Transcript
Last November, a computer from China ranked first on the TOP500 list of the world's most powerful supercomputers. The Tianhe-1A system at the National Computer Center in Tianjin, China was the first Chinese supercomputer to take the top spot on the list.
But now, University of Florida researchers have created a supercomputer, called Novo- G, that can perform some important science applications even faster than the Chinese supercomputer. According to Alan George, professor of electrical and computer engineering, Novo-G is the world’s fastest reconfigurable supercomputer.
George, who is also the director of the National Science Foundation’s Center for High-Performance Reconfigurable Computing, known as CHREC, explains that the TOP500 list ranks supercomputers by their performance on a few basic routines in linear algebra using 64-bit, floating-point arithmetic. But many key applications in health and life sciences, signal and image processing, and financial science do not follow that standard.
What is truly significant about Novo-G is that it is not just another incremental innovation in supercomputing. It is an entirely different technology.
As reported in the scientific journal IEEE Computing in Science and Engineering, most of the world’s computers, from smart phones to laptops to Tianhe-1A, are based on microprocessors with fixed-logic hardware structures. Because software applications for these systems must conform to these fixed structures, a lot of energy is wasted and the potential for faster speed is sacrificed.
In a reconfigurable machine, however, the architecture can actually adapt to suit the demands of each application, which translates into more speed and less energy consumption. In fact, the researchers claim that the Novo- G, which uses 192 reconfigurable processors, "can rival the speed of the world’s largest supercomputers at a tiny fraction of their cost, size, power, and cooling."
Many of today’s supercomputers are as big as an entire building. They burn millions of watts of electricity, which produces enormous levels of heat. By contrast, Novo-G is only about twice the size of an ordinary refrigerator, and it uses less than 8,000 watts of energy.
Even better, unlike conventional supercomputers, Novo-G can easily be scaled up. By the end of this year, the University of Florida team plans to double its capacity by making only small upgrades in its size, power, and cooling.
This technology promises to revolutionize many areas of science, including gene sequencing, cancer diagnosis, plant science, and the ability to analyze large data sets.
|
