Slashdot Mirror
China Switching To Home-Grown Chips For Supercomputers
rubycodez writes "The Tianhe-1A system will be the last Chinese supercomputer to use imported Intel and AMD processors. By years end, China's own 64 bit MIPS-compatible 65nm 8-core 1GHz version of the Godsen (Longsoon family) processors will be used, including 10,000 of them for the 'Dawning 6000' supercomputer. Yes, the chips can and usually do run GNU/Linux, but also can run FreeBSD, OpenBSD, and NetBSD."
I am not sure if I should be worried of relieved now that we get Chinese instead of American spying equipment inside our processors.
Probably correct. The 19th century belonged to Europe, 20th to North America, and 21st to Asia. History keep changing, but considering that the population of Asia is so large and that China does not really rely on superstitions a Chinese hegemony may last longer than any based on European/Middle East traditions.
That sounds like a good thing. These guys actualy make stuff instead of trolling the world with ever more batshit insane *CTA treaties.
As a Chinese American, I'm glad to see China using their own technology, but it's hardly any sign of world domination, especially when the Chinese chips aren't anywhere close to Nehalem, Fusion, or Sandy Bridge. China has already forced Microsoft to hand over the source code to Windows previously, and being aware of exactly what you're getting from a foreign company or agency is a wise move for any developing nation. Remember the big debate over the NSA_KEY variable a while back?
Besides, when it comes to spying, I would take Mossad over Intel or AMD anyday.
The processor family is called Loongson and not "LongSoon" as summary says. But the typo is funny in its own way.
China does not really rely on superstitions
Have you seen Chinese medicine?
Stick Men
Speaking of which, it does make me wonder about all this fuss over 64 bit ARM chips for datacentres. There are already high performance, low power 64 bit MIPS chips and have been for years
Not really. Low power MIPS64 chips use 10-20W. Low power ARM chips use under 1W. They're both low power within their various domains, but the ARM chips get a lot more performance per Watt. Most of the time, the MIPS chips are more interesting for supercomputing, because they have better floating point, better interconnect (there's a lot of experience floating around building large MIPS systems, a lot from ex-SGI people), better toolchains (MIPS has been in HPC so long that it's a standard target for compiler in that market), and better overall performance.
The ARM chips are interesting because a lot of server tasks are not CPU-bound. You can stick 64 ARM SoCs, each with enough flash and RAM to run a small business server, in a 1U case and not worry about heat. You can connect it to a big SAN for storage of data (just put the OS and apps on the flash). Idle power usage can be a few mW per server, power usage under load is basically the power usage of the SAN - the rest of the hardware is adding 1W or so.
It's a mistake to confuse the server and HPC markets. They have very different requirements.
I am TheRaven on Soylent News
I wonder how well these chips compare to the R16000's?
"To those who are overly cautious, everything is impossible. "
and anyone want to fill me in why 10,000 8-core MIPS chips at 1ghz can be expected to outperform 12,000 12-core x86 chips at 2.1ghz?
I missed that claim in TFA. There are very good reasons for wanting to use their own chips though. They have a lower power envelope (around 20W for the quad-core version), but more importantly they are helping to ramp up the economies of scale for the production. 10,000 is about the smallest run you can do for a CPU and it be cost effective. A single order of that size makes the per-unit cost low enough that it becomes attractive for other companies (or projects within China), which helps fund future development, rather than sending the R&D money overseas to the USA / Israel.
It's worth noting that these chips are now produced with all of the relevant patents licensed from MIPS Technologies, so they can legally be sold in the USA.
