Supercomputer to Hit 1.6 Petaflops With 16,000 Cell Chips

tygerstripes writes, "IBM has announced that they are gearing up to build the world's fastest supercomputer, more than four times faster than the reigning champ, IBM's BlueGene/L. Nicknamed 'Roadrunner,' the new machine will be a hybrid of off-the-shelf CPUs and Cell chips designed for the PS3. Roadrunner is to be installed at Los Alamos National Laboratory, occupying 1,100 square metres of floorspace (that's a square about 110 feet on a side). According to the BBC: 'The computer will contain 16,000 standard processors working alongside 16,000 Cell processors... each Cell is capable of 256 billion calculations per second.'"

So the price was

[CrazyJim1]

16,000 *600$= 9.6 million. That doesn't seem like much for the biggest super computer.

Re:Feet/Metres/Meters

[jelle]

"Why mix the units like that? It's either 33 meters a side, or its 12,100 square feet. Mixing units is the sort of thing that can only lead to errors."

'Non-metric people' are used to units being mixed up... The solution taken by many is to either give up and think that 'math is difficult', or to only use rounding/approximations for 'quick calculations': '5000 feet per mile' (instead of 5280)...

Rounding like that is what results in allowing space for the 4 interns... Hurrah for rounding: Where would the interns sit otherwise? ;-)

For example, how many ounces does a cubic foot of water weigh, and how many gallons is that by the way? It's not trivial, you'd have to memorize or calculate it from other numbers you memorized. So a cubic foot of water weighs 62.31 pounds, or 996.96 ounces, which is 7.78875 gallons of fluid, oh no 7.48051945 gallons because there are two meanings of 'ounce'.

Allways there is either an apparently random number to memorize, or maybe one that is easier to memorize if you give up on being exact and round it, and there is more confusion wherever possible. In the metric world, it's much simpler: A cubic meter of water weighs a (metric) tonne (1000 Kg), and is 1000 Liters. You still have to memorize relationships, but the numbers are nicely rounded. It's easier because all metrics more easily related.

Re:But is it Vista Ready

[Anonymous Coward]

I'll take part of the bait - anonymously anyway...

Compiz [wikipedia.org] - nothing else comes close to virtual desktop in 3D, Video demonstration of Compiz on Xgl [freedesktop.org] (linked at the bottom of wikipedia page direct to video)

I've been using Windows since 95 on a P1 all through 98, NT4, ME, 2000, XP, and my current XP2003 with 200 and 300 gig hdd's, Athlon 64 X2 4200, XFX GeForce 7900 GT, 4 gigs of ram, and twin/dual display 17" lcd's, - and Linux is great...

BTW: did I mention I also put Linux on my laptop (an Athlon 3200 with a gig of ram). Not everybody that runs linux cheaps out on hardware - in fact, more windows users cheap out on machines than linux - what can you say about Dell's running XP? not shit. On the other side, what do you have that can compete with supercomputers such as the one this article is supposed to be about - IBM's next 1.6 petaflop supercomputer - which will run what? LINUX!!!

Re:Hmmm

[suggsjc]

Forgive the oversimplification...
GHz is a measurement of how fast the clock cycle of a processor is. This system will have lots of processors, that will contibute to the computing power (# of flops [floating point operations]) of the overall system. So, GHz isn't a good measurement. However, I'll try to give you a meaningless comparison.

From wikipedia:

A relatively cheap but modern desktop computer using, for example, a Pentium 4 or Athlon 64 CPU, typically runs at a clock frequency in excess of 2 GHz and provides computational performance in the range of a few GFLOPS.

A GFLOP is 10^9 FLOPS or 1,000,000,000 FLOPS
A petaFLOP is 10^15 FLOPS or 1,000,000,000,000,000 FLOPS

So for the comparison purposes (assuming the few GLOPS from above is 5) this would be like a standard desktop running at ~4,000,000 GHz

I hope I did that math correctly...

Re:Flops? CPS?

[adam31]

The flop, of course, came from floating point operation. Even then it's vague--is it single, double or double-extended?

I was thinking the same thing. Running the numbers, 256 GFlop * 16,000 => 4.096 PFlop @ single precision. So if IBM means SP flops, something is slowing its theoretical max down by 2.5x. But Cell's DP perf yields 18.2 * 16,000 => .292 PFlop @ DP. So that's not it either.

It's long been rumored that a post-PS3 Cell is in development that can pipeline DP flops. Its max theoretical DP perf would still be half of SP because it's just 2 DP values per 128-bit register instead of 4. AND, if you figure they lower the GHz to 3.2 to cut the heat output in half, you arrive at the magical number... 1.638 PFlop.

So can we take this as evidence that there now exists a Cell that performs DP calculations pipelined?

Re:PS3 delayed?

[Anonymous Coward]

Or comon news articles [google.com], take your pick.

No, definite ass-pulling is occuring. The somewhat substantiated rumors of low yields don't mean there's a supply problem. You can increase wafer starts if the low yields are a surprise, or if it's not really a surprise, you've already planned accordingly. After all the hoo-ha resultingly from an IBM VP mentioning low yields, saner minds did point out that the chip is frigging gigantic, after all.