GENERATION OF GRAPHIC OBJECTS IN THE GODOT ENGINE USING PARALLEL COMPUTATION METHOD

This article presents a study of graphic object generation methods in the Godot game engine using parallel computing. The main focus is on the efficient use of computational resources to improve the performance of graphic content generation in dynamic game environments. The research conducted a detailed analysis of four main categories of graphic object generation algorithms actively used in the Godot Engine ecosystem: noise-based algorithms (Perlin, OpenSimplex), which are used to create natural landscapes and textures; cellular automata, which allow modeling organic structures; Wave Function Collapse (WFC), which is used to generate structured levels with connection rules; and L-systems, which are effectively used to generate fractal structures, particularly vegetation. For each type of algorithm, the potential for parallelization and factors limiting their efficiency on multi-core systems were determined. The problem identified during the analysis is the presence of a significant proportion of sequential computations in generation algorithms, which, according to Amdahl's Law, substantially limits the maximum possible speedup even when increasing the number of computational cores. A mathematical model is proposed for estimating potential speedup in the context of graphic object generation, allowing more accurate prediction of the performance of different algorithms on multi-core systems. As part of the research, an innovative optimization approach was developed, based on the chunk generation method, where graphic objects are divided into spatially localized elements with minimal inter-element dependencies. The key feature of the proposed approach is asynchronous chunk generation and localization of coordinate calculations, which significantly reduces the proportion of sequential operations and, consequently, greatly increases the potential speedup according to Amdahl's Law. For an objective evaluation of the developed method's effectiveness, a series of experiments was conducted, including benchmarking on different types of graphic objects using the built-in Godot Engine tools. The measurement methodology was based on decomposing the algorithm into logical phases and analyzing changes in execution time with a gradual increase in the number of threads. The experimental results demonstrated a significant reduction in the proportion of sequential computations for all types of graphic objects when applying the chunk approach, confirming the effectiveness of the proposed method. The research results have practical significance for optimizing game development based on the Godot Engine and can be adapted for other visualization systems with similar architecture.