The “Big Ben” Cray XT3 at University of Pittsburg Computer Center runs at 10 Teraflops
Ok if you aren’t computer savvy. you might think “tri-tera-flops…a dinosaur?” or maybe “Tera Flops” which could be a headline about a movie star with a bad gig.
But FLOPS which stands for FLoating Point Operations Per Second is a measure of the numerical calculation power of a computer, be it a PC or a mainframe. Supercomputers are often measured in Teraflops or Trillions of Flops per second.
And now, its all in one chip. Intel has produced a prototype, shown below:
The Teraflops chip is built on a single die composed of 80 independent processor cores, or tiles as Intel is calling them. The tiles are arranged in a rectangle 8 tiles across and 10 tiles down; each tile has a surface area of 3 square millimeters. And that isn’t all, in a unique manufacturing technique, each die is “stacked” in 3 dimensions, making the chip multi-layed with components not only in breadth, but in depth.
The chip can operate at a number of speeds depending on its operating voltage, but the minimum clock speed necessary to maintain its teraflop name is 3.13GHz at 1Volt. At that speed and voltage, the peak performance of the chip with all 80 cores active is 1 teraflop while drawing 98 Watts of power. At 4GHz, the chip can deliver a peak performance of 1.28 TFLOP, pulling 181 Watts at 1.2Volt. On the low end of the spectrum, the chip can run at 1GHz, consuming 11 Watts and executing a maximum of 310 billion floating point operations per second. Heat dissipation for this chip will be huge, so I expect a water cooling system will be used to acheive the peak performance.
It may be 1 or 2 years, but you’ll soon be able to have your own Teraflop Desktop.