Partitioning Algorithm of 3-D Prestack Parallel Kirchhoff Depth Migration for Imaging Spaces

In the process of treatments to large-scale 3-D prestack Kirchhoff depth migration, the required memory for imagining spaces is large, even more than the total memory of the single node. So this paper proposes the partitioning algorithm based on the process groups in the environment of distributed memory. Partitioning to the imaging spaces is going on in a group, and the processes of the group achieve dealing with migration to the entire imaging space through cooperation with each other. The treatments to the imaging spaces among groups are independent, and the dealing is to the different seismic data. The actual test results show that the proposed partitioning algorithm of the imaging spaces in the paper is feasible. It can efficiently solve the problem of memory needs for large scale 3-D prestack Kirchhoff depth migration.