NVIDIA’s 18K+ Tesla K20 GPUs claims ‘world’s fastest open-science supercomputer’
NVIDA has just announced the completion of ‘the world’s fastest open-science supercomputer’ at the Oak Ridge National Laboratory in Tennessee this month.
Dubbed as the ‘Titan,’ this supercomputer will allow researchers to compute and explore physical and biological phenomena at a peak performance of more than 20 petaflops (20 billion floating-point operations per second), with 90 percent of this brain power coming from 18,688 K20 GPUs.
The Tesla K20-powered supercomputer, NVIDIA claims, is 10 times faster and 5 times more energy efficient than the previous x86-based AMD Opteron 2.3 petaflops Jaguar system.
“The new Tesla GPU accelerators offer the performance and energy efficiency that enable Titan to scale to unprecedented performance levels without consuming the energy equivalent of a small city,” said Jeff Nichols, associate laboratory director for computing and computational sciences at Oak Ridge National Laboratory.
Oak Ridge’s new system now boasts more than 18,000 K20 GPUs, 710 terabytes of memory, and a conversion of the system’s 200 cabinets to a Cray XK7 supercomputer.
NVIDIA hopes that Titan will usher in a new era of supercomputing in which the 1,000 petaflops mark will become a reality within the next few years.
“You simply can’t get these levels of performance, power-and cost-efficiency with conventional CPU architectures. Accelerated computing is the best and most realistic approach to enable exascale performance levels within the next decade,” said Steve Scott, chief technology officer of NVIDIA’s GPU Accelerated Computing.