In a statement, Advanced Micro Devices Inc and Hewlett Packard Enterprise Co stated, they had won a $600 million deal to deliver a supercomputer that will be used by the U.S. Department of Energy’s nuclear security arm to support the nation’s nuclear arsenal.
The supercomputer, dubbed El Capitan, named after the famous rock face in California’s Sierra Nevada mountains, will be housed at the Lawrence Livermore National Laboratory. It will also be used by two other national nuclear labs, Los Alamos National Laboratory and Sandia National Laboratories.
Winning the chip contract is a big win for AMD, which has now deals for two of the three most powerful computers under development by the U.S. government.
According to supercomputing research group TOP500, all but one of the world’s current top 10 fastest supercomputers use central processor chips from either International Business Machines Inc (IBM) or Intel Corp. The world’s third fast super computer, located in China, uses a domestically developed chip.
Following the end of live detonation tests, using computers to model nuclear blasts has become a core part of weapons design. The enormous computing power required has also become a source of international competition and geopolitical tension.
“That is why it is so important to push the edge” of computing technology, said Bill Goldstein, director of the Lawrence Livermore National Laboratory, at an event in San Jose, California announcing the news. “To make sure we’re there.”
Washington has barred U.S. firms from doing business with select Chinese supercomputer groups following U.S. assessements of them having ties with the Chinese military.
According to Pete Ungaro, senior vice president and general manager of high-performance computing and mission-critical solutions at HPE, El Capitan is likely to be the size of two basketball courts and weigh as much as 35 school buses when complete; it will also be nearly 30% faster than its predecessor thanks to AMD’s forthcoming “Genoa” central processor units (CPUs).
AMD’s graphical processing units (GPUs) will also feature in El Capitan to speed up artificial intelligence tasks, said TOP500.
According to analysts, AMD makes both chips and a system for them to share computer memory, which is more difficult for systems that use chips from separate providers like Intel and Nvida Corp.
“AMD is the first to market with them both in same package,” said Addison Snell, chief executive of Intersect360 Research.
According to HPE, the likely deliver date of El Capitan is in 2023.
If everything goes as per plan, El Capitan will have a speed of 2 exaflops, and would be faster than the current top 200 supercomputers in the world combined.