Parallel approximation of RSA encryption

RSA is the most popular public key cryptography algorithm. Therefore, one of the major weaknesses of RSA is that it requires a lot of computing time compared to secret hey cryptography algorithms. For example in hardware, some authors claim that RSA is around 1000 times slower than DES (Data Encryption Standard).
In addition, current consumer architectures embed several computing units, distributed on processors and possibly on graphics cards. These resources are now easily exploitable thanks to parallel programming interfaces like OpenMP or CUDA. This article proposes parallel versions of RSA allowing to take advantage of the whole of the computing resources, in particular on multi-core architectures with shared memory.