基于多层次并行算法的共成像点道集抽取

Common Image Gathers Extraction Based on Multi-level Parallel Algorithm

研究地表区域勘探和开发,应依靠地震偏移成像算法。在地震勘探中,利用叠前深度偏移抽取的共成像点道集为修正偏移速度提供了一种工具,然而庞大的计算量严重影响生产速度。为提高数据处理速度,提出了多层次并行共成点道集抽取算法,在集群节点间利用MPI并行,每个节点内部采用OpenMP并行,从而充分发挥每个节点的运算性能。实验结果表明,基于多层次并行的共成像点道集抽取具有良好的并行加速比和并行效率,有效提高了偏移速度分析的计算效率,为地震资料处理提供了更优的并行策略,并为油气勘探提供有效参考依据。

In seismic exploration,common image gathers extracted by pre-stack depth migration provides a powerful tool for correcting migration velocity,but the huge computation seriously affects its application.Based on data independence of pre-stack depth migration and common image gathers,the paper presents the method of multi-level parallel common image gathers extraction.The method processes in parallel using MPI among cluster nodes,and using OpenMP in every node,thereby developing the operational performance of each nodes.The experimental results show that the method has better parallel speedup and parallel efficiency,improves computational efficiency of migration velocity analysis,provides a better parallel strategy for seismic data processing,and has important significance on actual production.