论文部分内容阅读
为了提高机载设备振动环境实测数据处理效率,提出了一种并行化数据处理和振动环境谱编制方法。在SMP(symmetrical multi-processing)集群系统的多级体系结构下,以Welch(改进周期图法)算法为基础,进行了振动数据处理模型的并行化分析,对于并行化过程中存在的并行I/O、负载平衡等关键问题进行了讨论,提出了相应的解决方案。最终选择基于MPI/Open MP混合方式实现了算法。在搭建的集群环境下测试表明,12核参与运算的条件下,最高加速比可以达到7.4,有效提高了运算效率。
In order to improve the measured data processing efficiency of airborne equipment vibration environment, a parallel data processing and vibration environment spectrum compilation method is proposed. Under the multi-level architecture of SMP (symmetrical multi-processing) cluster system, the parallel analysis of the vibration data processing model is carried out based on the Welch (Improved Periodogram Method) algorithm. For parallel I / O, load balancing and other key issues were discussed, put forward the corresponding solutions. The final choice is based on the MPI / Open MP hybrid implementation of the algorithm. Tests in a clustered environment show that under the condition of 12 cores participating in the operation, the highest speedup ratio can reach 7.4, which effectively improves the operation efficiency.