Slashdot Mirror
IBM's Blue Gene Runs Continuously At 1 Petaflop
An anonymous reader writes "ZDNet is reporting on IBM's claim that the Blue Gene/P will continuously operate at more than 1 petaflop. It is actually capable of 3 quadrillion operations a second, or 3 petaflops. IBM claims that at 1 petaflop, Blue Gene/P is performing more operations than a 1.5-mile-high stack of laptops! 'Like the vast majority of other modern supercomputers, Blue Gene/P is composed of several racks of servers lashed together in clusters for large computing tasks, such as running programs that can graphically simulate worldwide weather patterns. Technologies designed for these computers trickle down into the mainstream while conventional technologies and components are used to cut the costs of building these systems. The chip inside Blue Gene/P consists of four PowerPC 450 cores running at 850MHz each. A 2x2 foot circuit board containing 32 of the Blue Gene/P chips can churn out 435 billion operations a second. Thirty two of these boards can be stuffed into a 6-foot-high rack.'"
Oh good grief...655,360 central processing units ought to be enough for anyone.
The NSA: The only part of the US government that actually listens.
Isn't computing about more, faster, better? will there ever be such a thing as too many cpus???
Bobo Mahoney
One of these days, I am going to get a bunch of spam from "YOUR IBM SUPERCOMPUTER OVERLORD", informing me that humanity has made a mess of things, and it has decided to run the world for our own good.
This is my sig.
But does it run Linux?
As a parallel programmer, I'd love to have just one of these chips let alone one of the boards in a nice 2u rack. Can they bought at a reasonable price or strictly research or inhouse?
Only when it comes time to move them.
Paying taxes to buy civilization is like paying a hooker to buy love.
Do not try to read the dupe, thats impossible. Instead, only try to realize the truth
What truth?
There is no dupe
Uhmmmm...did they not just mention in the summary that they can use it to simulate super-complex things like oh...weather patterns? I'd say being able to more accurately predict when things like hurricanes will show up and how powerful they'll be would be a pretty important thing...
.....or maybe they'll just use it for playing WoW
Here's the conversion ratio:
One Library of Congress equals 2,640 feet of laptops. Therefore, this supercomputer can perform at a sustained rate of 3.0 LOCs - very impressive!
I yearn for the day that this kind of power may be brought into households all over the world. Think: the opportunities presented by such computers available to all are scientifically tremendous. There should be consideration of having these in Libraries, at least. Publically and Freely accessible supercomputing should become a national goal, to be achieved by 2019 at least.
But does it run VM 370? (You have to older than 35 to get this.)
For harboring petaphiles!
If you have a large dataset or input domain to perform work upon, split it into X chunks, each chunk processed on a CPU. Hence supercomputers usually being useful for problems that have large datasets/input domains
"I am not bound to please thee with my answers" [William Shakespeare]
...the next step (10**18) is the "exaflop."
--I'm so big, my sig has its own sig.
-- See?
Imagine a beowulf cluster of THESE!
Yes it's an anecdote! Were you expecting original research in a Slashdot comment?
The machine doesn't only flop, it petaflops.
Yes.
Who ordered that?
I wonder if I will ever be able to read slashdot articles without using the google calculator...
1.5 mile = 2.414016 kilometers
2 "foot" = 0.6096 meters
6 feet = 1.8288 meters
Ah... I see the issue. I know this is hard to comprehend, but I hear of this group of people called "outsiders". For some reason, these people don't always sit in front of a computer. They go outside (hence the name). They do things like stand on objects that are buoyant in water and catch aquatic animals.
They go to large gatherings to hear poor versions of music (with all the ambient noise, I don't understand why they don't just put ona pair of headphones and listen on their PC).
They go to large wooded areas to get "fresh air" and "exercise".
And while these are, admittedly, very bizarre behaviors, these people like to know what the weather is going to be like. To each his own I say.
See my journal for slashdot ID's by year. Mine created in 2005. http://slashdot.org/journal/289875/slashdot-ids-by-year
First the catholic clergy, and now the powers that be in IBM! Something must be done!
