US government agency launches Titan supercomputer with thousands of GPUs

The US Department of Energy has launched a new supercomputer that is alleged to be among the fastest in the world for open science research. The machine, dubbed Titan or the Cray XK7 supercomputer, will work in conjunction with the department’s nationwide network of research labs and is operated in Tennessee by the Oak Ridge National Laboratory.

Titan, an upgrade to the Jaguar computer, uses a combination of discrete graphic chips and traditional microprocessors to boost the amount of tasks the computer can complete while reducing the amount of power it needs. Accordingly, it’s estimated to be 10 times more powerful and five times more energy-efficient than its Jaguar predecessor.

It is powered by nearly 300,000 AMD Opteron microprocessors, almost 19,000 Nvidia Tesla K20 ‘Kepler’ GPUs, and some 710TB of memory, allowing it to run at speeds of up to 20 petaflops. Combined, it has the computing power of 500,000 laptops.

“We can augment the CPUs to provide power efficiency that is far better,” said Sumit Gupta, general manager of Nvidia’s Tesla GPU business. “This is like putting 500,000 laptops in a space the size of a basketball court.”

“Basing Titan on Tesla GPUs allows Oak Ridge to run phenomenally complex applications at scale, and validates the use of accelerated computing to address our most pressing scientific problems,” said Steve Scott, chief technology officer of the GPU Accelerated Computing business at Nvidia. “You simply can’t get these levels of performance, power- and cost-efficiency with conventional CPU-based architectures. Accelerated computing is the best and most realistic approach to enable exascale performance levels within the next decade.”

Titan will be used by different groups interested in high-level research, including government agencies, academics, and global industry, to work on such complex, computing-intensive subjects as biofuels, climate change, and nuclear energy.

"Why care about these big computers? It's really because society's got big problems," Scott said in a recent interview. "There are healthcare issues, lots of diseases across the board, an aging population. Energy is a huge problem facing the world and our country. ... Increasingly computers are used to solve these problems."

But as powerful as Titan is, the work doesn’t stop here. Researchers at the Oak Ridge lab are hoping to keep improving on the system, with plans for an exascale computer that will reach a performance of 1,000 petaflops by 2020.