Comparative Study of GPU Performance and Energy Efficiency Across Generational Architectures: A Systematic Literature

In this study, the evolution of GPU architectures across multiple generations is explored, focusing on their computational performance and energy efficiency. Using a systematic literature review and comparative analysis, this study examines key advancements in GPU technology, particularly in AI workloads, ray tracing, gaming, and scientific simulations. The analysis reveals that modern GPUs, such as those based on the Ampere architecture, demonstrate significant improvements in both performance and energy efficiency, with a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{3 0 \%}$</tex> increase in AI task performance and a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{2 0 \%}$</tex> reduction in power consumption compared to previous generations. This study also highlights how innovations in specialized cores and optimized memory management have enabled GPUs to excel in domain-specific applications. Additionally, the broader implications of these advancements for other computational fields, including gaming and scientific research, are addressed. By providing a comprehensive view of these developments, this review offers valuable insights into future GPU design trends and the ongoing challenges in balancing high performance with energy efficiency.