This summer a student and a graduate from the Department of Computer Science took a very interesting road trip. They are part of Memorial's research team working in the Landmark Visualization Laboratory and in August research assistant/programmer Dave Churchill and computer science/pure math undergraduate student Matthew Hamilton, along with their supervisor Dr. Phil Bording, Memorial's Husky Energy Chair in Oil and Gas Research, travelled to IBM office in Cambridge, Mass. to work with IBM's Deep Computing research team. Dr. Kirk Jordan of IBM hosted their visit. Because Memorial's visualization lab is driven by IBM-based systems, the students were hoping to get some knowledge and experience in the fields of scientific parallel computation and large scale three dimensional visualization. They want to build systems to allow researchers to more readily perform useful visualizations. “Visualization can be a powerful tool but we must learn how to use the facilities we have in order to make most effective use of this tool. That was our goal when we visited IBM,” said Mr. Hamilton. The students worked side by side with IBM researchers and staff on two IBM systems.

The first of these systems is IBM Deep Computing Visualization (DCV), coupled with Scalable Visual Networking (SVN), a method that allows large scale three dimensional visualizations to be viewed on multiple synchronous displays. “We used the DCV system coupled with four nine million pixel IBMT221 displays running on a quadruple screen to perform extremely high resolution imaging of Dr. Bording's seismic models - a data cube representing the waves generated by a seismic shot record using software we had written prior to our trip. At IBM we got a high-quality display that we can use as a test bed in a less careful and expensive fashion than the visualization lab. But it gave us very valuable information we can now apply to the Landmark Visualization Laboratory here at Memorial,” said Mr. Churchill.

The second of these systems was the IBM Blue Gene supercomputer at IBM's TJ Watson Research Center in New York. According to TOP500 Supercomputer Sites, this is ranked the fastest in the world, at approximately 140 teraflops, or 65,000 times faster than your average desktop computer. “A smaller test version of the Blue Gene machine was used to run Dr. Bording's 2-D and 3-D seismic modelling code and it would have taken weeks, if not months or years on a normal computer but only an hour to run on this supercomputer,” said Mr. Hamilton. Mr. Churchill and Mr. Hamilton met with several experts in various fields of computation and say the experience will help them in building applications for Memorial's visualization laboratory.