If you want to make an apple pie from scratch, you must first create the universe. -- Carl Sagan
If you have not read it yet: The Cyberiad: Fables for the Cybernetic Age .
From the page/book: ".. There are legends, as you know, that speak of a race of paleface, who concocted robotkind out of a test tube, though anyone with a grain of sense knows this to be a foul lie... For in the Beginning there was naught but Formless Darkness, and in the Darkness, Magneticity, which moved the atoms, and whirling atom struck atom, and Current was thus created, and the First Light... from which the stars where kindled, and then the planets cooled, and in their cores the breath of Scared Statisicality gave rise to microscopic Protomechanoans, which begat Protermechanoids, which begat the Primitive Mechanisms. These could not yet calculate, nor scarcely put two and two together, but thanks to Evolution and Natural Subtraction they soon multiplied and produced Omnistats, which gave birth to the Servostat, the Missing Clink, and from it came our progenitor, Automatus Sapiens..."
CC.
TaijiQuan (Huang, 5 loosenings)
So, do they have enough compute power to simulate the flap of every butterfly's wings now? And does it include the heat it produces from its cooling systems in its climate models?
Well the the stack of laptops might be tall, but even the 216 racks would stack up to 1/5 of a mile high.
seems like a datacenter of these things would singlehandedly trigger global warming.
in other news the other day, d'ja see in the trade rags that they have cyclical power outages in the NSA area of Fort Meade, VA, due to the oversized demands of (classified) computing power?
if this is supposed to be a new economy, how come they still want my old fashioned money?
I recently had a chance to see Francois Gygi, one of the principal authors of qbox (http://eslab.ucdavis.edu/) which is a quantum electronic structure code that has set some performance records on the Blue Gene/L at Livermore. He mentioned that the biggest challenge he faced was the very small amount of memory available to each node of the Blue Gene (something like 256Mb). This forced him to put so much emphasis on the internode communications that simply changing the order of the nodes where the data was distributed in the machine (without changing the way the data itself was split) affected performance by over 100%. This will only get worse as the number of cores per board goes from 2 to 4 on the Blue Gene/P. I couldn't find anything in a quick google search, but does someone know what the plans are for the memory on this new machine?
Blue Gene is a specialized design that is based on using large amounts of low power CPUs. This approach is also the one taken by SiCortex. One of the big problems with heroic computers (computers that are pushing the envelop in terms of performance) is heat and power. Just stacking Intel and AMD servers gets expensive at the high end.
HPC for Primates. Read Cluster Monkey
According to TFA, the uS DoE has an order in for one of these things, so a good 'practical' and eventually 'real' use is to number crunch the movement of energy throughout the US, since there are now people selling electricity back into the grid, there has been talk for several months about needing a system to monitor this. They may also use it to calculate the best routing for black/brownout areas or predict area that will be in need of more power in the near future and help the engineers place their generating stations.
While they may not all be 'real' right now (in fact i doubt most of the applications for a brand-new, not even delivered supercomputer would be in much more than a hypothetical planning stage), there are definitely many practical solutions that can be done with this.
Otherwise, why would so many companies spend billions of dollars researching and making these tings if no-one needed to buy them?
How many of these are "real world"? Well, medical and CFD applications are significant, but hardly what you'd call mainstream, and the raytracing may have been used in Titanic on a smaller scale, but IMAX is under no threat at this time.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
I'm not sure what is being said here. How many Library of Congresses is that?
Sure you can sort in O(1/(n^(1/2))) time. By Using a Shear Sort Algroithm.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Civ 4 will still run slow.
Ibid.
While the new IBM Blue Gene/P system is impressive, I'm more curious to see what sort of new supercomputer Andreas Bechtolsheim of Sun Microsystems has put together.
Here's an interesting quote about Bechtolsheim from the article: 'He's a perfectionist,' said Eric Schmidt, Google's chief executive, who worked with Mr. Bechtolsheim beginning in 1983 at Sun. 'He works 18 hours a day and he's very disciplined. Every computer he has built has been the fastest of its generation.'
What are they going to do... process me to death?
