IBM To Build 3-Petaflop Supercomputer

angry tapir writes "The global race for supercomputing power continues unabated: Germany's Bavarian Academy of Science has announced that it has contracted IBM to build a supercomputer that, when completed in 2012, will be able to execute up to 3 petaflops, potentially making it the world's most powerful supercomputer. To be called SuperMUC, the computer, which will be run by the Academy's Leibniz Supercomputing Centre in Garching, Germany, will be available for European researchers to use to probe the frontiers of medicine, astrophysics and other scientific disciplines."

Comments

Re:i now play chess vs the world

[Anonymous Coward]

1: it's a dick-wagging contest to have the best supercomputer in the world.

2: huge supercomputers can be leased out to cycle-hungry organizations the same way one would lease office space in a skyscraper.

3: each incremental advancement represents overcoming various hurdles faced by all computing technology; the simple needs of common folk will become that little bit easier as a part of our constant forward march in technological advancement.

Re:3 Petaflops

[gmhowell]

So how long would that take to sort 3 petafiles?

Hi, I'm Stone Philips with Dateline NBC...

Not POWER7, Not BlueGene,

[pip-PPC]

From the article:

"The system will use 14,000 Intel Xeon processors running in IBM System x iDataPlex servers."

IBM has two in-house HPC platforms that could both reach 3 PFLOPS (BlueGene/Q and POWER7), but instead they're building a Xeon cluster. I'm surprised that they would want to put a machine near the top of the TOP500 that wasn't a full-on IBM benefit--maybe IBM Germany is the contractor, and they don't have the R&D expertise? Or the Xeon cluster is cheaper/easier to program and maintain?

Re:Why I love Moore's law

[NewtonsLaw]

So our desktop computers will wait thousands of times faster than they do today... for the next keystroke or mouse button-press. :-)

Re:Why I love Moore's law

[FeepingCreature]

Moore's law is ok. I prefer that law, forget the name right now, that says that as computational power increases, windows will require ALL of it to run, greatly increasing demand for CPU and RAM, and lowering the cost of hardware just behind the curve for the rest of us.

As Intel giveth, Microsoft taketh away.

Re:Not POWER7, Not BlueGene(BlueGene/Q)

[Required Snark]

Here is a look at the guts of the IBM next generation BlueGene/Q. http://www.theregister.co.uk/2010/11/22/ibm_blue_gene_q_super/page2.html

The Sequoia super that Lawrence Livermore will be getting in 2012 — IBM said it'd be in late 2011 back when the deal was announced in February 2009, so there's been some apparent slippage — will consist of 96 racks and will be rated at 20.13 petaflops. Argonne National Laboratory said back in August that it wanted a BlueGene/Q box, too, and it will have 48 racks of compute drawers for a total of 10 petaflops of floating-point power.

Both the Chinese machine and the German machine are not cutting edge designs. They represent what you can do with near commodity hardware and good but not fully custom packaging. They may look like top end machines today, but by 2012 they will not be in the top ten.

How should we measure supercomputers now?

[Entropius]

Once upon a time, supercomputers were bunches of general-purpose cpu's, and you made them faster by connecting up more of them.

Now people have realized that massively parallel special purpose chips (like Cell and, even more so, GPU's) can be used to do general-purpose computing, and have started to add those to clusters. But those chips have a lower bandwidth:flops ratio than the x86 etc. CPU's that have been historically used; the gap between a computer's "peak" FLOPS (on an ideal job with no communication requirements to either other nodes or to memory) and the performance it actually achieves is wider using something like CUDA than on a standard supercomputer. CUDA machines are so bandwidth-limited that people use rather hairbrained data compression schemes to move data from place to place, just because all the nodes have extra compute power lying around anyway, and the bottleneck is in communication. (The example that comes to mind is sending the coefficients of the eight generators of an SU(3) matrix rather than just sending the eighteen floats that make up the damn matrix. It's a lot of work to reassemble, relatively speaking, but it's worth it to avoid sending a few bits down the wire.)

CUDA is wonderful, and my field at least (lattice QCD) is falling over itself trying to port stuff to it. Even though it falls far short of its theoretical FLOPS, it's still a hell of a lot faster than a supercomputer made of Opterons. But we shouldn't fool ourselves into thinking that you can accurately measure computer speed now by looking at peak FLOPS. It makes the CUDA/Cell machines look better than they really are.

Re:How should we measure supercomputers now?

[afidel]

No, the computers that the term supercomputer was coined for were all special purpose vector machines that couldn't even run an OS, they had to be fronted by a management processor. Only much later were clusters of commodity machines (often with specialized interconnects for high bandwidth and low latency) accepted as contenders for the name. Now with Cell and GPU's we are getting back to fast vector machine with a management computer in the front but now the front end computer is capable of computations (at least in the case of the GPGPU machines) and each machine is a few rack units instead of a couple racks.

Oh, and the measure you are looking for are Rmax to Rpeak which will tell you how efficient the machine is (at least for LINPACK which may or may not track with your own code depending on how chatty it is in comparison to the benchmark).

Re:How should we measure supercomputers now?

[chichilalescu]

I had to really think about measuring the efficiency of a simulation and I came up with a single answer: money. I was at a lecture about gyrokinetic simulations, and when I heard about the amount of resources being used for some simulations, I asked "how much does one of these simulations cost, in euros?". Luckily for me, the guy knew (large simulations cost in order of thousands), and he also knew how much an experiment on ITER will cost (order of a million); his argument was "it's obviously efficient to run a thousand simulations and pick the most relevant set of parameters for an experiment afterwards".
I think this is the way to go with comparing supercomputers: "In order to simulate experiment X, we needed N1 euros for the developers, and N2 euros of electricity to run the code on a machine that cost N3 euros to build". If you want to be thorough, add some maintenance costs. It's a bit complicated, because developers might actually be researchers, and it's not very clear how much of their time goes into writing code... but we don't really have a better way of measuring efficiency.

Re:Why I love Moore's law

[Esvandiary]

I believe May's Law is the one you're referring to; a corollary to Moore's Law, stating that software efficiency halves every 18 months (or two years).

Re:germany

[Anonymous Coward]

Germany, FUCK YEAH!
Coming again, to save the mother fucking day yeah,
Germany, FUCK YEAH!
Federal parliamentary republic is the only way yeah,
Computations your game is through cause now you have to answer too,
Germany, FUCK YEAH!
Das Land der Dichter und Denker,
Germany, FUCK YEAH!
What you going to do when we come for you now,
it's the dream that we all share; it's the hope for tomorrow

FUCK YEAH!

BMW, FUCK YEAH!
Mercedes, FUCK YEAH!
Porsche, FUCK YEAH!
Engineering, FUCK YEAH!
Efficiency, FUCK YEAH!
Claudia Schiffer, FUCK YEAH!
Mozart, FUCK YEAH!
Bach, FUCK YEAH!
Einstein, FUCK YEAH!
Heisenberg, FUCK YEAH!
Max Planck, FUCK YEAH!
Von Braun, FUCK YEAH!
Berlin, FUCK YEAH!
Bismarck, FUCK YEAH!
German food ... (Fuck yeah, Fuck yeah)