On Friday IBM and the U.S. Department of Energy’s Oak Ridge National Laboratory (ORNL) announced the unveiling of Summit, the world’s fastest supercomputer. IBM claims that it is a world-class supercomputer with thousands of processors.
For IBM, it is a big step forward unveiled at Oak Ridge National Laboratory in Tennessee. Currently, it will have a peak performance of 200,000 trillion calculations per second, thus making it the “most powerful and smartest scientific supercomputer.”- Dave Turek, vice president of high-performance computing and cognitive systems at IBM. This performance will ensure a position at the top of the Top 500 supercomputer ranking that will be published later this month. Since 2012, this will be the first time that a U.S.-based supercomputer has held the top spot on that list.
Summit features 4,608 computer servers with two 22-core IBM Power9 chips and six Nvidia Tesla V100 GPUs, each with 10 petabytes of memory. It’s of no surprise that because of the Nvidia GPUs, the system is meant to be used for machine learning and deep learning applications.
Jeff Nichols, ORNL associate laboratory director for computing and computational sciences, announced that the summit’s AI-optimized hardware “gives researchers an incredible platform for analyzing massive datasets and creating intelligent software to accelerate the pace of discovery.” It will be used for designing chemical formula, exploring new materials, studying links between cancer and genes on a very large scale, and so on.
It takes up an eighth of an acre that is the size of two tennis courts. Its peak energy consumption of about 15 megawatts can provide power to more than 7,000 homes.
IBM has been continuing to build DOE supercomputers, and Summit isn’t the end of its invention; instead, it is a step forward in reaching the later goals of one that can perform about 5 times that of Summit.
