Oak Ridge unveils Titan, the world's most powerful supercomputer
The U.S. Department of Energy's Oak Ridge National Laboratory (ORNL) has just introduced Titan, the world’s most powerful supercomputer. The size of a basketball court and using enough power to run a small town, the water-cooled circuits of Titan are capable of 20 petaflops or 20,000 trillion calculations per second. This makes Titan ten times more powerful than ORNL’s previous computer, Jaguar and 200,000 times more than the average PC. What’s more, it achieves this through components originally created for gaming computers.
Also known as the Cray XK7 system, Titan has 18,688 nodes. Each of these consists of a 16-core AMD Opteron 6274 processor and an NVIDIA Tesla K20 graphics processing unit (GPU) accelerator. It also has 700 terabytes of memory.
GPUs are designed to handle computations simultaneously at high speed and make excellent supercomputer components. The 299,008 CPU cores in Titan are used to guide the simulations, while the GPUs do the brute-force number crunching.
"One challenge in supercomputers today is power consumption," said Jeff Nichols, associate laboratory director for computing and computational sciences. "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."
The increased speed and computational power of Titan will allow for faster, more detailed simulations of a wide range of phenomenon. ORNL and Cray are still working on final system acceptance, but Titan is already working on projects, such as those from the Department of Energy's Innovative and Novel Computational Impact on Theory and Experiment program, better known as INCITE.
The INCITE projects that Titan will work on include nano-scale study of magnetic materials for new motors and generators. Another project is the creation of computer models of the turbulent combustion inside an internal combustion engine and large-molecule hydrocarbon fuels to improve fuel efficiency. It will also be used to simulate nuclear fuel rods to study extending reactor life and the development of more accurate climate modeling to not only determine if the climate is changing, but where and how.
But it's not only high-end scientific research that's set to benefit from the development of this monster computer – the GPU technology is also likely to find its way into consumer graphics cards in the future.
Update: This story has been modified in response to an error pointed out by a reader regarding the computing power of Titan compared to an average PC.