OPTIMIZATION OF DESCRIPTOR ENTITY COMPARISON ALGORITHM USING WORKER POOL PARALLEL COMPUTING APPROACH

The article considers the analysis and optimization of the algorithm for comparing descriptor entities. The temporal characteristics of the algorithm's operation are studied on different input data when using the worker pool approach with different numbers of workers and without it. Variants of the algorithm's operation in a single thread are investigated, as well as optimal transition points for using a larger number of threads. The problem of overhead costs when using more than one thread is solved. The developed solution was tested on several local machines and showed an increase in performance when working with various descriptor entities, which allows users to save time when working with the algorithm.