China Building a 100-petaflop Supercomputer Using Domestic Processors 154
concealment writes "As the U.S. launched what's expected to be the world's fastest supercomputer at 20 petaflops, China is building a machine that is intended to be five times faster when it is deployed in 2015. China's Tianhe-2 supercomputer will run at 100 petaflops (quadrillion floating-point calculations per second), according to the Guangzhou Supercomputing Center, where the machine will be housed. Tianhe-2 could help keep China competitive with the future supercomputers of other countries, as industry experts estimate machines will start reaching 1,000-petaflop performance by 2018."
And, naturally, it's planned to use a domestically developed MIPS processor
Yeah right (Score:1, Funny)
They're using domestically built copies of MIPS processors they copied from someone else (usually wrongly), stringing them together and proving that 2+2=5.
Re: (Score:2, Informative)
Copying the instruction set is not the same as copying the processor.
Re: (Score:3)
You're right, but it looks like they've done the latter. http://laotsao.wordpress.com/2011/10/29/sw1600-and-alpha-21164/ [wordpress.com]
Could You Clarify Something for Me? (Score:4, Interesting)
You're right, but it looks like they've done the latter. http://laotsao.wordpress.com/2011/10/29/sw1600-and-alpha-21164/ [wordpress.com]
It says right in the ShenWei Wikipedia article that it's based on the DEC Alpha but something that strikes me as curious is that your article refers to a chip that lasted from 1995 to 1998 [wikipedia.org]. So I am to believe that by outright copying a fifteen year old chip from a processor line that has been extinct for a decade or more [wikipedia.org] has yielded a modern day competitive multiprocessing chip?
You can convince me they copied DEC's work. You can convince me they violated IP laws. You can convince me that it is their societal norm to ignore restrictive IP laws. Hell, I'll tell you that right now. But to say that they are doing no work to build on top of these chips feels like it must be erroneous unless what we see is 1990s technology in the ShenWei processors.
This isn't a black and white scenario here. Yes, it's bad that IP laws have been violated. Yes, it's bad that DEC won't see a dime from any of their work being used. But it is also a good thing to have a competitive architecture arise in the world of computing and also it feels good to have a race with other countries for computing power. I can only hope our super computing budget is considered part of the onerous "defense budget" and our leaders who are concerned with a dick measuring contest can dump tons of money into supercomputers for modeling and simulation to scientists while at the same time being able to give the hallowed talking point of "I increased defense spending."
You can start with someone else's good idea, turn it into a great idea and share some credit, right?
Re:Could You Clarify Something for Me? (Score:5, Interesting)
Why is it bad that DEC (a defunct company) can't profit from their imaginary property in a country that has no protection from or laws against such use. The US Government can also appropriate technology and materials from private corporations within their sovereign state without compensation.
DEC attempted to market their solution and failed miserably, they made their money back by selling it to Intel (so actually it's currently Intel's Imaginary Property). If someone else can improve upon their design (which was quite good actually especially in floating point operations) then I can only applaud their work. Just because it's the boogeyman-du-jour that's developing it doesn't make a difference to me.
Re: (Score:2)
You're obviously a stupid shit.
The status of DEC is not based on opinion. The corporation no longer exists.
Re:Could You Clarify Something for Me? (Score:5, Informative)
That old Alpha chip did about 5 GFLOPS with a single core on 666Mhz, so 16 of these at 1.1Ghz would go up to about the 140GFLOPS that are stated on the wikipedia on the ShenWei SW1600. Thats about twice as fast as an i7-930@4.2 Ghz.
So so much for decades old technology. It was just abondoned because there was to little market for it, but that doesn't mean it's bad stuf.
Re: (Score:2)
+1
Re:Could You Clarify Something for Me? (Score:5, Informative)
It was just abondoned because there was to little market for it,
No, it was abandoned because HP/Compaq ended up owning the Alpha and PA-RISC and Intel convinced them that they could lower costs by outsourcing their CPU design and use Itanium instead. There were still a lot of people who wanted to buy new Alphas, and they got stuck with Itaniums instead. The ones that weren't on VMS or NonStop just gave up and switched to commodity x86 and some open source *NIX.
