The performance analysis and research of sorting algorithm based on OpenMP

On the basis of OpenMP relevant technology, under the condition of combination of OpenMp with the Bubble Sort algorithm, OpenMp scheduling and load balancing issues are discussed. The parallel programming of Bubble Sort algorithm was implemented using the oMPi compiler in the Linux Ubuntu environment.Taking advantage of the program level instrument method to monitor the total time consuming and each thread time overhead under four kinds of scheduling strategy in a parallel program segment. The big difference in these four kinds of scheduling strategy is shown. From the result of the experiment we can see, load balance spending has greater impact on performance than scheduling overhead in the decreasing cycle with small loop body and many iterations. Dynamic scheduling with smaller block should be taken to release load imbalance and the execution efficiency and performance of the algorithm will be enhanced greatly.