Oak Ridge National Laboratory (ORNL) has received a handful of the GPUs that will power their upcoming Titan supercomputer. Jeff Nichols, the lab’s scientific computing chief, confirmed that 32 Kepler processors from NVIDIA have been installed in the “development platform” when speaking with the Knoxville News Sentinel last week.
ORNL expects to receive roughly 1,000 more Tesla K20’s this week, enabling the next phase of the upgrade process to begin in earnest. But that’s still less than 10 percent of the total number of GPUs that will be present in the final system.
When all is said and done, the Titan supercomputer will perform at an estimated 20 peak petaflops. This will be achieved with 18,688 nodes running the latest 16-core AMD CPUs and 14,592 K20 GPUs. Each node will have 32 GB of memory, supplying 2GB per CPU core.
The upcoming system is expected to provide some healthy competition to the Sequoia supercomputer at Lawrence Livermore National Laboratory. Sequoia is currently ranked first in the world in total performance (16.32 Linpack petaflops) and energy efficiency (2,100 megaflops/watt). Titan has a shot at topping Sequoia in both categories.
The official target date for Titan to be available to end users is March 2013. “We will change the name (from Jaguar to Titan) once we have gotten through acceptance (most likely sometime between December 2012 and March, 2013),” said Nichols. Given that timeframe, it’s still up in the air whether Titan will appear on November’s TOP500 rankings.