Re: (Score:1)
Re:Could You Clarify Something for Me? (Score:5, Informative)
That old Alpha chip did about 5 GFLOPS with a single core on 666Mhz, so 16 of these at 1.1Ghz would go up to about the 140GFLOPS that are stated on the wikipedia on the ShenWei SW1600. Thats about twice as fast as an i7-930@4.2 Ghz.
You're right. Alpha CPUs were, AFAIK, quite well-suited for multicore operation, though the Chinese must have created some impressive glue logic.
The original 21164 was implemented using a paltry (by today's standards) 10 million transistors. Using 350 nm technology, at that. The Chinese are capable of reducing that by about an order of magnitude, achieving a significant speedup because of the smaller gates - that's just by using the new cleanroom microfabrication tech.
Actually, I'm wishing good luck to the Chinese engineers. And a big fat "fuck you" to the managers/CxOs that doomed the amazing technologies from DEC (Alpha wasn't the only one that died on the chopping board of corporate stupidity).
Re: (Score:2)
Re: (Score:2)
It's hard to make a fast chip that uses lots of power that doesn't fry the chip.
Fast means small. Using power means heat. Small + heat means thermal noise + atom migration (gates dying). For each size, there's a maximum reasonable heat dissipation. (And, yeah, less is better, if you can keep the speed that you gained by going smaller.)
LOTS of tradeoffs here. And I'm not a hardware guy, so if one corrects me, believe him. But those are genearlly accepted rules of thumb.
Re: (Score:2)
Re:Could You Clarify Something for Me? (Score:5, Interesting)
Have you ever used one of those things? They were amazing chips for their time, especially with their bus architecture. They were great for SMP usage.
AMD and Intel managed to get ahold of a lot of the developers and IP related to these chips, and wedged it into their systems. I'd be very surprised if you still couldn't find traces of it in their systems today. I know not long after AMD got their share, their SMP performance shot up massively, and when it comes to SMP use, they are still better at it than Intel (though, their per-core lack of performance, sadly makes up for this).
So, yeah, with a die shrink, I could see these being amazing for a multi-core behemoth, and competitive with anything extant on the market right now. The only reason we don't see these today, I suspect, is because Intel got most of the IP, and used it to make the Itanium, and the wouldn't produce a competitor for their pet pink elephant.
Re: (Score:2)
Nobody said anything about being a modern day competitive chip.
I'll agree they've probably shrunk the die and increased the clock speed.
But, since they haven't increased the L1 and L2 caches over the late 90s DEC version, I doubt they've done anything else radical either.
Re: (Score:2)
Yes, it's bad that IP laws have been violated. Yes, it's bad that DEC won't see a dime from any of their work being used.
If the chip in question came out in 1995 and the supercomputer is scheduled to come online in 2015, what exactly is the problem?
Re: (Score:2)
How does a copy of a DEC Alpha processor end up with a MIPS instruction set??
Re: (Score:2)
I haven't seen MIPS mentioned outside of TFSummary, so I think that was an editorial brain fart.
Re:Yeah right (Score:4, Insightful)
Re:Yeah right (Score:4, Insightful)
Building massive, highly functional supercomputers is not child's play, regardless of your beowulf clusters of hot grits down Natalie Portman's pants experience. It's one thing to cluster a few computers together and to have very specific programs that do very specific kinds of jobs, it's another matter entirely to have hundreds or thousands of microprocessors working in tandem and to be able to simply even allocate their tasks, let alone program for them. There's a reason why every city government has their own supercomputer, they're difficult.
The Chinese government has the resources to build such a computer, but only if they work against corruption and don't delude themselves when they have difficulties in an effort for every middle manager to safe face.
Re:Yeah right (Score:5, Insightful)
And please, although the Chinese government is very corrupt, it is not more corrupt than US government or US corporations.
Re: (Score:1, Funny)
Re: (Score:3)
Re:Yeah right (Score:5, Insightful)
And please, although the Chinese government is very corrupt, it is not more corrupt than US government or US corporations.
