Supercomputer POWER

Hendrik_2000

Lieutenant General
China surely can possibly a higher quantity of quality engineers than the United States, but software developers? I don't think so...

I don't think that is true anymore. software industry in China goes by leap and bound averaging 20-30%/ year.

China quadruple the science and engineering graduate every years more than any country in the world 2011 number is probably around 6 million
Plus young Chinese programmer win competition after competition . Now how can Huawei become the 2nd largest communication company in the world without software

Or the large Kanji base business software or defense for that matter

China dominates NSA-backed coding contest
Chinese student, 18, wins, prompting call for earlier math and science education in U.S.
By Patrick Thibodeau
June 8, 2009 04:24 PM ET
12 Comments
Please, Log in or Register to view URLs content!

Computerworld - Programmers from China and Russia have dominated an international competition on everything from writing algorithms to designing components.
Bin Jin
Winner: Bin Jin, or 'crazyb0y.'

Whether the outcome of this competition is another sign that math and science education in the U.S. needs improvement may spur debate. But the fact remains: Of 70 finalists, 20 were from China, 10 from Russia and two from the U.S.

TopCoder Inc., which runs software competitions as part of its software development service, operates TopCoder Open, an annual contest.

About 4,200 people participated in the U.S. National Security Agency-supported challenge. The NSA has been sponsoring the program for a number of years because of its interest in hiring people with advanced skills.

Participants in the contest, which was open to anyone -- from student to professional -- and finished with 120 competitors from around the world, went through a process of elimination that finished this month in Las Vegas.

China's showing in the finals was also helped by the sheer volume of it
 
World class developers do not waste their time on algorithms and programming competetions. In industry, you have very, very few people working in areas where you really need those crazy mathematic and algorithmic skills. And it is not surprising for China to be able to dominate those competetions just due to the sheer number of computer science, math, and engineering graduates it is producing every year. However, American schools are still producing more high-caliber developers and architects than China for the time being, and will probably continue to do so until the end of this decade.
 

Schumacher

Senior Member
Well, Chinese CPUs are based on MIPS, and Japanese on SPARK architecture. Neither is indigenous.... Also China's 45nm CMOS technology is licensed from IBM, there is still a long ways to go before China could catch up. Though Interestingly, in the early 1960s, China isn't that much behind the US in terms of semiconductor and computing technology. That's what 10 years cultural revolution can do to you....

The Loongson CPU is MIPS based, not the ShenWei CPU. You're overestimating the role of the the RISC. There're reasons US is limiting the exports of completed supercomputers and chip making equipment to China but didn't stop the sell of MIPS's rights, at a not so expensive price, to Loongson.
With the same RISC, some can design good CPUs while others cannot. You don't need to reinvent the wheel and when commercial/compatibility considerations are involved like with Loongson, you want something existing like MIPS.
China's indigenous chip tech is 65nm, about five years behind the US mainstream of 32nm. Generational gap in chip equipment is not as long as in other tech.
 

Hendrik_2000

Lieutenant General
Anyone discounted super computer as gimmick should read this .Where is challenge aka Richard Que
Supercomputer is one of those enabling technology without which no country can claimed advanced technology. From weather forecast to Aerospace technology, New Drug, Security, etc

Please, Log in or Register to view URLs content!

2011-12-30 (China Military News cited from China Daily and by Wang Tao) -- When the Tianhe-1 supercomputer was ranked No 1 in the world last year, it suddenly and greatly raised the profile of supercomputing in China. This year, with the announcement of the Sunway Blue Light, which uses domestic processors, in Jinan, there are three petaflop-level supercomputers in China including Tianhe-1 in Tianjin and Nebulae in Shenzhen. It is clear that China's computer science and computational science is making great progress.

