ANALYSIS AND RESEARCH OF MODERN COMPUTER ARCHITECTURES

Abstract

The article presents a comprehensive analysis of modern computer architectures and their development trends in the context of contemporary computational challenges. The research explores the evolution of processor design principles from the classical von Neumann model to modern heterogeneous and AI-oriented systems. Particular attention is paid to the fundamental architectural paradigms, including RISC (Reduced Instruction Set Computing), CISC (Complex Instruction Set Computing), and various models of parallel computation such as SIMD (Single Instruction, Multiple Data), MIMD (Multiple Instruction, Multiple Data), and GPU-based architectures. The study also examines the emergence of specialized accelerators designed for artificial intelligence and deep learning tasks, such as Google TPU and Apple Neural Engine, which significantly improve performance and energy efficiency.

The analysis highlights the key problem of balancing high computational performance with energy efficiency and scalability, which has become increasingly relevant due to the physical limitations of semiconductor technologies. Modern computing systems are moving toward heterogeneous integration, combining CPUs, GPUs, and AI accelerators within unified architectures to achieve optimal task distribution. The paper reviews current scientific approaches to optimizing computational throughput, reducing energy consumption, and improving hardware-software interaction in complex systems.

The research concludes that the evolution of computer architecture is characterized by the transition from universal to specialized and adaptive computing systems. Future progress is expected to be driven by the development of open architectures such as RISC-V, near-memory computing technologies, and neuromorphic processors that imitate biological neural structures. The presented findings contribute to understanding current architectural trends and form the basis for further research aimed at designing efficient, scalable, and intelligent computing platforms.