I am TheRaven on Soylent News
you are completely wrong. this processor has over 200 x86 emulation instructions, allowing it to run x86 code with only a 30% performance penalty, under qemu. it also has two 256-bit vector pipelines that provide SIMD floating-point operations so powerful that a single 1ghz core can do 1080p at over 100 frames a second. to claim that "it will never work" in the face of evidence that you simply haven't looked at is ridiculous. look up the specifications on the GS464V, please. also, you are not aware that the Chinese Government has purchased 25% of MIPS, and is working with the MIPS teams in the U.S. to create this processor. this processor *IS* MIPS's high-performance, low-power 64-bit MIPS chip.
the article has missed out some important information, which is that they are planning two versions of the CPU. the first is a Quad-Core 65nm, and the second is a 16-core 28nm, which will use the same amount of power (about 12-15 watts). hopefully they will also do a Single-Core 28nm which would be under 1 watt, because at 1ghz the SIMD units are so powerful they can do 1080p at 100 frames per second. really, this CPU design is a game-changer. i've been advocating their use for some time - http://lkcl.net/laptop.html
Taking a snapshot of where the Longsoon is now and comparing against where AMD and Intel are now is flawed. The processor business chases moving targets, rather than comparing single samples you need to look at a longer history to try to estimate the rate of change.
Intel started 30 years ago. The Longsoon project started 9 years ago. In that time they have closed the gap on Intel to about 3 years. This 65nm design is comparable with something from about 2007 (the clock speed is lower but having 8 cores helps a lot). The real question is where they will go next.
If they meet their stated plan they are going to skip the 45nm node and make the Longsoon 3B on a 28nm process. They are aiming at a higher clockspeed, more cores and a large integrated vector co-processor that would rival Fusion or Larabee. If they can do what they claim then they are in the process of overtaking Intel and AMD now and we will see the effects on the world processor market over the next five years.
Whether or not they can do this is a big question, and according to the stories in the press it caused quite a debate at HotChips when they announced these plans. It's not clear who will be licensing them a 28nm fab, or quite how they've packed that much into a design. It's not clear how AMD and Intel will respond to a new competitor with state backed funding and a huge protected market.
The next five years will be interesting times...
Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
Taking a snapshot of where the Longsoon is now and comparing against where AMD and Intel are now is flawed. The processor business chases moving targets, rather than comparing single samples you need to look at a longer history to try to estimate the rate of change.
I'm sorry, but it's ridiculous to think that because Longsoon starts today at 1GHz, that they will be able to accelerate faster than Intel and eventually overtake them. The rate of change has got nothing to do with the starting point. A 1GHz MIPS core is easy to make by today's standards, so it just doesn't mean anything.
And the men who hold high places must be the ones who start
To mold a new reality... closer to the heart
It also allows you greater flexibility in chip design. The Japanese are still convinced that vector processors are still the way to go. The earth simulator had a lot of Japanese-designed vector CPUs and the K computer is no different, it has 2x as many SIMD units per core as the Intel/AMD CPUs. There are lots of benefits to using the vector CPUs in parallel computing, but the problem is that there is very little demand in the personal/corporate world for them(outside of a few specific applications). By designing your own chips you can incorporate the vector CPUs into your chip, but at a cost, both money and an opportunity cost. The money is quite obvious, designing and fabbing your own chips is not only expensive, it takes a long time. Often times by the time you have designed, verified, and fabbed your chips advances in commodity CPUs often obviate the gains made by using your own chips.
That being said, low-power supercomputing is going to be the way forward as often times the cost of operating the computer dwarfs the initial design and manufacturing costs, especially if you are able to sell your cpus on a large enough scales. By including the vector processor on the CPU(instead of say in the GPU), you have the potential to save massive amounts of power.
Monstar L
You make a good point. One of the tragedies of the US is our frontier and pioneering spirit. (Not that other people and countries don't have the exact same thing. But the US just happens to be the biggest right now.) We do the hard work of inventing a lot of things, the hard work of refining the processes. And then other countries and peoples learn from our mistakes and do "better" than we did at it.
Of course, it would probably have been a lot harder for the Chinese if some Intel or AMD partner hadn't sold their fab plant to the Chinese.
I'm not complaining- I'd still rather be in the US. But it is galling to hear people make comparisons that just don't work. It is easy to improve upon something that someone has already been busting their asses on. This doesn't make the Chinese "better" (nor does inventing it first make the US better). It's just a different thing.
At work the other day, someone was pounding their head against the wall trying to figure out why their computer wouldn't read a DVD. Hours. They ask me what I think might be wrong. "Maybe the disk is corrupted?" Sure enough, the disk was (functionally) blank. They effused and groveled at my "genius". Fuck no. They did all the other things, I just identified (guessed) at the thing they didn't try.
