Performance analysis of Intel Ivy Bridge and Intel Broadwell microarchitectures using Intel VTune amplifier software

摘要

In recent years the thread level parallelism has emerged as more competent technique for enhancing the processor's performance. As, it was not possible to exploit the instruction level parallelism any further and the heat dissipation problem associated with complex larger processors had become a serious issue. Thus, the multithreading approach was encapsulated in the processor architectures to exploit the thread level parallelism together with instruction level parallelism for enhancing the processor performance. This has encouraged the development of chip multicore processors. The current study analyzes the ability of the Intel Ivy Bridge and Intel Broadwell micro-architectures to handle the branch instructions and cache memory. In this study, all the considered processors were subjected to test runs under the observation of Intel VTune amplifier software that was used to study the branch mispredictions, cache misses and to gather the respective data.