It's very fashionable to overstate the problems of the US. Even with all it's problems it remains one of the most successful systems in the world on any number of levels.
That said, the rampant corruption on China isn't the kind that will interfere with things like building a supercomputer. Quite the opposite in fact. Need a neighborhood demolished or workers expropriated? No problem.
Where as the much smaller level of corruption in the US is almost precisely targeted to screw these kinds of projects. Congressman can't tack on some random spending for their district? Screw over the whole project just to build a reputation so everyone bends over next time.
Re: (Score:3)
Re: (Score:2)
It's been getting less successful with increasing speed over the last couple of decades. Not only on a relative scale, where it's success growth is startlingly negative, but even on an absolute scale. You can lay much of the problem on existing IP laws, but there are other laws favoring corporations over others, and other laws favoring the established over the challengers, and other laws favoring the powerful over the weak...do you see a pattern? These aren't the same laws, and they each have different j
Re: (Score:2)
It's very fashionable to overstate the problems of the US. Even with all it's problems it remains one of the most successful systems in the world on any number of levels.
And it gets very unsucessful for others. USA life expectancy is now lower than Cuba [wikipedia.org], for instance.
Re:Yeah right (Score:4, Insightful)
the natural tendency is for them to surpass stagnated US sooner or later
Probably later. Its common for Chinese fanbois to paint the US as some stagnant, bloated, lethargic country, and in some ways it is, but not in technology. The US still leads the world in technical innovation, and China is still playing catchup, and will for some time. Militarily China is 20 years behind in submarine technology, has one aircraft carrier (Russian surplus), is just now introducing stealth technology in its aircraft, and still sends most of its elite students to US schools for hi-tech education. Oh, and lets not forget the army of hackers the Chinese government STILL employs to spy on American hi-tech corporations right now.
NATURAL tendancy? How is that? 100 years ago China was nation of drug addicts beholden to the British Empire. Natural tendancy my ass. The US is mired in debt and a stuck bureaucracy right now, but to count it out is a bit premature.
Re: (Score:2)
Re: (Score:2)
Militarily China is 20 years behind in submarine technology, has one aircraft carrier (Russian surplus), is just now introducing stealth technology in its aircraft,
And no military significant enemies where they'd be useful, nobody would be crazy enough to attack 1.35 billion people with tons of manufacturing capability and 240 nukes unless China goes on the aggressive and say invades Taiwan and the US gets involved. And in that case it won't be about subs or aircraft carriers or stealth planes, it'll be about the US having many, many thousands of nukes and how poker face they can play another Cuban missile crisis. Any other country? Draft a hundred million soldiers an
Re: (Score:1, Insightful)
USA is stagnant but not because it lacks military technology, it does not. It spends more on military than anybody. USA spends more on military than the next top 10 countries combined [aol.com].
So what? That's not the point, USSR spent everything on military, I mean every non-trivial productive facility had some military application. Where is the country?
To have a sound economy the point is not to spend insane amounts of money on military, the point is to allow the people to save their money and invest it in any w
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Its common for Chinese fanbois to paint the US as some stagnant, bloated, lethargic country, and in some ways it is, but not in technology. The US still leads the world in technical innovation, and China is still playing catchup, and will for some time. Militarily China is 20 years behind in submarine technology
To label proponents as "fanbois" does not exclude yourself from being the same for the opposing camp.
You seem to be under the delusion that there is some divine gift to America by some higher being for technology savvy.
Anybody who actually partake in these activities (and I would suggest that you are probably not one such practitioner) knows that such exclusive gift exist mainly in the minds of born losers.
America happened to offered a number of economical advantages in the twentieth century that fostered i
Re: (Score:2)
America happened to offered a number of economical advantages in the twentieth century that fostered innovative activities.
"Happened?" THAT is where you fail, grasshopper.
Bull (Score:1)
And please, although the Chinese government is very corrupt, it is not more corrupt than US government or US corporations.
Transparency International says otherwise. Way, way otherwise.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Now, say you're the next someone else. You also have to find a way to demonstrate success, and even if you find the flaw you might not be able to push it back on t
Re: (Score:2)
Re: (Score:2)
As to Communism, the idea itself is great, but requires vast amounts of resources and requires everyone to accept being on an exactly equal playing field with everyone else. Life itself has found itself in competition for its very existence, so it's not a surprise that we take competition to artificial levels, beyond merely providing each of us with just enough, but to then start hoarding, not just upper-middle-class hoarding eith
Re: (Score:2)
The main difference between China and US is that China is going from communism to a more free-market oriented economy with minimum interference from the government, which worked qui
Re: (Score:2)
Your comments are interesting, but I'm not sure they describe the current systems. China seems to be producing more capable scientists than does the US, and possibly than does the rest of the world combined. This doesn't prove that they dynamic that you are proposing isn't present, but perhaps it isn't any stronger than the CYA policies of corporate life. And clearly most US students look at a technical degree and figure that it's not worth the time and effort. I, personally, look at it and consider tha
Re: (Score:2)
I'll be even more impressed when I can buy said processor and use it in one of my designs. I suspect that may be a while.
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
You crack me up sir. You make it sound like those are two different groups.
Re: (Score:1)
And please, although the Chinese government is very corrupt, it is not more corrupt than US government or US corporations.
Yes, because in the US, bridges regularly collapse because they are made of garbage [weirdasianews.com] and corruption measurements support your view [wikipedia.org]. Oh wait, that was China. And they don't.
Re: (Score:2)
US has its share of awkward moments, too, regarding misuse of public money and badly constructed structures, and not every (nor even most) structure in China collapses as you seem to think
Re: (Score:2)
China has less government interference than you seem to believe, and the US has a lot more.
Of course, this would only make sense if I were to believe in a free market, but it has long been obvious that not only do *I* not believe in a free market, nobody powerful in either politics or the economy does either. What matters is not how much government control there is, but what the governmental controls are. Of course, micro-managing is always destructive, so you need to keep the regulations within sane limi
Re: (Score:3)
I seriously like people who whistle in the dark to overcome their angst.
Re: (Score:2, Insightful)
There's no reason why the West shouldn't accept China as an equal. They're rational, hardworking, and contribute to modern life. They can be our friends.
At the beginning of the 20th century, the UK became worried and self-concious about the rise of Germany. Soon enough, they were diplomatic enemies, and then the war started. And yet, after the war - and its followup - the UK found out that life didn't end when Germany was as strong or stronger than them.
This is the beginning of the 21st century, and the US
Re: (Score:2)
It wasn't just the rise of Germany as a Great Power, it was Germany's not so subtle designs on Britain's allies and economic interests.
Re: (Score:1)
Germany's designs came out of its rise as a Great Power, and its rise as a Great Power came out of fundamental economic and political forces - industrialization and unification.
In the end, there was nothing Britain could do to stop Germany's rise, so Britain should have accepted a diminished place in the world and come to a new understanding with Germany regarding the balance of power in the world.
This could have been achieved diplomatically, if both sides had been honest enough with themselves to see reali
Re: (Score:2)
Bollocks. On the eve if WW1 was the supreme naval power and had a far more expansive empire. Germany was left with the leavings from the older colonial powers. What you were describes the Ottoman Empire and to a lesser extent France.
Re:Yeah right (Score:4, Informative)
Re: (Score:2)
Honestly, given the nature of Chinese copying, plus the kind of overstated and shoddy output we've historically seen from State-Capitalist projects from Soviet governments, I think that the burden is more on the Chinese to make this boast a reality.
Ditto. Sounds like just more goose-stepping communist propaganda to me - if they can actually do it, good on them, but I'll believe it when I see it. For now, move along - nothing to see here.
Re: (Score:2)
Building massive, highly functional supercomputers is not child's play
Yep, it is a wonder that US manages to pull that one even after it had run out of German scientists. Seriously, next time you want to make fun at China for being unable to come up with anything without using knowledge from another country, look at who managed your space program during the Cold War...
Re: (Score:2)
Pentium 5s then?
Re: (Score:2)
Re: (Score:2, Interesting)
You know a handful of British guys with a BBC Micro designed and taped out the ARM, right? That was 25 years ago, before FPGAs. I don't know why anyone thinks China in 2012 can't put together enough math & EE knowledge to prototype and tape out a microprocessor. Half the people on Slashdot could do it if they had state funding and limitless time. You could probably do it yourself, using Wikipedia to research architecture, and Xilinx WebPack to program the FPGAs. Start with an first-generation 32-bi
Re: (Score:2)
Regarding the summary, how is it "naturally" that they're using a custom CPU? I haven't followed the supercomputing field that well, but I was under the impression that most of them used "standard" CPUs like x86 and IBM PowerPC. The Titan SC uses Opterons and NVidia "GPUs". Well I just RTFA and it has been reported before; "A clear example of this [investment in homegrown tech] was when last year China's Sunway Bluelight supercomputer grabbed headlines for using a domestically developed processor, the Shenw
Re: (Score:2)
Re: (Score:2)
But Americans will sell them the insurance (Score:4, Insightful)
Because that's all America does anymore.
MIPS and 5x (Score:2)
5x faster sounds ambitious, but not off the wall for 2015. It gives them 3 years during which time one might expect a 4x increase in speed. I expect that they will get the top spot for a bit before being passed by something else relatively quickly.
The top10 is quite interesting.
There's a bunch of PPC BlueGene, UltraSparc VIIIfx, Xeons, Opterons and GPUs.
As for the processors, this will be an interesting workout for them. I think like the UltraSparc VIIIfx, it will be hard for them to match the Xeon or Opte
100% correct (Score:2, Funny)
I Spy (Score:1)
The reason for the size is that the censor chips will have to watch what the working chips try to share with other working chips.
Because they're not going to let us do to them (Score:2)
Because they're not going to let us do to them what they know they've done to us.
Hey free trade everybody.. get your free trade riiiight over here! Free movement of goods services and people ... hey ... whatsamatter with you??? Doncha' love FREEDOM????
Re: (Score:2)
As long as there are companies trying to reverse engineer the Chinese chips then if there is a backdoor, it will be found. If a backdoor is found it will be really damaging to their reputation.
It's one thing to be suspected of espionage, it is another thing altogether to be caught putting backdoors into their products.
Re: (Score:2)
It doesn't seem to have damaged Microsoft very much.
Re: (Score:2)
And the evidence of their "backdoor" is what again?
By 2015 (Score:2)
We should be introducing out first Exaflop machines.
To preserve national security (Score:2)
Re: (Score:2)
Re: (Score:2)
100PF project in Santa Fe (Score:1)
Re: (Score:2)
But only Red Flag Linux
Re: (Score:3)
A domestically developed MIPS processor?
I wonder how much of that "domestic development" was looking at what foreign companies were making in their country and surrounding areas?
Has anybody even been working on a MIPS processor that they could outsource, for anything larger than little router boxes and things, since DEC died a horrible death ages ago?
Re: (Score:2)
since DEC died a horrible death ages ago?
What has DEC got to do with MIPS except that they used them in pre historic, or possibly paleological times. SGI worked on MIPS for years after DEC abandoned them. The lates generation R??000k were pretty snappy and scaled up very well.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
http://en.wikipedia.org/wiki/Loongson [wikipedia.org]
http://en.wikipedia.org/wiki/Ingenic [wikipedia.org]
Re: (Score:2)
Like it or not, the Chinese have as much right to take what others have made and build on it as the US had for most of the last century. Remember, modern rocket technology was invented in Germany and the computer was invented in UK; and a lot of other things Americans think were their own inventions came from elsewhere.
Re: (Score:2)
Unclean hands aren't an excuse for further wrongs.
Re: (Score:2)
Oh, give it a rest, will you?
This is very much the way things work in real life. If I see somebody doing something that I think I would like to do, of course I'll learn from it and, in your words "steal their idea".
And, at the same time, it looks like a lot of people on this list find it obviously right to make a 'backup copy' - or 10 or 100 - of any music or data CD, more or less. File sharing is commonly seen as acceptable; but if the Chinese, the devils, do something similar, then it is THEFT and deeply
Re: (Score:2)
Not even close.
You'll be lucky to simulate a gram of matter to any significant accuracy, and the higher-level simulations (neural networks, etc.) are seriously lacking in their ability, and to model how a brain of any significant size works you're honestly looking at supercomputers the size of the planet with current technology.
Just throwing power at a problem like modelling the brain isn't going to make anything happen any time soon.
If we can accurately model an ant's brain, down to the individual neuron,
Re: (Score:3)
Re: (Score:3)
No, your phone does edge detection to get a series of numbers. There's a HUGE difference between that and actual intelligence.
I can program a computer in seconds to do maths that's beyond any human's capability to work out in their lifetime. That's not intelligence.
Now if you could say on each image that "this is me", "this isn't me", and it could build up a database of people that should and should not be authorised and OVER TIME learn on it's own without just having a bunch of statistics like "> 20%
Re: (Score:3)
Re: (Score:2)
> Now if you could say on each image that "this is me", "this isn't me", and it could build up a database of people that should and should not be authorised and OVER TIME learn on it's own without just having a bunch of statistics like "> 20% green = > 90% probability", then you'd have some mild form of intelligence. Otherwise you just have heuristics, which are 0.000001% of how actual intelligence operates.
Cute. From which orifice did you pull the 0.000001%? Do share with the rust of us your elabo
Re: (Score:2, Funny)
If we can accurately model an ant's brain, down to the individual neuron, in the next 100 years or so, I'll be impressed.
If you can do that without violating any patents, EVERYONE would be impressed.
Re: (Score:2)
Re: (Score:2)
You don't carbon-copy things. That just makes a brain, or a virtual brain, not a computer-that's-as-good-as-a-brain, which is what we're discussing.
But the neuronal connections (10,000 per neuron, constantly changing) and time-slices that you'd need to get even close to simulating the number of interactions in an ant's brain when it smells a certain scent-trail? You might JUST about be able to approach that complexity if you turned the whole Earth into a computer with the nanoscale technology.
It's not a q
Re: (Score:3)
But the neuronal connections (10,000 per neuron, constantly changing) and time-slices that you'd need to get even close to simulating the number of interactions in an ant's brain when it smells a certain scent-trail? You might JUST about be able to approach that complexity if you turned the whole Earth into a computer with the nanoscale technology.
I'm sorry, but I'm not buying that. Are you suggesting that the difference of the density of features between brain tissue and anything that we can reasonably manufacture is on the scale of 10^10? (6*10^24 kg for Earth and 10^-6 kg for an ant brain, with negligible difference in density between the two, so 30 orders of magnitude of difference in volume) This is ridiculous, that would mean that we would have to simulate a single neuron (20 um?) using a 100km-sized device. That seems really ridiculous, no mat
Re: (Score:3)
You are way off. 1st off the neuro nets in use today doing amazing things are based off a really really simplistic abstracted concept of how neurons work. You don't have to simulate the whole thing, you can crudely approximate it and get world changing results and we've already done this using a concept from the 1940s. With advances in our knowledge we can make the approximation more accurate without simulating all the details--- in fact, I worked with a guy for a short bit who was working on a mainfram
Re: (Score:1)
Re: (Score:2)
I don't think that by "brute force" he meant doing quantum state modeling.
FWIW, there is currently in progress an attemt to model the entire nervous system of C. elegans. Admittedly that's only 35 neurons or so, but once that is done we should know what parts of the neuron it's important to model.
So here we have the need for a neuron, or perhaps synapse, level model of the neural system as a "brute force" approach. I.e., we model the neural system as it exists, without understanding it.
OTOH, I've never ma
Re: (Score:2)
China is second world. Commies are second, democracies and those aligned are first, third is everyone else, it's political alignment, not social or economic.