What would you do with a supercomputer hooked up to a five- by 12-foot screen?
Probably something less altruistic than the staff at the University of Toronto‘s Molecular Design and Information Technology (MDIT) centre, which specializes in bio-molecular structure-based research. Through computer-assisted design, the staff tries to understand drug interactions and designs.
Less than a year old, the centre is already getting an upgrade, the school said Wednesday. The Silicon Graphics Inc. (SGI) Onyx 3800 visualization supercomputer is getting eight more 600 Mhz processors (bringing the total to 44), an Octane2 workstation and monitors.
MDIT director and U of T assistant professor Dr. Lakshmi Kotra says the nature of the research demands these kinds of improvements.
“”The majority of our work focuses on understanding bio-molecules and complex receptors,”” he explains.
“”These are extremely complex structures and in our type of business we have to be able to see them to get a grasp on these systems to design a new compounds and understand what is going on at the atomic level.””
“”The University of toronto is really a leading organization. They’ve made a commitment to exploring how high-end visualization is going to make a difference,”” says Juli Moultray, biology market manger at SGI.
The payoff of this kind of computing and graphics power is the ability to do collaborative work. The system can produce and project stereo images onto the big screen where groups of 10 to 15 research can examine the results, according to Kotra.
Moultray says MDIT is pushing the boundaries of drug research, making it difficult to predict future computing needs.
“”What we don’t know is where this will take us?”” Moultray says. “”For these leading edge customers we cannot predict what their growth challenges are going to be.””
While the future needs of the centre are unclear, the present ones are clear. IDC Canada analyst Alan Freedman says this kind of research demands systems with a lot of throughput, stability and processing power. “”A lot of the more intensive workloads like this are moving towards resource sharing. In most cases that would be a grid system where you can dynamically control the workload going through,”” Freedman says.
Kotra says part of his research includes designing anti-cancer compounds and insulin mimetics. He adds none of this research would be possible without partnering with Tripos Inc. and funding from the Ontario Innovation Trust.
Comment: [email protected]