如何剖析OpenMP瓶颈 [英] How to profile OpenMP bottlenecks

问题描述

我有一个已经由OpenMP并行化的循环，但由于任务的性质，有4个临界子句。

I have a loop that has been parallelized by OpenMP, but due to the nature of the task, there are 4 critical clauses.

什么是最好的方式来配置加速，找出哪些关键子句（或也许非关键（！））占用了循环内的最多时间？

What would be the best way to profile the speed up and find out which of the critical clauses (or maybe non-critical(!) ) take up the most time inside the loop?

我使用Ubuntu 10.04与g ++ 4.4.3

I use Ubuntu 10.04 with g++ 4.4.3

推荐答案

OpenMP包括函数omp_get_wtime ）和omp_get_wtick（）用于衡量时间效果（ docs这里），我建议使用这些。

OpenMP includes the functions omp_get_wtime() and omp_get_wtick() for measuring timing performance (docs here), I would recommend using these.

否则尝试一个profiler。我更喜欢可以在此处找到的google CPU分析器。

Otherwise try a profiler. I prefer the google CPU profiler which can be found here.

还有这回答。

这篇关于如何剖析OpenMP瓶颈的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！