At the end of July, President Obama signed an executive order to set up the National Strategic Computing Initiative (NSCI) with the intention of building the world’s first exascale supercomputer by 2025. Currently, the use of supercomputers is fairly limited, but the initiative seeks to make the U.S. a world leader in their use by expanding the applications of high power computing (HPC) and generate demand for HPC software and hardware developers.
Various programs are already underway that involve processing huge amounts of data for scientific and defense research projects, for instance the Large Hadron Collider which gathers data on high-speed collisions of subatomic particles and The Human Brain Project that uses computers to mimic the workings of the brain. They’re also used by aerospace engineers to model planes and weapons, and by climatologists to forecast the short-term impact of hurricanes and analyze the long-term consequences of climate change. Having more powerful supercomputers in use would help stimulate these kinds of initiatives, as well as generate yet un-thought of projects. Industry could also make extensive use of HPC, such as enabling the testing of automobile designs or the fluid dynamics of new aircraft. If more testing were possible in a virtual setting, it would lower costs and hasten product development.
The proposed exascale computer would be able to complete one quintillion floating point operations per second (1018 FLOPS, or one exaflop), making it 30 times more powerful than China’s Tianhe-2, the world’s current leader. Global data consumption is already measured in exabytes. In fact, global Internet traffic is expected to surpass one zettabyte (1021 bytes) in 2016. That kind of scale necessitates a new kind of supercomputer, capable of analysing such a vast quantity of data. The U.S. want to be at the head of that race, instead of lagging behind of China as they are currently doing.
The new supercomputer will have to be newly designed for purpose, and will require so much energy to operate that it will need its own exclusive power plant. The Department of Energy, the Department of Defense and the National Science Foundation will jointly fund the initiative. Various government agencies, most notably the Department of Energy have been heavily involved in the development of supercomputers over the last several decades, but they’ve largely worked separately . The primary goal of the initiative is to get scientists and engineers from different agencies working together on shared goals.
After an exascale computer has been built, designing programs that can take advantage of its power will be the next technological hurdle. The NSCI has explicitly stated, however, that its goal is to “make HPC resources more readily available so that scientific researchers in both the public and private sectors have ready access”.