Despite everything that’s known and theorized about space, anyone can
tell that the information is just a fraction of what’s really out there.
Getting the answers is the main goal of ASTRON, the Netherlands
Institute for Radio Astronomy, when it called for IBM’s help in creating
a supercomputer.
Why a supercomputer? Why not a satellite or a telescope? Because a telescope already exists- it’s called the Square Kilometer Array or SKA, a square kilometer of flat land full of antennas that reach the millions. While it’s still in development, it will never be able to look into the far reaches of outer space if the 20 funding countries don’t find some way of getting something capable of creating the processing power the SKA needs to work, and this is where ASTRON and IBM’s collaboration comes in. It’s called DOME, named after the Swiss mountain and the covering over telescopes, and it’s basically an exascale computing machine that will be a hundred times faster and more powerful than the fastest supercomputer to date, Fujitsu’s K computer, a petascale running at 10 petaflops.
For those who don’t understand technical jargon, FLOPS is short for floating-point operations per second, the measurement of a computer’s performance. Petascale computing is the next step up from terascale and a step below exascale, which is far more than what normal PCs can boast as they’re only at gigascale computing. A better analogy was presented by IBM- imagine that the daily traffic of all the computers in the world is doubled, and that will be the rough estimate of DOME’s own daily output.
With DOME, scientists can analyze, run and store the huge amount of data that the SKA will receive with its square kilometer of antennas that collect radio waves, which is expected to amount to a few exabytes per day or about 1500 petabytes per year. Obviously, borrowing or replicating Fujitsu’s toy for the project is out of the question.
Of course, creating DOME will not be an easy task for both ASTRON and IBM, so they’ve launched a five-year collaboration to create the supercomputer before 2024, the year the SKA is supposed to be finished. If they get this done in time, and the two work as expected, it might not be long before we find out if there really is other life out there.
Why a supercomputer? Why not a satellite or a telescope? Because a telescope already exists- it’s called the Square Kilometer Array or SKA, a square kilometer of flat land full of antennas that reach the millions. While it’s still in development, it will never be able to look into the far reaches of outer space if the 20 funding countries don’t find some way of getting something capable of creating the processing power the SKA needs to work, and this is where ASTRON and IBM’s collaboration comes in. It’s called DOME, named after the Swiss mountain and the covering over telescopes, and it’s basically an exascale computing machine that will be a hundred times faster and more powerful than the fastest supercomputer to date, Fujitsu’s K computer, a petascale running at 10 petaflops.
For those who don’t understand technical jargon, FLOPS is short for floating-point operations per second, the measurement of a computer’s performance. Petascale computing is the next step up from terascale and a step below exascale, which is far more than what normal PCs can boast as they’re only at gigascale computing. A better analogy was presented by IBM- imagine that the daily traffic of all the computers in the world is doubled, and that will be the rough estimate of DOME’s own daily output.
With DOME, scientists can analyze, run and store the huge amount of data that the SKA will receive with its square kilometer of antennas that collect radio waves, which is expected to amount to a few exabytes per day or about 1500 petabytes per year. Obviously, borrowing or replicating Fujitsu’s toy for the project is out of the question.
Of course, creating DOME will not be an easy task for both ASTRON and IBM, so they’ve launched a five-year collaboration to create the supercomputer before 2024, the year the SKA is supposed to be finished. If they get this done in time, and the two work as expected, it might not be long before we find out if there really is other life out there.
0 comments:
Post a Comment