Abstract—This paper discusses the technologies of parallel processing of signals and images on dual-core and quad-core processors. The issues of accelerating the computation by the parallel vector-matrix organization procedures when performing spectral analysis algorithms in different basic systems Fourier and wavelet transform are consider. The techniques used for parallel computation on multi-core processors at the level of the algorithms are considered and highlighted features implementation of a parallel version of the FFT algorithm on sites that consume at the implementation of algorithms for the highest amount of CPU time. In the paper presents the tools that were used by the authors during the experiments are given. The use of these funds provided to simplify writing multi-threaded applications and provided an opportunity to objectively assess the resulting acceleration. For the implementation of the pilot studies, the authors created a software package which includes direct and inverse spectral transforms used in all basic systems. The results showed that multi-core processors in the processing of signals and images partition in too large parts do not allow the processors to load evenly and to achieve a minimum computation time and splitting into smaller parts increases the share of unproductive expenditure. 
 

Index Terms—signal, images, multicore processors, parallel programming, spectral transformation

Cite: Musaev Muhammadjon Mahmudovich and Berdanov Ulug’bek Abdumurodovich, "The Technology of Parallel Processing on Multicore Processors," International Journal of Signal Processing Systems, Vol. 4, No. 3, pp. 252-257, June 2016. doi: 10.18178/ijsps.4.3.252-257