With the increase in data-intensive applications, enterprise information technology (IT) companies are focusing on the performance of processor hardware to determine the optimal central processing unit (CPU) and memory selection.
The authors investigate the use of performance monitoring counters (PMC) for CPU performance profiling, and characterize the performance trends with different analytical models such as cache miss ratios and processor service time.
Overall, this paper is a good starting point for researchers mainly focusing on cluster performance dependent on the CPU and memory selection for on-demand data-intensive applications.