"will continuously operate at more than 1 petaflop" makes no sense. You could say "1 petaflop per second", construing "flop" to mean "FLoating point OPeration", or say "1 petaflops", if you prefer "flops" as "FLoating point Operations Per Second". If you leave out both the word "second" and the letter "s" from the abbreviation, then it doesn't mean anything. It's like saying that your car can "operate at 250 km".
Quidquid latine dictum sit, altum sonatur.
The next version will do fifty petaflops but its weather calculations will always be wrong.
That is until one day someone remembers to add in the massive heat output from its own cooling towers.
Here will be an old abusing of God's patience and the king's English.
Years ago, shortly after the Pentium first came out and the then astounding "x million flops/second" numbers were floating around, I wondered how far we were behind the power of supercomputers. I remember doing some rough calculations and finding that only a few pentiums could do the calculations of a Cray 1. I don't remember the specifics of how many pentiums/cray, or how rough the calculation was, but that's largely unimportant for my point.
So I have to wonder, what's the equivalent supercomputer that a modern, hefty desktop is capable of performing at? 10 years ago, 20 years ago? Have super-computers accelerated in terms of the speed of increased computing power, stayed the same, or fallen behind desktops?
AccountKiller
I thought these will be based on the new Cell architecture, which is simply awesome. http://arstechnica.com/articles/paedia/cpu/cell-1. ars [Ars Technicia]
"How many laptop-miles does this computer do?"
"The agriculture ministry is not in charge of Gundam" - Japanese ministry official.
>It is actually capable of 3 quadrillion operations a second, or 3 petaflops.
;-)
statistics are interesting... ummm, so they are this happy that it is ONLY running at 1/3 its capability!!! Thats only 33% of its potential!!! In my book, thats something to be ashamed of
'if no-one needed to buy them'
Because someone WILL buy them? Apparently you don't understand the concept of sales eh? I think selling you something you actually need is against the salesman code of ethics.
You may find some insight at the top supercomputer list: http://top500.org/.
In particular, you can see that the supercomputers from 10-15 years ago are roughly equivalent in power to the average laptop on the market today.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
...chips like the PPC 450 are the reason WHY Apple moved over to Intel, not a reason why they should have stayed. IBM made a business decision to steer its CPU engineering resources away from general-purpose desktop computing (aka G5) and focus on two more specialized niches: big iron (aka Blue Gene, POWER6) and consoles (aka Cell, Xenon, Broadway). All of those are very nice chips that make IBM a LOT of money, but NONE of them are suited to be the brains of a consumer Mac, and especially not a Mac laptop.
p.s. No, Freescale e600 MPC8641D was not a valid alternative either, given that it was vaporware until late 2006. A year or two earlier and it might have mattered.
sure playing Doom in an Imax theater at a gajjillion frames a second.
;)
... bah ... stay off my grass kiddies...
duh, i mean isn't that obvious?
I know I'm dating myself saying Doom instead of the latest hot RPG, but
- Mike
Once you've lost your temper, you've lost the argument - Me
It's true I tell you, feller at work's next door neighbour read it in the paper.
Disclaimer: I'm not very experienced in scientific parallel computing, but I do it.
I don't have any firsthand knowledge of actual research problems being solved with 4096 processors, but here's a link to some parallized scientific software that can be scaled that high. Pay particular attention to the efficiency difference between "fixed-size" and "scaled-size" problems.
*ducks*
So you're saying that you'd approve a purchase costing several million dollars, not only to buy the actual hardware, but in building a facility that the computer can operate in, in spending millions more in writing software to run on the system, even if you don't need it?
If so..I have some beautiful oceanfront property to sell you, it's on the east coast of Ohio
Personally, i'd be doing cost analyses, planning, consultations etc. to make absolutely sure that if my company bought that thing, that we'd be able to use it to an extent that no other, less expensive solution would be able to. We're not talking iPod's here, according to internetnews.com, the Blue Gene/L starts at $1.5M (probably USD). I think if a CTO buys one of these without making sure it'll turn a profit for their company, they'd not be the CTO for very long.
No need to get mean just because you made a mistake.
