Oct
29
2012

US Has The King of Supercomputers again!

Great to hear that we have the new fastest supercomputer, and we plan on keeping it that way. Titan, the U.S. Department of Energy’s top open science computer, is going live on Monday with an upgrade that will likely make it the fastest supercomputer on the planet. At 20 petaflops — that’s 20 quadrillion calculations each second — Titan outperforms by four petaflops the DOE’s Sequoia supercomputer, which has held the crown since June. What is really crazy is what this guy is made of, gaming computer GPU’s! That’s right, all us gaming dorks are helping the DOE get the fastest supercomputer on the planet.

Titan replaced its predecessor’s 224,256 central processing units (CPUs) with 299,008 faster CPUs made by AMD along with 18,688 graphics processing units (GPUs) made by Nvidia. The GPUs serve as accelerators to the CPUs. That’s why Titan has just a third more central processors and the same number of computing nodes and cabinets as Jaguar, but delivers 10 times the performance. In other words the GPU’s give them a boost, and with an energy savings too! If they had just added plain CPU’s, a 20 petaflop machine would have required 60 megawatts of power, at a cost of $60 million. That would have been a dealbreaker.

This global race to have the fastest computer at all times will not be ending soon. Europe, China, and Japan are running right along with us, and this is going to be a marathon instead of a sprint. Yet we plan on setting the pace. By 2016, the Department of Energy will be upgrading Titan to its successor, which it hopes will reach 200 petaflops — 10 times the speed of Titan.

“Demand will never stop,” Scott says. “Once we’re on the verge of an exaflop” — that’s 1 quintillion calculations per second — “scientists will be talking about their demand for a zettaflop.

1 Comment + Add Comment

  • I think robotics and ariiatcifl intelligence will be one of the main uses for increased numbers of processors. Computer vision systems especially can make use of multiprocessing. The ability to produce photorealistic 3D models from camera images in real time is one possible application. Such models can be used for robot navigation, scene understanding and simulation of the possible outcomes of different actions before they’re implemented (a kind of visual imagination).Computer games will be the other main beneficiary, with the ability to model very realistic physics effects and scenery (especially flexible objects with very organic shapes such as trees, vegetation and water).

Hub Page


Check us out on Thumbtack!