Slashdot Mirror
China Releases 2nd generation MIPS Chip
eldawg writes writes with news of the launch of a second-generation Chinese 64-bit MIPS CPU. "The Godson-2 or 'Dragon' went into production last week. News reports indicate that, 'The CPU is 95% MIPS compatible using an unauthorized and unlicensed variation of the MIPS architecture, which is owned by the American company MIPS Technologies...The Godson-2 is pretty much a copy of the MIPS R10000 which makes it on par with 1995 technology.' The Chinese plan on using these chips in consumer electronics for the local market, but one can assume that they will eventually end up in exported electronic goods. I wonder if MIPS Technology will sit idly by when this happens?"
News reports indicate that, 'The CPU is 95% MIPS compatible using an unauthorized and unlicensed variation of
Unauthorized and unlicensed - duh, of course it is. That does NOT per se make it illegal and it certainly does not mean it is "stolen". Anyone can implement an instruction set (there are decades of precendent for this) - while our system may be really fucked up when it comes to thing like business method patents, on processor architecutre (and electronics in general) it is clear: it's the implementation that counts, NOT the idea.
the MIPS architecture, which is owned by the American company MIPS Technologies...
Do you mean "implementations of which have been successfully licensed by MIPS, but frankly it's a well documented and relatively simple RISC instruction set
that a single person with a few years VHDL experience can implement"? See OpenCores for an example.
The Godson-2 is pretty much a copy of the MIPS R10000 which makes it on par with 1995 technology.'
So WTF are the latest Opteron processors? On par with 1978 technology?
The Chinese plan on using these chips in consumer electronics for the local market, but
one can assume that they will eventually end up in exported electronic goods.
One can be assured that cheaper processors will find their way into everything. Nice try insinuating that the EVIL CHINESE are deliberately out to screw us by EMBRACING CAPITALISM!
I wonder if MIPS Technology will sit idly by when this happens?"
I wonder if MIPS has a choice. See AMD vs Intel ca. 1991
The Godson-2 is pretty much a copy of the MIPS R10000 which makes it on par with 1995 technology.
Which is excellent for vintage music lovers like myself, because all the hardware I've used since 1996 and on has absolutely refused to play my Ace of Base MP3s.
"The Register" has a better write up on this story (sorry guys). Apparently they've managed to get Windows CE, Linux, and VxWorks up and running on the CPU.
As for the patented instruction sets, apparently they aren't used in the chip. (Supposedly that's why it's 95% compatible).
Currently the chip clocks in at 400-500Mhz, but the next generation is going to be around the 1Ghz mark - by which point China is going to be spitting out all manner of sub $200 computers I imagine.
READY.
PRINT ""+-0
If you've ever used an R10k Silicon Graphics workstation, you'll know that these MIPS chips are pretty beefy. The floating point performance in particular on modern MIPS chips is spectacular. (R14k chips are used in Tezro currently)
There's a REASON Silicon Graphics used MIPS chips in their systems until recently. (and they only switched to x86 stuff due to economic pressure, not performance...)
I have a dual processor R12k SGI Octane on my desk and it still beats my brand new P4 out on a LOT of tasks. And that's a seven year old machine....
Plus, these are 64-bit chips.
Sure, the R10k processor is "from 1995". But SGI's policy at the time THEY were using MIPS R10k chips in their $50k workstations was to factor of ten beat everything else on the market. Meaning, their systems were engineered to be at least ten times as powerful as the competition (and ten times the price to boot).
So... Knockoff R10k MIPS chips, built with modern advancements, smaller dies, and scaled to higher clock rates, will perform VERY WELL comparatively. In fact, for some tasks, (floating point) the chip should compete quite well with a P4 1.5 Ghz... and probably be a whole hell of a lot cheaper. And 64 bit I might add.
And since there are already designs for systems with massive numbers of MIPS R10K nodes (Origin 2000 for example) which are considered to be "junk" it's not hard to imagine knockoff supercomputers....
It's just exploiting cheap labour and nothing short of a revolution will make things change for them.
As a sinologist and former resident of China, I disagree whole-heartedly.
