Forward propagation algorithm in the dataflow paradigm

The article discusses the issue of developing and implementing the forward propagation algorithm in the dataflow paradigm. The principles of dataflow computing differ significantly from traditional control-flow computing, just as the dataflow programming paradigm differs from the imperative one. Programs created in the dataflow paradigm are initially parallel. The article provides a detailed description of the forward propagation dataflow algorithm and the program that runs on the parallel dataflow computing system "Buran". The dataflow program is compact and versatile in its code. It automatically scales to the entire system, and its program code does not contain references to library functions and is based solely on basic arithmetic operations such as addition, multiplication, and comparison. The program is capable of processing perceptrons of any dimension without modification and recompilation of the program code. The size and structure of the processed perceptron is determined by the initial data. The experimental part of the article demonstrates the high adaptability of the dataflow program to processing large amounts of continuously incoming data. An assessment of the applicability of the dataflow computing model for solving neural network problems is given.