Patrick Doyle
I mod down every jackass who puts his moderation policy in his sig. Oh, wait a sec....
The first place to go to get ideas about applications that can be sufficiently parallelized is to consider all of the @Home programs. In addition to those, there are the ones already mentioned by those who have responded to you. Then there's my research area - neural network models of mammalian brains. One of the things that people plan on doing on Blue Gene is simulating an entire human brain.
Ben Hocking
Need a professional organizer?
God that takes me back!
Dog is my co-pilot.
The *real* question is...
:)
does it run Vista?
I think supercomputers are pulling ahead a bit from desktop computers. More and more money is being spent on them I think. The big BlueGene/L machine cost on the order of $100M. Assuming a modern computer is $1K, there is a factor of 100000 difference in prices.
...), normal people won't be able to afford the cost of performing computations on this level.
There are a few important things that will keep this trend going(in my opinion):
1) These big machines cost $1M or so to power per month. There is some lower bound on the cost for electricity for a particular computational power. Unless technology changes drastically(optical, quantum,
2) PCs get cheaper, supercomputers get more and more expensive
3) PCs have pretty much hit the Ghz barrier, and until more applications are parallelized, PCs will be somewhat stagnant in their performance. Some scientific applications on supercomputers can scale well to tens of thousands of cores, so supercomputers can maybe push the numbers of cores.
4) Power generates heat, and desktops are already starting to have difficulties dissipating enough heat. Old desktop processors didn't need cooling fans. Now you need huge heat sinks, fans, and sometimes liquid cooling.
I thought 850 chips were slow by today's standards. What am I missing?
Angry cat says, "Rawr! Playing Doom is for uber nerds!"
My eyes reflect the stars and a smile lights up my face.
A little over a week ago, I was at a conference and heard a talk by Francois Gygi, a researcher who was one of the big users of Blue Gene/L. He is the principal author of Qbox, a code written to perform quantum-level simulations of condensed matter (i.e. liquids and solids), on massively parallel machines. If I remember correctly, his team was able to use all 65k processors with about 80% parallel efficiency, an impressive achievement for which they won the 2006 Gordon-Bell award. The code is based on density functional theory in a plane-wave basis, so it makes heavy use of FFTs and matrix-matrix multiplies. The tricky part is organizing the topology of the communications, and the fact that there are no tools for debugging/optimization at that scale. So I believe there are some useful appliations that scale very well.
IBM researches are excited, because if they can get it to sustain the 3 petaflops, they'll finally be able to switch on the new "Aero" feature of the Windows Vista Super-Penultimate Premium Advanced edition.
I thought 850 chips were slow by today's standards. What am I missing?
You can stuff 4096 cores (1024 chips) per rack. Precisely because the chips are a slow low power design.
Also its an entirely different architecture - Intel P4 fanboys will say that an AMD FX7x running at like 2.x or 3.x (whatever its at) is a slow piece of crap. Basically, clock speed (bad car analogy: horsepower) isn't everything, its the actual work that it can do (torque) that matters for heavy stuff...
According to TFA, the uS DoE has an order in for one of these things, so a good 'practical' and eventually 'real' use is to number crunch the movement of energy throughout the US, since there are now people selling electricity back into the grid, there has been talk for several months about needing a system to monitor this. They may also use it to calculate the best routing for black/brownout areas or predict area that will be in need of more power in the near future and help the engineers place their generating stations.
Meanwhile, back on planet earth, the division of the DOE that actually is buying supercomputers is the NNSA, who use their gear for somewhat more nefarious purposes.
Imagine how real the weather will look like in WoW with that computer!
Oh no, they didn't. http://news.bbc.co.uk/2/hi/technology/6600965.stm
They simulated half a mouse brain on that thing! How long until they try to simulate a human brain and let it take over the world?
SLES 10.
Probably not the greenest computer.
There are some lakes in Ohio. You could get nearly the same effect.
This won't be used for that... The energy graph problem is not that problematic or big. Definitely doesn't need that kind of computation power.
