AMD Opteron and Nvidia Tesla to shatter supercomputer records

Cray is planning to break all supercomputer speed records by using a cocktail of AMD Opteron processors and Tesla graphics processors from Nvidia.

The plan is to create a supercomputer, dubbed Titan, with a peak speed between 10 and 20 petaflops for the US Dept. of Energy’s Oak Ridge National Laboratories.

The computer will be built by early 2013 and will manage a peak performance of 10 and 20 peteflops which is 10 and 20 million billion calculations per second.

Currently the record for supercomputer speed is held by the K Computer at the Riken Advanced Institute for Computational Science in Kobe, Japan, which was clocked at more than eight  petaflops.

The Titan is an upgrade to Oak Ridge Labs’ Jaguar supercomputer, the Cray XT5 system and can manage 2.3 petaflops. In addition to speed Titan Cray XK6 system will offer advanced capabilities in modelling and simulation.

According to a press release,  Titan will have 299,008 cores and 600 terabytes of memory. It will be used to run complex number juggling for the Department of Energy.  This will include assisting in the commercially viable production of biofuels and biomaterials.

Simulations will also aid the development of new materials for photocells that will convert more sunlight into electricity and new battery technology to store that energy for use when the sun isn’t shining. Titan’s users will also study safe extension of nuclear power plants, according to Oak Ridge.

The initial system upgrade will replace all of the Jaguar processors with the latest AMD Interlagos Opteron processors. This will take the system from two six-core processors per node to one 16-core processor per node. It will mean that it will be possible to install graphics processing units.

The Tesla GPUs can perform many more calculations for the same amount of power as conventional microprocessors and will work hand-in-hand with CPUs to deliver new levels of energy-efficient application acceleration, the statement said.

There will be between 7,000 and 18,000 of Nvidia’s next-generation Tesla GPUs will be added to the system, next year increasing the peak performance to between 10 and 20 petaflops.