10-20 Petaflops in 2012?
The “Fastest Supercomputer” title may move 6,940 miles (11,167km) eastward in 2012 from Kobe, Japan to a small Tennessee town.
That’s if the folks at the Oak Ridge National Labs (ORNL), along with Cray and AMD, can pull off a massive upgrade of the existing Jaguar system. They’ll replace existing nodes with the new Cray XK6 nodes, which will run the new 16-core AMD Interlagos chips.
In the second half of 2012, these will be augmented by dual NVIDIA Kepler GPUs. Interlagos CPUs should be around 3x faster than their current hex core processors and the addition of Kepler GPUs (and lots of ‘em) will really crank up the performance potential.
Overall, they’re expecting a 9x increase in speed, which will put the system in the 10-20 PF range when it’s completed near the end of 2012.
This new system will carry the moniker Titan, which could allude to the largest moon of Saturn or to the Titans, who in Greek mythology were a race of gods who appear in a lot of old poems and probably dismembered and ate Dionysus. Either way, the system will be a screamer.
Jaguar is hardly a slouch in the performance department now. It was the fastest system in the land back in 2009 and currently holds down third place on the Top500 list, delivering 1.759 PFs from 224,162 cores.
If it can hit the midrange of their 10-20 PF performance estimate, it has a fighting chance to unseat the Fujitsu K supercomputer at the top of the list. However, K is still growing, and assumedly can also use the Kepler GPUs when they become available, so it’s not assured that Titan can top K.
Our pals at InsideHPC have a link to the ORNL slides. You’ll see a good discussion of the hurdles facing system scalability along with details about the new Titan box. ®
I don't visit slashdot any more so
"Just imagine a beowulf cluster of these things!"
"...Kind of a moot point however, as the sort of computations run on these things are often embarrassingly parallel or, at the very least, amenable to parallelization..."
Yes, and because Linux is very good at horizontal scaling (scales good in a cluster) we see that Linux is often used. Workloads that suits clusters is what Linux is good at.
AMD CPUs with nVidia GPUs?
Vote of "No Confidence" in what's left of ATI, then?