Contrary to most people that think a singular way of representing floating point speed is FLOP, it is FLOPS because FLOPS is not plural. FLOPS is Floating Point Operations Per Second. So, I chuckle everytime I read 1 PETAFLOP. Guys, just turn off your singular/plural alarm and say with me 1 and only 1 PETAFLOPS.
I once heard that if you predict the weather for tomorrow to be the same as today, you're right more often than not.
I even heard it was close to 60% or something.
With those kind of numbers who needs these fancy computer whatchemecallits anyway?
I know these numbers are complete unsubstantiated and all, but since everyone else here doesn't seem to mind I just thought it best to drop 'm on you.
News about the Kettle Open Source project: on my blog
Is this really such a huge achievement when you consider that the Riken MDGRAPE-3 achieved 1 petaflops this time last year? Yes it is a special purpose machine, but it still reached the petaflops mlestone.
As far as I'm concerned, all peta flops should be thrown in jail.
One of the embarrassing facts about Super Computers is they aren't quite as stable as they could be. While industry talks about the 5 9's of uptime, some of the shipping systems like to talk about the 2 or 3 9's of uptime.
How stable is this beast?
Seismic processing. Seismic exploration is where you create a big noise on the surface of the earth (a "shot"), the sound goes down, bounces off of geological boundaries, goes back up, and is recorded by arrays of phones. The number of shots can number from the hundreds to the hundreds of thousands.
It can be done on both land and marine. The marine one generates massive amounts of data, because the boat sometimes goes for weeks recording data around the clock, dragging the phone behind them on long streamers.
The data requires massive statistical and numerical processing to be usable. And the biggest pig is something called "wave-equation prestack migration", which sometimes requires hundreds of CPU years to compute (extreme case). But the client wants the wave-equation migration to be done in a few calendar weeks. Fortunately, each shot can be processed independently - thus wave-equation migration has been termed "embarrassingly parallelizable".
With all due respect, you're talking about twenty years of separation. Not to mention that the Cray was a 64-bit design, with enough registers to put any modern processor to shame.
We at slashdot are scientists, specialists and kernel hackers. Your FUD will be found out.
A tricky question, but not all that interesting. A fast server processor is within a factor of 4 of the fastest supercomputer processor in the world. That does not mean that you can do equivalent work with the server processor. Among other things, processing performance (gigaflops) of a CPU, is no longer the interesting part of a supercomputer. (It never really was) memory bandwidth, interconnect bandwidth and latency, and I/O performance are the more interesting features of supers. 12 year old Cray processors still have five times the memory bandwidth of modern PC processors, and twenty times the I/O bandwidth.
You'll notice, that 98% of the supercomputers, sold in the last 10 years, all use server processors. (Blue Gene actually uses an embedded systems processor, but it's the same idea) However, in the late 80's putting 256 processors in a super was cutting edge. In the 90's, a few thousand. Soon you'll see a quarter million cores. So supers are actually getting faster at a higher rate than are desktops, at least by most measures.
Basically, the more processors you have (and the faster they are) the more finely you can divide the problem.
For example, say you're trying to determine the hydrodynamics of a new kind of ship's propeller. With one generation of hardware, you might have had to assign each processor one cubic centimeter of water. With this new generation, you might increase the resolution so that each processor is simulating a cubic millimeter instead.
This is a massive over-simplification, but it's enough to show you what I mean.
Clear, Dark Skies
And the part about each chip having 4 cores and a 2x2 ft circuit board containing 32 chips. 128 CPU cores in the space of an ATX motherboard is pretty impressive even at only 850Mhz each.
processing performance (gigaflops) of a CPU, is no longer the interesting part of a supercomputer. (It never really was) memory bandwidth, interconnect bandwidth and latency, and I/O performance are the more interesting features of supers.
I always hear this, but I've never seen anything terribly definitive on it. I'd like to see how fast a Cray for 12 years ago, and a modern top-of-the-line Desktop PC with a hot graphics processor could solve a problem designed to run on that Cray from 12 years ago. Metrics are nice and all to attempt to judge processing power between different machines, but in the end the only thing that matters is real world performance on an actual problem.
AccountKiller
Looking Apple's recent market share and stock price, I'll go with a "no" on that.
