  Donghui Guo,et al.  Efficient implementation for MD5-RC4 encryption using GPU with CUDA , 2009, 2009 3rd International Conference on Anti-counterfeiting, Security, and Identification in Communication.

[3]  Jason Sanders,et al.  CUDA by example: an introduction to general purpose GPU programming , 2010 .

[4]  Sam S. Stone,et al.  Program Optimization Study on a 128-Core GPU , 2011 .

[5]  Joseph White,et al.  Breaking Weak 1024-bit RSA Keys with CUDA , 2012, 2012 13th International Conference on Parallel and Distributed Computing, Applications and Technologies.

[6]  Elliot B. Koffman,et al.  Problem Solving and Program Design in C , 1979 .

[7]  Marco Cavazzuti,et al.  Design of Experiments , 2013 .