Impact of Data-Oriented and Object-Oriented Design on Performance and Cache Utilization with Artificial Intelligence Algorithms in Multi-Threaded CPUs

Researchers published a comprehensive performance analysis on arXiv (paper 2512.07841v2) comparing Data Oriented Design (DOD) with traditional Object-Oriented Design (OOD) in multi-threaded environments, specifically examining cache utilization efficiency as the performance gap between multi-core CPUs and main memory continues to widen. The study developed and compared four distinct versions of the A* search algorithm, including single-threaded OOD (ST-OOD), with the goal of understanding how different design paradigms affect performance in modern computing architectures. This research addresses a critical challenge in contemporary software development where traditional programming approaches often fail to fully leverage the capabilities of multi-core processors. The growing divergence between CPU processing capabilities and memory access speeds has created an urgent need for software design methodologies that are more cognizant of hardware constraints. By focusing on cache utilization, the researchers tackle one of the most significant performance bottlenecks in modern computing systems, where inefficient data access patterns can negate the benefits of multi-core processing. The findings from this study provide valuable insights for developers working with AI algorithms and other computationally intensive applications, offering concrete guidance on how to structure code for optimal performance on contemporary hardware architectures.