With computing technology continually getting smaller, thinner and faster in order to fit in our pockets and on our laps, you may look at modern day room-sized supercomputers and balk. But if you're impressed with all the things your tablet can do, imagine the power of a computer so big that it needs its own room.
The announcement that the U.S. Department of Energy is planning to spend $425 million to build two supercomputers that are 5 to 7 times faster than any supercomputer in history is big news, and it could lead the way to major advancements in science research, reports Reuters.
Most of that cash, $325 million of it, will be designated for the construction a supercomputer to be named "Summit," for Oak Ridge National Laboratory in Tennessee, and another to be named "Sierra" at the Lawrence Livermore National Laboratory in California. The remaining $100 million will go toward research into future supercomputing advances to eventually build computers that are even faster.
Summit and Sierra will operate at 150 petaflops and 100 petaflops respectively, which is the unit of measurement for supercomputer performance. Basically, that means that the machines can perform as many as 150 million billion floating-point operations (or FLOP) per second. By comparison, the world's highest performing computer at present, the Tianhe-2 in China, performs at "just" 55 petaflops. The first-ever supercomputer to reach 1 petaflop was built as recently as 2008. So Summit and Sierra are significant upgrades.
Plans are for Sierra to be the exclusive domain of The National Nuclear Security Administration, which will use its new super-toy for ensuring "the safety, security and effectiveness of the nation’s nuclear deterrent without testing," according to a press release by Nvidia, one of the manufacturers of the new computers' components. Summit, however, will be made available to researchers worldwide who can apply for time with the powerful technology.
This will be a huge boon to science and technology, as researchers can use Summit to model everything from climate change and weather behavior to materials science and nuclear-weapons performance. The faster the computer, the more detailed models that can be created. Natural systems, such as Earth's climate, are extremely difficult to model because of their immense complexity.
"There is a real importance of having the larger systems, and not just to do the same problems over and over again in greater detail," said Julia White, manager of a grant program that awards supercomputing time at the ORNL and Argonne National Laboratory in Illinois. "You can actually take science to the next level."
Related on MNN: