Specifically, up to 18,000 of NVIDIA’s

Clearly enjoying the fruits of its labour in the wider chip-making market, one-time graphics specialist NVIDIA is to have its GPU processing technology utilized in Oak Ridge National Laboratory’s upcoming supercomputer.

Specifically, up to 18,000 of NVIDIA’s Tesla processors are to be adopted during the second phase development of the Titan supercomputer (Cray XK6), which is being built to churn through a variety of important scientific research projects—including the creation of commercially viable biofuels.

“Oak Ridge’s decision to base Titan on Tesla GPUs underscores the growing belief that GPU-based heterogeneous computing is the best approach to reach exascale computing levels within the next decade,” commented Steve Scott, chief technology officer of Tesla products at NVIDIA.

For those not immediately familiar with such heavy-handed parlance, “exascale” computing refers to performance levels in the region of 1,000 petaflops.

According to Scott, NVIDIA’s processors will provide more than 85 percent of Titan’s muscle when running at peak performance (approx. 20 petaflops), adding that it isn’t possible to achieve such levels in a power-efficient and cost-efficient way with just CPUs.

In the first phase of Titan deployment, which is currently underway, Oak Ridge will upgrade its existing Jaguar supercomputer with 960 Tesla GPUs based on the NVIDIA Fermi architecture. These GPUs will serve as companion processors to multi-core CPUs in the Cray XK6 supercomputer. 

If the Oak Ridge National Laboratory is able to make good on Titan’s considerable promise, it will emerge boasting twice the performance of Japan’s K supercomputer, which is currently the world’s fastest chunk of supercomputing hardware. It would also be three time more energy efficient than the K.