What this system does is allow scientists to see and hear their research.
Like Mr. Sulu piloting the Enterprise, one researcher guides the docking of an experimental drug molecule to its human protein target. What happens on screen is exactly what would happen at the molecular level inside a human body. On screen, thousands of bio-chemical reactions unfold as a single event, in real time, as they would inside you or me. How will the protein react to the new drug? What effects will it produce? What unexpected results?
Few devices on earth can solve for the immense complexities of biochemical interactions. Fewer still can show them, interactively, in real time. This is the kind of job that PowerParallel systems like Deep Blue and CAVE were created to do. It is a short step from the intricacies of grandmaster chess to the almost infinite compexities and interactions that are part of the daily life (and death) of even the smallest living organism.
For researchers - in AIDS, say - it is handy as well as humane to have a machine simulate some of these effects. As Kevin Kelly points out, "computerized entities don't have bodies."
Computers have been cranking away on biomedical problems for years, but until parallel computing became practical in the 90s, the most important problems defied the biggest machines. A processor that does only one thing at a time - even at a rate of millions per second - is going to take a while when there are 1023 or 1035 things to do. Maybe forever. But get a whole swarm of processors working in parallel, all at once, and you can chainsaw through some pretty big numbers.
Parallel computing brings insoluable problems within solution-range; and makes solutions to difficult problems routine. None of the 128 PowerParallel RISC processors in the CAVE system is particularly exotic or expensive. Nor is it a major deal to add more to the gang, stepping up system performance by orders of magnitude. Deep Blue is 1,000 times faster than its immediate predecessor.
Creating software for a massively parallel system is almost biologically complex itself. So it's no coincidence that parallel systems are playing a role in accelerating the development of code for ... parallel systems! Funny how it all seems to fit, with an almost ecological rightness. Listen to MIT's "computer/biologist" Tom Ray "Look," he says, "ecological interactions are just parallel optimization techniques. A multi-cellular organism essentially runs massively parallel code of an astronomical scale."
Some of humanity's oldest unsolved problems - and some of the newest, too, like AIDS - are now beginning to yield to systems like Deep Blue's cousin at CAVE. Which makes you think. "Virtual reality" can provide scientists with powerful insights to the real thing. And Garry Kasparov is righter than he knows when he says that the chessboard and life are the same.
[ The Game | The Program | Commentary | Grandstand | Deep Blue ]
[ IBM home page | (C) | (TM) | IBM Small Planet Pavilion | Internet 1996 World Expo ]