For most people, supercomputers have an air of mystery, yet such computers have a close bearing on many people's lives. Many fields demand the help of supercomputers. A simple definition of supercomputer is "computers connected by a network". These computers, called nodes, could be personal desktop computers, servers, workstations, and even notebooks. The network connecting these nodes could be an Ethernet, a high-speed network or some specially designed network. Usually when people talk about supercomputers they are referring to the system with many high-performance nodes, a high-speed network, large memory and disk space so that the system has large processing capacity and high computational speed.

Another expression closely related to supercomputing is parallel computation. In normal computers, most programs run sequentially. For sequential programs, supercomputers offer no significant advantage over normal computers. However, for parallel programs or parallel tasks, the computing nodes in supercomputers can work together by using many processors and exchange information through a high-speed network. That means computation can be sped up.

Supercomputers are widely used in scientific research, industrial engineering, information processing, finance and public affairs. They have become one of the most important tools in research and innovation, greatly speeding up research and development and improving product competitiveness. In basic scientific research, a lot of physical or chemical behavior can be observed through experiments, but only at considerable cost. Supercomputers offer the possibility of simulating such behavior so that it can be more easily studied and better understood. In astrophysics and environmental science, people use Newtonian mechanics to simulate large-scale phenomena such as the evolution of galaxies, climate change, and ocean circulation.

In micro fields, quantum mechanics and molecular dynamic methods are used to stimulate atomic or molecular behavior such as protein folding, gene sequencing, nano materials and devices, catalytic processes, and chemical reaction processes. Indeed, the demand for computing capacity is endless in basic research fields. For example, at the Shanghai Supercomputer Center, which has the broadest and most successful supercomputer applications in China, almost 80 percent of computing resources are used in basic scientific research. In industrial engineering fields, supercomputing is mainly used in computer aided engineering (CAE). By using CAE, the product development cycle and cost are reduced greatly, and product competitiveness is improved significantly.

In aviation, people use supercomputers to carry out aerodynamic design and large-scale, detailed and precise computational fluid dynamics (CFD) simulation. In the design and development of China's first new regional jet the ARJ21-700, the Shanghai center's supercomputer was used to finish the design requirements determination, preliminary conceptual design and detailed design. During the conceptual design stage and the preliminary design stage of the ARJ, UCFD analysis was made to reduce the number of wind tunnel tests and comply with fast and highly iterative engineering design. During the detailed-design stage, CFD analysis and design optimization were made for various components and component combinations, for instance, aerodynamic shape and design optimization of airfoil, wing, nose, tail, nacelle and pylon. In one project now being carried out on the China-made large passenger aircraft C919, the Shanghai supercomputer has been used for configuration design and aerodynamic design, which includes the nose, wings, wingtip, nacelle, and body design optimization.

In the automobile industry, the supercomputer can be used for finite element analysis and structure analysis. In car design, passive safety is very important. In the research and development of vehicles, the virtual crash analysis of safety design can be done on a supercomputer. With such computing resources, whether it be the number of virtual crash tests, analysis accuracy, analysis precision or design cycle, we move close to the world's leading automotive research and development levels. Thus, the passive safety performance of vehicles can reach the highest level, and the product development cycle is shortened and development costs are reduced. In other areas of vehicle design, structural analysis using supercomputers is also widely used, such as in sub-frame design, developing new child safety seats, and durability analysis.

In the marine industry, CFD simulation on a supercomputer is an important research and development tool. In modern shipbuilding, CFD technology on supercomputers is used to inspect flow phenomena and principles of the multi-propeller propulsion system, provide details on the performance and mechanism of such a system, improve designers' awareness of performance, and provide a technical consulting service for related development and designs.

In the iron and steel industry, numerical simulation on supercomputers is used to improve equipment and production. For example, a simulation of steel flow, solidification and temperature distribution in molds was used to analyze how different entry nozzles, sections and drawing speed affect the flow field, temperature field and the thickness of solidified matter.

