摘要
本文详细阐述了并行计算技术及其在地球物理勘探数据处理中的发展现状和发展趋势,分析了几个代表性的并行算法实例.这些结果表明,在拥有强大的并行机的基础上,基于并行计算开发环境(MPI和PVM等)设计高效的并行算法,通过分配合理的并行粒度、通信开销、负载平衡等执行高效的并行计算,可以有效加快处理速度、降低成本.目前,并行算法在地震数据处理中应用已较为成熟,近年来向更实用的基于PC机群的并行技术发展.然而,在非地震方法中,并行算法应用较少见文献报道,研究尚处于初级研究阶段.在大地电磁的二维和三维正、反演问题上,并行计算技术逐渐得到越来越多关注和重视.随着资源和能源需求的增长,地球物理勘探向深度和广度快速发展,大幅增长的数据量使得高性能并行计算机和高效的并行算法在勘探地球物理学中的发展和应用将占据愈来愈重要的地位.
We present the parallel computing technology and its development status in geophysical exploration data processing,and analyze some representative parallel algorithm examples in geophysics.Parallel computing can effectively speed up the processing and reduce the cost based on powerful parallel computer,design efficient parallel algorithm in the parallel computing exploiting environment(MPI,PVW,etc)by allotting rational parallel granularity,communication cost,and load balance.The application of parallel algorithm is quite mature in the seismic data processing,which is developing ahead toward the more practical parallel technology based the PC cluster in recent years.But in the non-seismic aspects there are few reports on the parallel algorithm,and the research is still at the initial stage.For the 2D and 3D MT inversion,the parallel technology is getting more and more attention and recognition.With the demand growth of resources and energy,the exploration development to the depth and extent has resulted in the rapid increase of the data,which promotes the development and application of high performance parallel computing(HPC)technology in geophysics.
出处
《地球物理学进展》
CSCD
北大核心
2010年第2期642-649,共8页
Progress in Geophysics
基金
油气资源与勘探技术教育部重点实验室(长江大学)开放基金项目K200608
关键词
并行计算
机群
MPI
PVM
地球物理勘探
数据处理
parallel computing
cluster
Message Passing Interface(MPI)
Parallel Virtual Machine(PVM)
geophysical exploration
data processing