Dr. Bording presented a presentation on a technology, which is quite different from most popular computer technology. He began his presentation by explaining the materials that he had on display. Some of the display included a Cray YMP CPU, a XMP circuit module and the CPU of an early model of an IBM system. He talked about the advances that are currently being made in hardware and software development and some of their limitations. One of his main concern was with main-memory Vs cache and how industry leans toward the idea that the more speed you have, the more power you have. He then pointed out that the access rates to main memory have changed very little over the last 15 - 20 years. Dr. Bording pointed out that one method that can be used to improve the performance of software via software is by changed the way inner loops are handed, i.e. loop unrolling. He referenced a book, "Inner Loops", by Rick Booth which describes how Mr. Booth, who develops video games, was able to speed up his process time from 23 to 2.9 cycles by modifying the loop structure in his assembly code.

Dr. Bording's main idea is that he believes that for certain applications, i.e. seismic processing, massively parallel machines with large numbers of processors will become the method of choice instead of machines with faster and more level of cache - the tortoise Vs the Hare. After giving us a review of the history of hardware development (that was enlightened with first hand experience), he discussed Penta-flop computing and how it relates to sizing the seismic problem to fit the computing resource. We were informed that recently a replica of the Babbage Difference Engine was built in England to celebrate the birthday of Charles Babbage who built the first digital computer, the Difference Engine, in the early 1800s. Many benefits have been obtain from the rebuilding of the replica which was build using the same machine tools used during the earlier era.

Dr. Bording concluded his presentation after he introduced us to the wave equation difference engine (WEDE) which is a prototype parallel computing machine constructed (by Dr. Bording) to solve a finite difference approximation to the acoustic wave equation. The WEDE prototype demonstrates:

• The notion of an application specific wave equation computer is feasible.
• It is now possible to build large scalable parallel wave equation engines for specific seismic algorithms.
• The existing difficulties of existing parallel computers for speedup can be overcome.

Dr. Bording predicts that application specific computing will become the method of choice for seismic processing in the future.