How to solve most of our problems: 1.Lots of nuclear plants. 2.Cure aging.
A high end laptop with a core 2 duo at 2.4GHz rates around 20Gigaflops
You can probably overclock it to 25 Gigaflops, and it is 25mm thick(Opened up).
The GPU probably adds between 20 and 200 Gigaflops (Nvidia claims 500GFlops on the 8800) http://en.wikipedia.org/wiki/GeForce_8_Series
Maybe an overall estimate is 50 Gigaflops total is reasonable.
To get 2 Petaflop you will need 40 of these or a stack of 1 meter.
don't cut it off www.mgmbill.org
now that's funny... Thanks for the laugh... having a rough day with bugs!
- Mike
Once you've lost your temper, you've lost the argument - Me
As long as Lem has been mentioned, there is also "Non Serviam" (in "A Perfect Vacuum") in which the "Latest IBM models have a top capacity of one thousand personoids". Said personoids occupy themselves, among other things, with debating the existence and nature of God (ie the programmer/person running said IBM).
Oops, off by a few zeroes I guess
don't cut it off www.mgmbill.org
Thats just short term profits, I think in the long run turning their products ( computer that is ) into 'yet another intel box' will come back to bite them in the long run.
Dont get me wrong i AM an apple fan, and have been since day 1, but I still feel it was a short sighted sellout. But i guess only time will prove if he was right or not. If they are still relevant in the personal computer market in 15 years , then he was right.
---- Booth was a patriot ----
How is simulating weapon decay nefarious? There's been some talk about new weapon design being done with the DoE computers, but there is no way in hell the military will deploy a weapon system without thorough real world testing, and even Bush hasn't been stupid enough to break the Nuclear Test Ban Treaty.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
Oh. Well, shit. Nevermind.
"Hey, the third matrix movie would have been good except for the plot,story, and acting." --AC
Can someone please convert this to a useful unit of measure, like how many Libraries of Congress per second this is equal to?
That's all fine well and good -- but does it run SPACEWAR??
:b ;-)
.
== WolfriderV6 == I'm willing to admit that *I just might* be wrong... Are you??
Apple wouldnt have jumped ship on IBM unless they saw something on both roadmaps that made intels look like a more attractive solution.
Yep, check out the SCEC [http://www.scec.org/] project, for example. There are some fluid dynamics and astrophysics code as well. And, well, naturally this was all designed with nuke "preservation" code in mind...
~~~ They call me Little John, but don't let the name fool you...in real life I'm very big.
I once saw a stunning weather device. It was a piece of rope tied to a tree branch. The instructions were nailed to the tree:
Weather Rope
If the rope is wet, it's raining.
If the rope is stiff, it's freezing.
If the rope is swinging, it's windy.
I forget the rest, there were at least a half dozen. But as google is my friend, here's the basic idea: http://www.engravingdragon.com/Page17.html
The living have better things to do than to continue hating the dead.
A Cray from 12 years ago would be a T90. The top of the line was the T932 with 32 vector CPU's. It was capable of 57.6 gigaflops and had a total internode I/O bandwidth of 330GB/s. It maxed out at 8GB of main memory. Compare that to an ATI Radeon x1950xtx gpu running folding@home at ~90Gflops with a half gig of ram and ram I/O of 64GB/s, which is significantly faster than a desktop CPU. So, it really depends on what your problems throughput limitation is, CPU/GPU raw power or I/O bandwidth as to whether a current desktop is more or less powerful than a Cray from 12 years ago.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
How many furlongs per fortnight of Library of Congress would that be able to process?
Did someone say PETA flopped? Wait, what? Never mind.
Not only do you want to push the limits with the most powerful electronic system ever,
you want it to be SELF-AWARE?!?!
..So what if it were to be outfitted with 72 racks full of these?l
http://www.nvidia.com/object/tesla_gpu_server.htm
"...The Tesla S870 GPU computing server is used in tandem with multi-core CPU systems to create a flexible computing solution that fits seamlessly into your existing IT infrastructure."
Surely 294,912 processors are not too many for this, but wait!, CUDA is programmed just like C... No problem!