A Chinese "supercomputer" has been confirmed as the fastest in the world, the first time the country has taken the honor. But the United States is currently working on two machines that could be 10 times faster.

The ranking comes from the TOP500 project, which produces a list twice a year to show how quickly computers have been proven to work for a sustained period. This is usually much lower than the computer's theoretical maximum speed (which is used in the rare case of tiebreak situations).

Japanese, American Dominance Ends

Since the list began in 1993, only the United States and Japan have housed machines that get the number one slot. The last three machines to hold that crown, dating back to 2004, have all been American-made. The "defending champion" going into this month's list was a machine named Jaguar, housed in Oak Ridge, Tennessee, which had a recorded speed of 1.759 petaflops. That speed represents 1.759 trillion calculations every second.

The new leader is the Chinese-made Tianhe-1 (which translates as "Milky Way"), which has a recorded speed of 2.507 petaflops. Reports of that speed had come out of China a couple of weeks ago, but have now been confirmed by the list's publication. The machine has undergone the mother of all upgrades, having only recorded a speed of 0.563 petaflops in May's list.

20 Petaflops The New US Goal

The Oak Ridge facility (the National Center for Computational Sciences) says it is working on a supercomputer that, in about two years time, will boast a maximum speed of 20 petaflops. That's the same speed and date target being pursued in a separate project by IBM in California. (Source: computerworld.com)

Amazingly, if the machines were able to reach 20 petaflops, they might only just be drawing level with the maximum "processing" speed of the human brain. Of course, the major difference is that the brain usually only carries out calculations at this speed for a matter of seconds while computers can go as long as they have power. (Source: ualberta.ca)

These supercomputers aren't simply designed to win prestige, though. They perform tasks that involve relatively simple calculations but model situations that have a gargantuan number of possible variables and outcomes. Common examples include modeling potential weather patterns or the effects of events such as flooding and other natural disasters.