The Implementation Of Parallel Processing Technology Based On MPI In Three Digital Signal Transformation Algorithms

In digital signals processing field, high speed and extremely high speed signals processing technology becomes an important method to solve many applications of scientific research and engineering fields. Because the digital signals contain abundant information and large volume of data. How to minimum the time overhead and space cost is the key point in digital signal transformation.The contradiction between the finiteness of single computer's technology development and the infiniteness of requirement for science computing determines that multi-computers' parallel processing will be the development of computer. and the development of parallel processing technology will be the main power to impel the constructing of national defense and the development of national science and technology. Accordingly, the increasingly demanding for science computing is also the main power to impel the flourish development of parallel processing technology in those twenty years. So the research of parallel processing technology in digital signals processing is very count.The paper will take parallel algorithm and its implementation in digital signals processing field as research object, the applications of parallel algorithm in FFT ,DCT and wavelet transform are especially discussed.Firstly,this paper introduces the theory of FFT and DCT, a parallel algorithm is mainly designed for the solution computation of FFT and DCT in cluster apllication . Design ideal and design strategy are described in detail, secondly, according to the theory of wavelet transform; a DWT parallel algorithm is designed ,detailed algorithm description and analysis is also given.The parallel algorithms in the paper have been implemented on MPI on Linux. It is proved that the expansibility and efficiency are both well.