Rick is there an email that I and Richard Chow can use to reach you?
Richard is a high school student and wants to do a research paper on grid interconnect computers like Godson and merrygoround interconnect like IBM and Intel.
Also do you know the email URL so that Richard can contact Weiw Hu?
China is for real and I just hope people will not be migrating to China in my generation. This country is summonsing feat after feat. Whether this chip has any impact or not, it is amazing what this nation has become. Intel is a monster and it is not possible to compete against them. However, if China gets better, the govt will help them prosper. After all, it is a state capitalism. America be better wake up. China is real.
I guess if china could compete with biggies like Intel....this is in general good for the industry. I think there are not many companies into CPU designs...but low cost products from these chinese could bring down the prices of the CPU'S drastically.
Silicon is but one leg of the tripod.
The other two legs are Price, and Software.
Expect the Chinese to be OK on price, but software is a tougher nut to crack.
Hu does seem to be somewhat skirting WinTel, so there could be niche-room for this.
China has tried multiple times in the past decades to produce its own CPU, as well as operation system, but have never succeeded commercially. There was a national R&D center in the school I have graduated trying to develop a Windows like operation system, they spent more than 10 years and was the biggest line item of the annual department budget but produced absolute no product, more than 80% of the participating members have eventually migrated to the US. The government controlled R&D model simply does not work.
I believe it's the evil side of human nature. Communism is fairly naive in realizing this compared with other western traditions/religions.
china paid dearly for this ignorance in the past, there have to be some mechanism to filter out more evil/sinful ppl in your org. to make it healthy such as market economy etc.
chinese govt. system use communism which could acts as a semi-religious group to promote human virtues. this system could achive some certain level of success ie. rockets, supercomputer etc.
but this system can't achieve the highest level of human cooperation, so china lags in some key area even after many years of struggle like jet engine etc.
about this cpu I bet it could achieve some level of success, hopefully some of its talents could find a way to form a real company to continue it like lenovo did.
I did know another school (namely peking u.) involved in similar acts years ago and generated lots fanfare. in the end it's just wasting money I guess.
Rick, at least when I was still in China, that was more than 10 years ago, the person who has enough clout to bring in huge amount of fundings calls all the shots. That person usually was well connected politically, but not necessarily technically savvy. I have involved in grant applications in both China and in the US, the review process in China was a joke, at least back then. The other team members had no role in deciding what to do and how to do it, and innovation spirit was largely suppressed because the big boss wanted all the credits.
China has improved dramatically in almost all aspects since. but the fundamental issue remain, on a grand scale it is not a meritorious society. Connection is everything. The Asian system, including China, Japan, and Korea, is good for catching up but not for leading, because once the goals are set, usually set by the US, all you need to do is organize a huge group of engineers to work toward it, which they know are solvable and the solutions can be looked up and copied from publications and, nowadays, on the internet. When I was in school, we were trained to focus on prove theorems that have been proven by ancient Greeks or 17th century Europeans, but not on how to work an open-ended problem. This is true to the entire Asia, not necessarily limited to China.
Will, I partly agree with you but I do not think it can all be blamed on the communist nature of the government, the Soviets had a much more repressive regime and they were relatively successful at science, at least they generated a few Nobel Prize winners. I think it has more to do with the culture. Taiwan and Korea have similar culture to China's, but they are much smaller in size, and the students returned from the West can make much larger impact to the society.
I took a look at Godson II chip spec in the web. The package design is totally of out whack on the EM issue, resulting less than 300MHZ ddr interface.
The chip physical design method is P&R and resulting large die and low performance, high production cost. Nobody designs high end cpu using P&R. The only reason to do this is to get result fast and get more money from the government.
The godson team may be able to design cpu which funcationally work. But they can not produce anything profitable within 5 years. The team is not technically prepared for it.
That was not the case hundreds of years ago when China led the way in innovation. It is only normal that after 200 years or so of stagnation, innovation and free enterprise stalled. However, you have to learn to read well before you can start to write. I believe China to be on the road to be a leader in innovation. Whether this processor proves to be a market hit or not remains to be seen, but it is an effort in the right direction.
Wrt your comment on grant application processes in China vs. US, the system that you describe (I would call it feudal) is also normal at this stage. With time, the cream will always rise to the top. As long as the Government goals are focused and well resourced, the system will improve its efficiency with time. On the way, setbacks will surely happen and money will be squandered, but that's part of the process....
I don't agree at all. In the case of China, they started their economic development late, compared with the US, Japan, Korea or Taiwan province. What China needs is time. It is just a matter of time before they catch with Intel, and Intel will be in some stiff competition when that happens. Once China's economy reaches fully developed status, we expect that it will an lead the world in at least some areas of technology.
