异构集群环境下逆时偏移任务调度算法-《计算机技术与发展》

文章信息/Info

Title:: Scheduling Algorithm of Inverse Time Migration under Heterogeneous Cluster Environment

作者:: 高新成¹ ; 刘德聚² ; 王莉利² ; 李强² ; 柯璇³; 1. 东北石油大学现代教育技术中心,黑龙江大庆 163318;
2. 东北石油大学计算机与信息技术学院,黑龙江大庆 163318;
3. 东北石油大学地球科学学院,黑龙江大庆 163318

Author(s):: GAO Xin-cheng¹ ; LIU De-ju² ; WANG Li-li² ; LI Qiang² ; KE Xuan³; 1. Modern Education Technology Center,Northeast Petroleum University,Daqing 163318,China;
2. School of Computer and Information Technology,Northeast Petroleum University,Daqing 163318,China;
3. School of Earth Science,Northeast Petroleum University,Daqing

Keywords:: heterogeneous cluster; inverse time migration; task scheduling; load balancing; utilization rate of system resource

摘要:: 逆时偏移方法作为目前最先进的地震资料成像方法之一,已经广泛应用于地震数据成像领域;基于地震资料的庞大数据量,该方法仍存在计算需求较大的问题,通常需要借助集群系统来完成运算。在异构集群环境中,各个节点的性能不同,节点的处理能力也会存在差异,在进行数据运算时容易出现负载不均衡的现象。为了提高并行计算的工作效率和异构集群系统资源的利用率,结合负载均衡技术,提出了一种异构集群环境下的自适应节点两级计算任务调度算法,将节点间和节点内的计算任务尽可能合理地划分。通过实验验证,同传统的 Min-Min 和 Max-Min 算法进行对比,在对逆时偏移数据进行处理时,该算法能够有效地缩短整体计算任务的完成时间,使得各个节点的计算任务分配更加均衡,提高了整个异构集群系统的资源利用率。

Abstract:: As one of the most advanced seismic data imaging methods,the inverse time migration method has been widely used in seismic data imaging field. Based on the huge amount of seismic data,this method still has the problem of large computational demand,usually needs to use the cluster system to complete the calculation. In the heterogeneous cluster environment,the performance of each node is different,and the processing capability of each node is also different. When performing data operations,load imbalance is likely to occur. In order to improve the utilization rate of resources and the efficiency of parallel computing, we propose a two- level adaptive node computing task scheduling algorithm for heterogeneous cluster systems, which divide computing tasks between and within nodes as reasonably as possible. Compared with the traditional Min-Min and Max-Min algorithms,the proposed algorithm can effectively shorten the completion time of the computing task,make the computing task of each node more balanced,and improve the resource utilization rate of heterogeneous cluster system in the data processing of inverse time migration.