Porting a simple 2-D migration program
to the Connection Machine

Steve Cole

Author has no known email address

ABSTRACT

As a learning exercise, I ported a simple 2-D implicit finite-difference poststack migration program to the Connection Machine. Getting the program up and running was not difficult, but figuring out how to get it to perform well was something of a challenge. Currently the program runs roughly 25 times faster than the equivalent well-vectorized program running on a Convex C-1. Since the algorithm itself is very familiar to most readers, those with no prior Connection Machine experience might find this review of my experiences a useful introduction.