On green and energy-aware GPU computing for scientific applications

摘要: Recently, modern graphics processing unit (GPU) has gained the reputation of computational accelerator that can achieve a significant increase in performance by reducing execution time for different type scientific application demand high computing. While GPUs reduce parallel as compared to CPU implementation, but this is sometimes achieved at an expense considerable power and energy consumption. This paper seeks characterize explore impression consumption GPU. We examine notion reviewing techniques used researchers analyze performance, power, characteristics are utilized These studies consider applications run on traditional setup, transformed applications, running hybrid CPU+GPU environment. indicated heterogeneous environment delivers energy-aware sustainable product much better than application.