Full text loading...
-
A Data-ware Scheduling Framework on GPU Grid for Kirchhoff Prestack Depth Migration
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 75th EAGE Conference & Exhibition incorporating SPE EUROPEC 2013, Jun 2013, cp-348-00198
- ISBN: 978-90-73834-48-4
Abstract
MapReduce is a famous programming model for processing large data sets on clusters of computers. However, for pre-stack depth migration (PreSDM), which is a classic migration imaging method in geophysical domain, it shows great inadaptation due to the computational characteristics of the problem. In this paper, an improved programming model is introduced especially designed for Kirchhoff PreSDM and a Data-ware scheduling policy is considered to maximized the utilization of computational and I/O capacities of GPU grid. The implemented framework is tested on actual examples and results show the effectiveness of our model and strategy in comparison with Apache Hadoop which is a popular free implementation of Mapreduce.