You have to remember -- when considering education for the Chinese people, the Communist Party has been a godsend. Under the communist government literacy has increased over a thousand percent.
Chinese culture, as the father of all East Asian cultures, holds education dear and promotes getting as much of it as possible. Their college system is still sub-par when compared to the rest of the world, and when compared to S. Korea or Japan, but it is rapidly improving. Their top schools compete with the world's top schools. Their local schools have been providing valuable training in business management, among other skills, that have allowed the Chinese economy to boom as it has been booming.
And that won't stop. In 50 years they will no longer be the cheap-labor capital of the world, because they will have raised the education bar to a level much higher than it is.
Only then will "revolution" make any sense. Anything before then will just put in a government that is MUCH worse than the current government.
If you want to understand a country's progression towards democracy, you should read books on international development -- especially "second track" or "citizens" diplomacy. The leaders in that field have studied successful migrations to democracy and have learned that democracy fails when "democratic norms" are not in place. Those include education and an entrepreneurial-type business culture (and a stable economy that isn't dependent on the government), among other things.
Until those democratic norms have been established, any democracy would collapse.
Look at Taiwan -- they were a military dictatorship until 1988, and the people who fled to Taiwan had, on average, very high education levels. Even then it took 40 years to bring democracy.
Look at Russia -- Putin is getting more and more powerful, and the people support him. Why? Because they would rather have a burgeoning economy and stability than have democracy.
What comes first, finding a teacher or becoming a student?
I don't get the reference to Hennesy and Patterson at all here, while they do indeed discuss the implementation of pipelined chips with the Mips as the running example they do it on a high level. Doing an efficient implementation of even only the parts discussed in the book requires doing the actual clever implementation work of large portions of logic not really touched upon. Even things that are discussed are in so general terms that the books worth as a practical guide to actually building a modern chip is fairly low (it is an introductionary book after all). In addition they don't even discuss the issues involved in a chip of this level, notable missing parts are OoO logic and the FPU and so on (again understandable since it is an introduction).
Don't compare the R10000 to the stuff that a CS class hobbles together (which also tends to be a very small portion of a complete chip), it is an insult to all of computer architecture if anything.
So, the R10000 was very much state of the art in 1995, and is still doing fairly well today (the R160000 is pretty much the same core, just shrunk and tuned). If China has made an equivalent it is proof enough that they can make a competitive chip.
A brief description with picture of the chip:u /0312/258718.html/
u /0312/258719.html/
http://www.pconline.com.cn/pchardware/foreline/cp
A 13-page write-up documenting the tough work and challenges faced by one of the chip scientists (e.g. pipelines/branch-prediction/cache design, packaging, etc...):
http://www.pconline.com.cn/pchardware/foreline/cp
Interesting bits from those Chinese sites:
- (back in 2003) they're already running Linux on it, with applications such as MP3 audio/mpeg movie playing, Mozilla, OpenOffice, games...
- (back in 2003) Max clock 300MHz, 1-2W power consumption, 1% CPU load for playing MP3, 23% for mpeg movie, SPEC_CPU2000 score of 300
- will reach 1GHz by early 2006
- it will be used in low-cost PC with price RMB1,000
- the 3rd gen of the chip will incorporate multi-core design
Not even the brightest CS junior, or even the brightest (lone) CS grad student in any institution all over the world can design something with the complexity of the R10000, which was a fairly sophisticated design. The Chinese have access to very recent process technologies and can easily design and build a simpler processor that would beat your puny FPGA-implemented broken design in a heartbeat. Their goal is to build a base on which they can build a competitive CPU design infrastructure for their local industry.
I found your post mildly disturbing, with an air of superiority that seems to assume that CPU design is an American specialty of some sort. Many of us who work in CPU design and implementation got our graduate degrees from American universities, where an overwhelming majority of the graduate students in our group were foreign. Your CPU's are already being designed by non-Americans, so this might be a good time to get over it. Also, there is much more to talking about microprocessor design than taking a junior level Verilog class, perusing the Hennessy & Patterson book and maybe reading a few ISCA papers. I suggest you take a look at the R10000 paper published in IEEE Computer some years back.
Zigbee Central: A Zigbee weblog