As China's economy grows, its culture is also changing rapidly. You seem to be fixated in the past, but you need to be prepared for drastic changes in China. If the past 30 yrs of Chinese economic growth was impressive, the next 30 yrs will be even more so.
You would imagine that developing a CPU is time-consuming and capital intensive and no private company can do it. So the government has to do it. Clearly, for the Chinese, it is very important for them to stay closely behind Intel technologically.
It would be nice if each Chinese student can buy a laptop with a Chinese CPU inside rather than a Intel CPU. That will save them a lot of money.
This putting together multiple core has become a fashionable feat these days. Do we really believe that creating a cpu with out of order instruction execution is all it takes to make it big? No, even though this is good as a research project it won't matter much commercially because of obvious reasons and intel is not in any danger from these efforts.
I agree this cpu will not pose any threat to intel etc.
best case is someone ll invest a company to utilize all the knowhows and end up with sth like huawei etc. which is challenging/beating lucent recently.
Let's be subjective and avoid political talks.
I dislike capitalism in the USA and communism in China. Either way is corrupted anyway...
Unless Intel makes many wrong moves and loses focus to position itself at the heart of its niche sector to fend off any competitor, China will certainly pose a threat, if and only if Chinese competitor(s) can safeguard its patents, core design expertise, production cost, consumer confidence in its branding.
More and more Chinese companies made its presence felt globally. Huawei, Legend, SMIC etc are making their way to the Tier 1 league. Their capitals are growing. Just China domestic market and economic regeneration is sufficient to spin the cycles, without export to USA.
More and more designers/engineers in American MNCs are Chinese. Less and lesser are American and Canadian, largely due to engineering is a lack luster as a career, compared to banking and other professionals. Look at the top US colleges, more and more are Chinese professors.
Every year, China can produce 400k engineering graduates. On the contrary, Germany, Japan and USA are falling in numbers.
With USA imposing strong protectionism and arrogant on global stage, it is in certain times USA will lose its shine. American need to re-think beyond technology and reflect upon morals and respect for other countries.
Being a superpower is more than just winning technological edge. With Hillary continues to anger more countries with her "airhead" remarks, that will further backfire American attempts to increase economic yields and definitely semiconductor sales.
The Russian move, is merely a request from a highly self-interested player, and it is doomed to fail. You cannot ban multiple process flows, using a single number, and the biggest irony is, Russia is getting the know-how from ST!!
Let's avoid the jingoistic tripe about countries. We should be welcoming Godson to the world stage, since anybody that's involved knows that it takes a world of talent to design and build a commercially viable world class processor. Let's look at some of the leaders
Intel - Designed in Israel, fabbed in US
Qualcomm - US-based company building a UK-based processor (ARM) SoC using a design team in India, fabbing in Taiwan
AMD - Designed in US or Canada (GPU), fabbed in Dresden Germany or Taiwan
Godson - Designed in China, based on a US-based instruction set architecture (MIPS), fabbed in France ?
Like its older versions, the new Godson will see no success outside of supercomputers and maybe workstations. It will remain a niche product. Besides hardware, you need software applications ported to the processor and that is where the platform is lacking.
For the fourth time this year China's central bank announced Sunday the biggest Chinese banks must hold greater cash reserves. China's financial leaders are struggling to keep inflation in check. A world addicted to cheap manufactured goods from China could possibly be in for a shock. Meanwhile, Donald Trump, pandering for a presidential bid, proposed a 25 percent tariff on Chinese imports that was roundly ridiculed by economists. I read this here:[url=http://personalmoneystore.com/moneyblog/2011/04/18/china-inflation-threatens-global-growth]China struggles to curb inflation that threatens global growth[/url]
Rick is there an email that I and Richard Chow can use to reach you? Richard is a high school student and wants to do a research paper on grid interconnect computers like Godson and merrygoround interconnect like IBM and Intel. Also do you know the email URL so that Richard can contact Weiw Hu?
It makes sense that if China is to make computers that its poorer consumers could afford then they need to avoid costly Western products.
This is happening with 'unconnected', standalone products in certain markets, but not yet fully with products that demand a connected ecosystem like CPUs.
But with device-independent technologies like Java there is less need for a standard architecture, and they can go it alone.
I wonder what their attitude is to IP, if their eventual new processor needs features that breach other's IP? Will they tough it out, or bend to international pressure to do the right thing, or will national pride force them to do a clean-room design that they can claim as their own?
a 464V core! Wow, that is high-powered! ;)
Sorry for adding another very scientific comment to the high-brow discourse here.
Interesting that it's MIPS-based.. It would be helpful to learn more about how they decided on the architecture and how they got where they are now.
Blog Doing Math in FPGAs Tom Burke 21 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...