Also, the pioneers get the arrows and the settlers get the land. Beware!
Intel and AMD are hampered by having to provide legacy compatibility, MIPS is a much newer designed architecture that should impose less bottlenecks on processor advancement.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Note that something to add is that until very recently ARM were 32bit only, which is not very good for datacenters.
They added a kludge on the ARM ISA (not as eleguant as the MIPS64 ISA) so now it isn't an issue anymore..
China has basically killed the rest of the world's economies with the poorly designed and cheaply manufactured products...
Or you could place the blame where it lies: With retailers who will lie to you with marketing about the quality of a shit product, and with consumers who lap up the shit gladly instead of doing some research to find a quality product. China would sell us quality goods if we refused to purchase their crap.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Yes, it's amazing how fast the chinese can reverse engineer old technology! Good thing there are strong copy-protection laws in force to prevent this sort of thing.
All snark aside, this does point out something very important; The Chinese can never surpass the performance of the people they're copying. On the other hand, they can price them right out of the market. The down side (for the entrenched powers) with the world going multicore is that you can solve problems by just throwing more cores at them. Granted, there are plenty of problems which can't be solved in this way, but even a really crappy CPU core of today is shockingly impressive by Ye Olde Tyme standards. When I think about the difference between my old Sun 4/260 and the cute little netbook on my lap with the 1.2 GHz 64 bit processor and the 2GB of ram, which is a kiddie class machine by modern standards, it makes my mouth a little dry.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
We do the hard work of inventing a lot of things, the hard work of refining the processes. And then other countries and peoples learn from our mistakes and do "better" than we did at it.
And all of mankind benefits. Too bad so many people are stuck in the "us and them" mindset.
c++;
Yes, because the chinese are stupid and they don't have any engineers.
What the hell are you people on? Can't you see the clear patterns? China began exactly like every other nation: first they copy, then the invent, then they lead. Compare with Japan. In the 60ies, you spoke of "cheap japanese copies". Then they took over, now you have Toyota and Sony.
Do you really really believe that a 5000 year old civilization with nearly 1.5 billion people, the highest average IQ in the world and lead by engineers, won't figure out how to design a CPU? What will it take for you to wake up?
c++;
The Japanese elite *may* have outlived the European/American elite but I'm gonna [citation needed] you on that one... The Japanese common man, however, certainly did NOT live longer or better than his Western counterpart.
I refer you to "Standard of Living in Japan Before Industrialization: From what Level did Japan Begin? A Comment" by Yasukichi Yasuba in The Journal of Economic History Vol. 46, No. 1 (Mar., 1986), pp. 217-224.
Yasuba takes to task the notion that life for the commoner in Japan was better than that in the West. While economic development HAD been ongoing throughout the Tokugawa shogunate, and circumstances had improved for the Japanese laborer, the reality of the situation is that farmers here and farmers there both were treated very poorly. He also points out, specifically, the flaw in Hanley's research (which estimated life expectancy to be around 40 years in Japan) specifically used a source which excluded year 0 deaths, and then substituted Western infant mortality rates in its place. At the time, Japan would be much closer to India than the West. By using data which matches temple records more closely, Yasuba suggests that the actual life expectancy of the time was around 35, which (again) puts it below the West.
present day... present time... hahahaha...
Your reasoning is impeccable, but I can't begin to count the number of times market forces, amortizing massive investments over huge economies of scale, have trumped common sense.
What's interesting here is how differently China plays this game. They're focused on long-term national prestige and influence, so they can tolerate being a few years behind by specifying the use of domestic products. That ensures the cash flow their enterprises need to catch up. That would. be unthinkable in the US, with the. exception of a few companies like Boeing, and. even then it's ideologically incorrect to be up front. about helping the chosen enterprise. The standard position is that the competitor kettle is blackened by government favoritism.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
A 1990s Toyota is "easy" to make today, but if China starts making them in bulk then Toyota wouldn't be happy.
In a world where slightly outdated chips are "good enough", and the marginal cost of making them is probably a few bucks, I'd be very worried if a really big competitor was breathing down my neck.
