Built by IBM and coming in at a monstrous 4,500 square feet at the US Department of Energy’s Lawrence Livermore National Laboratory in California, the supercomputer known as ‘Sequoia’ has taken back the United States’ number one spot for the world’s fastest supercomputer from Japan and Fujitsu’s K Computer.
Running the Linux operating system, the Sequoia carries out simulations of nuclear weapons tests and has been measured at 16.32 petaflops, or 16 thousand trillion calculations per second (sufficient to say, that’s fast!), dwarfing the K Computer’s 10 petaflops, which is a staggering achievement in of itself to say the least. What’s more, with over 1.5 million processors on board, IBM boasts the machine is capable of 20 petaflops, though this is as yet unconfirmed. Amazingly, Sequoia is also much more energy-efficient than the K Computer of Fujitsu, using just 7.9 megawatts of power compared to the 12.6 megawatts needed by its rival.
To put the power of the giant supercomputer into context, IBM’s creation is capable of calculating in just one hour of performance what would “otherwise take 6.7 billion people using hand calculators 320 years to complete if they worked non-stop,” according to the BBC. Quite. Moreover, it’s 273,930 times faster than Thinking Machines CM-5/1024, the very first computer to take top spot on the list, published every 6 months, of the world’s fastest supercomputers – that was in 1993. In a calculation which would have taken the CM-5/1024 three days to compute, the Sequoia can do in less than one second.
“Supercomputers such as Sequoia have allowed the United States to have confidence in its nuclear weapons stockpile over the 20 years since nuclear testing ended in 1992,” the lab said in a statement. National Nuclear Security Adminstration administrator Thomas D’Agostino added; “While Sequoia may be the fastest, the underlying computing capabilities it provides give us increased confidence in the nation’s nuclear deterrent… Sequoia also represents continued American leadership in high performance computing.”