Optimization of Reverse Time Migration for Fast and Scalable Seismic Imaging on Blue Gene

We present our optimization work on a Reverse Time Migration (RTM) code for fast and scalable seismic imaging on Blue Gene. RTM is a seismic imaging algorithm increasingly used in industry for oil exploration. Our work is novel not only in that it uses the Blue Gene supercomputing systems torus network, high inter-node communication bandwidth, and large resident memory in each node to convert an embarrassingly parallel problem into one that can be efficiently solved using massive domain partitioning; but it is also novel in that it develops effective schemes to keep intermediate data snapshots in internal memory and fast interpolation for sub-sampling that eliminate bottlenecks in the forward and backward pass and significantly improved performance while maintained image quality. With further optimizations in boundary condition calculation, OpenMP, load balance, MPI communications, etc., our final code achieved ~10X overall performance improvement over the base line code. The success of our optimization can be easily extended to next-generation imaging algorithms currently being developed. In the workshop, we will include a BG/Q architecture overview and comparative BG/Q results depending on availability.

By: Ligang Lu, Lurng-Kuo Liu, Karen Magerlein, Pascal Vezolle, Michael Perrone

Published in: RC25313 in 2012


This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.


Questions about this service can be mailed to reports@us.ibm.com .