In this paper parallel-vertical approach to realization of group summation has been analyzed. Analytical expression for synthesis of a 7-input single-digit adder have been realized. The structure and the program model for it implementation have been developed.
Requirements and design phases of hardware and software intellectual processing of intensive data streams in real time have been examined. Design principles of hardware and software tools have been proposed. Translation of processing algorithms into hardware and software tools has been described. The basic structure of parallel-flow system with data exchange through multiport memory has been proposed.
11 April 2022 in Circuits, Systems, and Signal Processing
Methods, algorithms and structures of neural networks were analyzed. Basic components of neural networks were defined and the principles of their development were chosen. It was shown that use of vertical-parallel method for implementation of work algorithms of neural networks basic components provides increased performance, reduce hardware costs and efficient VLSI implementation. Parallel-consequent code converter, which provides time alignment of the processes of data reception and formation of bit sections was developed. The methods of vertical parallel computing were developed: sum of difference squares, due to parallel processing of bit sections, provide performance increase. It was shown that use of developed basic components for neural networks synthesis will provide reduction of time and development cost.
In the paper scalar product calculation algorithms have been described, the parallel-How graph of the algorithm for calculating the scalar product using tables for the formation of macro-partial products has been proposed, the structure of the matrix scalar product calculating device has been created.