The Department of Energy’s Oak Ridge National Laboratory has launched Titan, a supercomputer capable of a peak output over 27 petaflops, which should put it at (or near) the top of the world’s fastest high-performance computers.
The Cray XK7 system contains 18,688 nodes, with each holding a 16-core AMD Opteron 6274 processor and an NVIDIA Tesla K20 GPU accelerator. In total, the hybrid system will contain more than 700 terabytes of memory and combine the specialized output of the graphics processing units (repurposed for optimized, repetitive calculations) and the general-purpose Opteron processors. It will take up 4,352 square feet.
“One challenge in supercomputers today is power consumption,” Jeff Nichols, associate laboratory director for computing and computational sciences for the ORNL, wrote in a statement. “Combining GPUs and CPUs in a single system requires less power than CPUs alone and is a responsible move toward lowering our carbon footprint. Titan will provide unprecedented computing power for research in energy, climate change, materials, and other disciplines to enable scientific leadership.”
Titan will replace Jaguar, the sixth-fastest supercomputer in the world, according to the most recent TOP500 list of the world’s fastest supercomputers.
The world’s fastest supercomputer, according to the June list, is Sequoia, also operated by the Department of Energy and containing 1,572,864 cores’ worth of 16-core, 1.6-GHz IBM POWER BQC chips. That system can provide a peak output of 20.132 petaflops, which should be easily topped by Titan. Until the list is released, however, it’s not entirely possible to say which system will come out on top.
ORNL will use the supercomputer to power research projects studying improvements in energy, climate change, efficient engines, materials, and other disciplines. The agency said that the “lion’s share” of access would be given to the Innovative and Novel Computational Impact on Theory and Experiment program, whose processor allocation awards include 65 billion processor hours to model the human skin barrier, 150 billion processor hours to model a weather simulation, and more. One simulation models the processes that take place inside the internal combustion engine of a car—obviously critical to the U.S. auto industry.
Titan will allow researchers to model large-molecule hydrocarbon fuels such as the gasoline surrogate isooctane; commercially important oxygenated alcohols such as ethanol and butanol; and biofuel surrogates that blend methyl butanoate, methyl decanoate, and n-heptane.
“These complexities pose challenges, but also opportunities, as the strong sensitivities to both the fuel chemistry and to the fluid flows provide multiple control options which may lead to the design of a high-efficiency, low-emission, optimally combined engine-fuel system,” Jacqueline Chen of Sandia National Laboratories, one of the team members on the project, wrote in a statement.
Titan is obviously a public relations boon for both AMD and Nvidia, as well as a potential source of conflict. Each manufacturer attempted to show that its chip was Titan’s heart and soul.
“You simply can’t get these levels of performance, power- and cost-efficiency with conventional CPU-based architectures,” Steve Scott, Nvidia’s chief technology officer, wrote in a statement. “Accelerated computing is the best and most realistic approach to enable exascale performance levels within the next decade.”
For its part, an AMD spokeswoman made sure that reporters knew that without AMD, there would be no Titan. “Let me be clear—there is no supercomputer without a CPU,” she added. “And there is no supercomputer based solely on GPU technology.”