|
Hyper-drive for computers By Brice Wallace Kent Gilson, chief technology officer for Star Bridge Systems, works on a program, designed in connection with the National Cancer Institute, that facilitates the study of the human genome. A Midvale company uses those key words and hopes someday to create a world just as revolutionary as the futuristic technology depicted in that old TV show, one that uses the company’s high-powered computing technology as ubiquitously as today’s world uses the microwave oven. Officials at Star Bridge Systems Inc., established in 1998, believe their Hypercomputers will quickly gain a market in business sectors requiring quick computation of masses of data. Armed with blazing speed, a relatively small size, low power consumption and programming flexibility, the Hypercomputers, they believe, can revolutionize concepts about what computers can do. “The essence of what we sell is time,” said Ed McGarr, vice president of sales and services. “We enable things that formerly would take months to calculate to be done in minutes.” “This is a technology that will permeate the tech industry over time,” said Brent Ward, vice president of operations. Lofty talk, but not without substance. While the company has spent most of its time in research and development, it has sold Hypercomputers to NASA, the Air Force, the U.S. National Security Agency and several universities and is partnering with the National Cancer Institute to develop advanced bioinformatics processing machines. Ed Bradley, senior engineer in the munitions branch for the Air Force Research Laboratory at Florida’s Eglin Air Force Base, obviously is a believer in Star Bridge’s technology, saying it “is eventually going to change the way everything happens in the computing world.” So, how fast is a Hypercomputer? Company execs say the technology is so different from conventional computers that performance measurements and benchmarks don’t apply. Kent Gilson, founder and chief technology officer, said Hypercomputers’ parallel processing allows them to do perhaps millions of operations simultaneously within 100 nanoseconds. Image Laura Seitz, Deseret Morning News This board is the “brains” of speedy and flexible computers developed at Star Bridge Systems in Midvale. The estimated peak performance of top-line Hypercomputers puts them among the world’s 200 fastest computers, comparable to multimillion-dollar systems. The heart of the Hypercomputer is its use of chips called field programmable gate arrays, or FPGAs. Seen as an eventual replacement for microprocessors, FPGAs have been around for nearly two decades, but Star Bridge’s technology allows them to tackle separate problems simultaneously rather than employing traditional serial, or one-at-a-time, computing. Gilson’s parallel processing approach was novel. “The analogy I like to use is that it’s like microwave technology, which existed for years and back in the ‘60s somebody said, ‘Hey, we can bake a potato with this sucker,’ ” McGarr said. “That was the birth of the microwave oven, and now, 40 years later, everything is microwave-ready.” What’s more, Star Bridge’s Viva software programming language and environment is designed to allow users to quickly develop new applications and improve existing ones without a lot of hassle. The thinking is that scientists can program a Hypercomputer for what they want or need, and where and when they want it done. McGarr described microprocessors as Swiss Army knives — jack-of-all-trades and master of none — but Star Bridge’s technology is a “silicon white board.” “You can design a specific program on our computer, then erase it and reconfigure it into something else. When you use a computer, it takes your computer and burns it into hardware while you need to use it. It ‘makes’ a computer every time. The software difference is a high-level, graphical user interface to allow people to create their programs that in the past have traditionally been very difficult and arduous to develop using these programmable gate arrays.” “We’re creating tools and applications that will allow solutions to problems that before were impossible or not practical to solve,” Gilson said. “This will allow us to translate our ideas into solutions as fast as we can think.” So how might this help John Q. Public? By quickly distilling data in life sciences research, Hypercomputers might help scientists be faster in investigating diseases and developing cures. If deployed on smart weapons, collateral damage might be minimized. Oil and gas exploration companies using data- and computation-heavy algorithms would be able to get results from their exploration tests quicker and more accurately, leaving consumers less likely to have to bear costs of unsuccessful exploration projects or poor reservoir management. “Our mission is to help everyone to gain knowledge — even truth — from all the information out there,” said Daniel Oswald, president and chief executive officer. “That’s what it’s for. In all these sciences, it’s become more and more computation and less and less scientific research in terms of lab work.” Because hardware chips can be reconfigured remotely, NASA sees possibilities of satellites or spacecraft being programmed after being deployed rather than having to stick solely to tasks programmed prior to launch. McGarr envisions cameras on car mirrors that would sound an alarm to awaken a drowsy driver if the device notices the driver’s eyelids getting heavy, or a sensor that could alert the driver if a child moves behind a car backing out of a driveway. Because of their relatively small size, Hypercomputers easily could be wheeled into operating rooms to monitor patients’ conditions. “Five years from now, we would expect to be the development platforms that would represent imbedded computing, from refrigerators to toilets to rocket ships to guitars — you name it,” he said. Star Bridge has come a long way since it was operated out of Gilson’s home. “I realized the only thing worth spending your time on were things that basically affected the basic economics of doing something that everyone has to do,” he said. “So I built a computational substrate that’s more efficient, that would change the envelope for doing computing.” Disbelievers abounded. “Not many people knew what we were doing. There was so much skepticism and so much new that people didn’t know what was going on. It was so out in left field, people thought it was a lie or somebody smoking wacky weed or raw insanity,” he said. Not only is the technology complex but, in essence, is revolutionizing computing from its basics. “We’re competing with 50 years and millions of people, and that’s an overwhelming infrastructure, and we’re coming in and saying, ‘There’s a better way to do it,’ ” Gilson said. But Ward, who was with Gilson during the company’s early days, said it’s starting to pay off. “I had the ultimate faith that this philosophy would blossom over time, and it’s getting there now,” Ward said. Star Bridge already is in the geophysics, biotechnology and national security/defense fields, and applications spun out from there will become products for commercialization. Those early adaptors of the technology, Oswald said, “become evangelists in their sphere of influence.” The company has seen its work force grow from 12 to 21 since November and already has sold second-generation systems — not much bigger than a desktop computer, scalable to their tasks and using regular electricity outlets. “It’s kind of a fairy tale,” Gilson said. “You’ve got the whole world saying it’s impossible or it’s a lie, and then you have the cream of the crop coming out saying, ‘Hey, that’s pretty cool.’ ” The company will grow as money becomes available, and Oswald said external funding will be sought this summer. The company also will seek partners that have a “long” view, he said. “We’re not saying our technology is where we want it to be. You can always improve. It takes a little bit of luck and hard work and a unified team. If it all comes together as we envision it, we’ll be successful,” Oswald said. “There’s a lot of work to do, and we’re just beginning this journey. As Kent says, it’s not a sprint; it’s a marathon. And we’re ready for the marathon.” And, Gilson suggests, it’s a marathon with many possible finish lines. “Hypercomputing basically changes the envelope for doing any kind of information system processing. If you’re optimizing for size or speed or cost or package size or whatever, this is a more economically viable solution to doing any kind of information system processing,” he said. “And it’s not just things we’re doing right now, but things we haven’t even thought of. Those are the really exciting things — things we can’t even think of because we’re climbing up the mountain in the fog and can’t see what’s ahead yet.” E-MAIL: bwallace@desnews.com Back to Publications |
|
