The world’s fastest supercomputer is designed to simulate atomic bombs

The top position in the league table of the world’s most powerful computers has changed hand, designed for a supercomputer that excludes another for US national security research.

Top500The fixed list of the most competent computers is based on a metric: how fast a machine can solve a huge number of equations, measured in the floating-point operation of the flop, or flop. A machine called Frontier manufactured in 2022 was first publicly accepted, reaching Excil – one billion billion flops.

The Frontier was designed by the Oak Ridge National Laboratory in Tennessee to work on a series of complex scientific problems including climate modeling, nuclear fusion simulation, drug discovery and national security applications.

Now, the Lawrence Livormore National Laboratory (LLNL) in California has created L Capiton, capable of 1.742 exflops, which is more than any other supercomputer.

The machine has been designed in collaboration with the National Atomic Safety Administration, which is dedicated to developing nuclear weapons under tight security of the Department of Energy. The agency was formed in 2000 after the discovery that the nuclear mystery was leaked in China from the Energy Department.

L Capiton will essentially provide the huge computational power required to ensure the effectiveness of the American nuclear preventive without executing physical atomic tests. LLNL claims that the complex, high-resolution 3D simulation of nuclear explosions that will take months on Sierra, its most powerful system will be done in a few hours or days on L Capiton.

The most powerful supercomputer in the world is a title that does not change the hands as many times. Simon Macintosh-Smit At Bristol University, Britain. He is developing the UK’s Isbard-AI supercomputer, which is expected to rank in the top 10 machines next year.

Frontier is sitting at number one for two years and L Capiton is expected to maintain his position for at least longer. McIntosh-Smith says, “There used to be more turnout, more turnover, but it all slowed down.” This is because it is known as Moore’s rule, in which the computing power of a new chip doubles every two years, no longer true.

“As Moore’s law slowed down, the way you compensate for it is that you make it big and big. Therefore a system of the same size in the past can double in performance in two years, ”says McIntosh-Smit. “It’s not now.” Therefore, in two years, 50 percent should be larger and 25 percent more power should be used. So it all becomes more complex, so it has become more expensive. All this became difficult. ,