Algorithms for multithreaded parallel processing of signals on heterogeneous digital telecommunication systems

摘要

The research carried out under this work is devoted to the problems of development and implementation of algorithms for multi-threaded parallel signal processing on heterogeneous computing systems, that is, in graphic processors. As a mathematical model for signal processing, the Haar wavelet function is used. In this paper are presented the results of evaluating the performance of a heterogeneous computing system, such as the average signal processing time, the usage of a graphics processor and the device’s global memory. The evaluation of the performance obtained with the implementation of the algorithms of wavelet transformation of signal on GPUs.