...
Bingo! The U.S. lost its consumer electronics industry to Japan in the 1970s because U.S. manufacturers were not concerned with low cost competition on the low end of the market. The low end is always pretty big and massive sales makes massive revenues and massive production; massive production and revenue perfect the industrial processes and leads to superior design and production technology; superior technology extinguishes competition who focuses only on the "high end". The low end will move up and cannibalize the high end market.
Second class citizen of the New Gilded Age
China could go into biofuel-from-algae and probably will if big oil loses much power.
They could likely go Soylent Oil, if they gave up the one-child-per-family law. (I am being modest.)
I feel fantastic, and I'm still alive.
You cant do separate branching *at all* between the multiple scalers within a SIMD vector. All the scalers have the same operations performed on them.
No, but you can do branching between each set of operations. If you're doing a matrix operation, then you can do a couple of SIMD operation on a row, then a branch based on the result. This is pretty fast on most CPUs, it's painfully slow on a GPU.
You seem to be confused about why the negative performance of branching matters on GPU's.
No, I'm not. One of the things I work on is a GPGPU compiler for HPC, so I'm intimately familiar with their strengths and weaknesses and when it makes sense to offload work to them from the CPU.
its not because it impacts their SIMD capabilities.. because it doesnt.. its because it impacts their CPU-like "GPGPU" capabilities... which means... what I said is 100% correct
I never said it did. I said that it affects their ability to handle instruction streams containing branches. A typical instruction stream coming from a piece of C code has a branch, on average, every 7 instructions. If you're doing vector operations on a scalar unit, then this may be every 20 instructions or so, but closer to 7 if you're using vector intrinsics. It needs to be over about 200 before you start to see the GPU being faster (and even that's highly dependent on other factors, including data independence and memory layout).
If you are doing heavy SIMD work, get a pile of GPU's.
Still not true. If you are doing highly parallel (SIMD or MIMD) work that has little locality of reference, predictable data access patterns, and very predictable code flow, get a GPU. If not, your CPU is likely to be faster. A GPU is not just a very fast SIMD unit, it's a processor that is insanely heavily optimised for a relatively narrow class of algorithms. The set of (useful) algorithms that benefit from SIMD is much larger than the set of (useful) algorithms that can run efficiently on a GPU.
I am TheRaven on Soylent News
Why hasn't Intel or AMD exposed a RISC interface? It would allow developers to target either the legacy one or the newer one.
It seems like this would be a great avenue for weaning people off x86 without incurring huge performance losses while at the same time keeping legacy support.
Mod me down, my New Earth Global Warmingist friends!
For one thing, if they don't expose it they can change it anytime they want. Translating to micro-ops isn't a huge performance hit, so being able to improve the underlying architecture without worrying about compatibility is something of an advantage.
Because if they do that, developers and users will EXPECT future intel processors to now support THAT instruction set. Which just make's Intel's problems worse if they change how the RISC instructions work internally.
Motorola and IBM said the same thing about PowerPC when they started. Over the following years the PowerPC got about 20-40% better performance at the same clock rate as the contemporary Pentium, SMP also had a similar performance advantage. However Intel was able to win with actual performance by achieving higher clock rates.
In the end, Intel won because of typical underhanded marketing strategies (which all businesses seem to do, naturally ) and its tie-in to the hardware of the Windows market. They simply ground Motorola to a halt over time. China won't face this problem, though, as they can simply do an end-run around Intel and AMD as well as tell Microsoft and Apple where to put it due to prices that are under their production costs. Maintaining legacy compatibility is killing them. In fact, dropping all of that off of a typical i3 chip would result in a fairly noticeable drop in heat and size as well.
I predict chips dropping in price to 20% of where they are today in ten years and IBM and AMD essentially going out of the chip business as a result(most of their money isn't made on PCs but on smaller embedded devices and business servers). Kind of like how electronic books are slowly crushing the life out of traditional bookstores as the average price heads towards 99 cents.
I know that I'd buy a 3Ghz i5 spec processor that ran Linux for $20 instead of $150-$200 retail. Total no-brainer.