high performance computing on graphics processing units: hgpu.org

With the development of the GPGPU (General-purpose computing on graphics processing units), more and more computing problems are solved by using the parallel property of GPU (Graphics Processing Unit). CUDA (Compute Unified Device Architecture) is a framework which makes the GPGPU more accessible and easier to learn for the general population of programmers. This is because it builds on C and hides many of the complicated details of how the GPU works from a CUDA developer. Using the unique properties of the GPU through CUDA has greatly increased the efficiency of many computational problems. The target in this project is to study and analyses the majority of algorithms related to the cryptography and then to design and make an implementation of algorithm in CUDA. Finally, this reach will compare the performance between the GPU implementation and the CPU implementation in order to look into the possibility of improving the performance of algorithms. The survey done in three cryptography algorithm AES, SHA, MD5.