Looks like Watson has a new, younger, but somewhat brawnier, family member.
On June 8 the U.S. Department of Energy’s Oak Ridge National Laboratory unveiled IBM’s Summit supercomputer and immediately billed it as the “world’s most powerful and smartest scientific supercomputer.”
As impressive as it certainly is, IBM and the DoE might get a legitimate argument about that lofty claim from other computers in the Top 500 around the world—namely in Guangzhou, China; Cern, Switzerland; and Japan. The Titan supercomputer in Oak Ridge, Tenn. was listed as No. 4 in the last ranking. We’ll have to see where Summit eventually ranks on the list, and it may well become No. 1.
Secretary of Energy Rick Perry attended the debut in Oak Ridge June 8 to meet with the ORNL team and see firsthand this monumental supercomputer that has 4,600 individual nodes—large enough to fill two tennis courts with racks of servers.
200 Quadrillion Calculations Per Second
With a peak performance of 200,000 trillion calculations per second—that’s an astounding 200 quadrillion calculations, or 200 petaflops—Summit will be eight times more powerful than America’s current top-ranked system, Titan, also housed at ORNL.
For specific scientific applications, Summit will be capable of more than 3 billion-billion mixed-precision calculations per second. Summit, IBM said, will provide unprecedented computing power for research in energy, advanced materials, and artificial intelligence (AI), among other domains. Summit’s power is expected to enable scientific discoveries that were previously impractical or impossible.
“Summit is also optimized for AI in a data-intense world,” IBM Senior Vice President of Portfolio and Research Dr. John F. Kelly III said in a media advisory. “We designed a whole new heterogeneous architecture that integrates the robust data analysis of powerful IBM Power CPUs with the deep-learning capabilities of GPUs. The result is unparalleled performance on critical new applications.”
“This project has always been about pushing the boundaries of innovation and technology to solve what was previously unsolvable. For instance, with this system we can make connections and predictions that will help us advance cancer research, understand genetic factors that contribute to opioid addiction, simulate atomic interactions to develop stronger, more energy efficient materials, and better understand supernovas to explore the origins of the universe.”
Can Analyze 30 Years’ Worth of Data in an Hour
Summit’s computing capacity is so powerful that it has the ability to compute 30 years’ worth of data saved on a desktop computer in just one hour, Kelly said. ORNL researchers have also figured out how to harness the power and intelligence of Summit’s state-of-art architecture to successfully run the world’s first exascale scientific calculation, or exaops, as DOE’s fleet of proposed exascale computing systems come online in the next five years.
From its start 75 years ago, ORNL has a history and culture of solving large and difficult problems with national scope and impact, ORNL Director Thomas Zacharia said.
“ORNL scientists were among the scientific teams that achieved the first gigaflops calculations in 1988, the first teraflops calculations in 1998, the first petaflops calculations in 2008, and now the first exaops calculations in 2018,” Zacharia said. “The pioneering research of ORNL scientists and engineers has played a pivotal role in our nation’s history and continues to shape our future.”
“For the first time, we (IBM) are making the same architecture that powers Summit available in commercial form,” IBM’s Kelly said. “Clients are already using the same hybrid architecture in our business product line with the IBM Power Systems AC922 system, and the family of new IBM POWER9-based servers. The result: business computing that can help every industry advance their products and services, from banking, to healthcare, to retail, to transportation.”
Summit will be open to select projects this year while ORNL and IBM work through the acceptance process for the machine. In 2019, the bulk of access to the IBM system will go to research teams selected through DOE’s Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program.
