Thought "dual core" microprocessors sounded impressive when they hit the streets a few years back? Forget about it. Intel (INTC) today introduced a chip with 80 "cores," or 80 distinct microprocessors arrayed in a microscopic grid, that collectively can churn through a trillion "floating point operations" a second. That makes it the world's first "teraflop" chip. FYI: "tera" comes from the Greek word for "monster."

OK, so this hot rod is not quite ready for prime time. It's still being shown off by the guys in lab coats. More importantly, John Markoff, among others, points out that nobody really knows how to write programs that take advantage of so many cores. It's a problem that has afflicted the world of "massively parallel" supercomputers for some time. It seems the hardware geeks are way ahead of the software types, and so Slashdotters have met today's news with a certain degree of frustration and self-flagellation. "We are doing something wrong," writes Ardor. "The human brain provides compelling evidence that massive parallelization works."

That's certainly what Intel believes, and it has been telling reporters it believes the new chip will excel at software problems involving "recognition, mining and synthesis." Not sure if all this is cool or creepy.

For the record, it's only been ten years (that's 1997) since the supercomputers first broke the teraflop barrier. The machine that did it was the ASCI Red at Sandia National Laboratories. Today's supercomputing champ is IBM's BlueGene/L, a machine capable of 280.6 teraflops. Want more? Check out The Smartest Machines on Earth....