In civil engineering, large-scale numerical simulation on a supercomputer can be helpful for building tunnels, bridges, large-scale building and so on. In the building of the Shanghai Yangzi tunnel, numerical simulation based on the Shanghai supercomputer helped solve problems that were the product of complicated geological and environmental conditions. Vibration analysis was also done on the supercomputer.

As supercomputers have become more and more important in people's lives over the past 10 years, China has made great efforts in their development.

From the 4000A supercomputer (ranked 10th in the world in 2004), Magic Cube (ranked first in Asia in 2008), to Tianhe-1A (first in the world last year), China's domestic supercomputer has made remarkable progress.

Computation capability and capacity has undergone tremendous growth. Up to last month China's supercomputer system number ranked second in the top 500 fastest supercomputers in the world, and total computation capacity ranked 3rd. However, even though supercomputer hardware has developed greatly, software and operations lag behind those of major Western countries. Most computation software packages running on domestic supercomputers are foreign.

In scalability, stability, functionality, acceptability and computational efficiency, domestic parallel computational software lags far behind its foreign counterparts. In operations, China still lacks a long-term plan or project and effective policy to support the development of a supercomputer center, which does nothing to help the long-term development of China's supercomputer.

The author is the manager and chief scientific computing engineer at the department of HPC Application Technology, Shanghai Supercomputer Center.
 
Last edited:

Quickie

Colonel
Well, Chinese CPUs are based on MIPS, and Japanese on SPARK architecture. Neither is indigenous.... Also China's 45nm CMOS technology is licensed from IBM, there is still a long ways to go before China could catch up. Though Interestingly, in the early 1960s, China isn't that much behind the US in terms of semiconductor and computing technology. That's what 10 years cultural revolution can do to you....

The difference here is the Loongson CPU microarchitecture is indigenously designed with own proprietary rights. The reason the cpu is not designed with a completely new software architecture and instead mips was chosen has to do with its existing popular usage.
 

Hyperwarp

Captain
..
China's indigenous chip tech is 65nm, about five years behind the US mainstream of 32nm. Generational gap in chip equipment is not as long as in other tech.

That gap is about to widen next year. Intel will release its 22nm based Ivy Bridge microprocessors (April 8th?). But these are not your regular planar transistors, they are new Tri-Gate Transistors
Please, Log in or Register to view URLs content!
,
Please, Log in or Register to view URLs content!
. Don't forget Intel also invented 'gate-last' as well. TSMC only recently moved to 'gate-last' (IIRC they have to pay royalties to Intel for that). Others like Global Foundries still use 'gate-first'. Intel's lead over others is quite crazy. Look how much trouble GF is having with its 32nm process.

PS:

Something interesting from
Please, Log in or Register to view URLs content!
-

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!
 
Last edited:

Igor

Banned Idiot
»Ø¸´: Supercomputer POWER

10 years ago china didn't have a viable indigenous microprocessor fabrication ability. In 10 years it will reach parity.

It's like calling the country's space program 40 years behind nasa after it lands men on the moon, yet nasa can't even put men into orbit anymore.

It's apples and oranges. There is no 'tech tree' to painstakingly go through. Chinese researchers need not reinvent the well, just quickly catch up by skipping redundancies and then keep up or forge ahead. That is occuring as we speak.

In the realm of supercomputers, the latest chinese interconnect handles 2x more data thoroughput than infiniband, a cpu networking architecture used in the west. The chinese have already leaped ahead in one area regarding supercomputing. Development is not sequential, it is in jumps.
 
Last edited:

CottageLV

Banned Idiot
Like I always said before, the most important factor in the cycle is the ability to make the machine that makes those things. In this case, it's the photolithography machine or the steppper.
 

Hyperwarp

Captain
Now here is something quite unique :
Please, Log in or Register to view URLs content!


More links :
Please, Log in or Register to view URLs content!
,
Please, Log in or Register to view URLs content!


This is full fusion, unlike AMD Brazos/Llano & Intel Sandy Bridge, which are APUs (Accelerated Processing Units). Impressive feat to say the least!

icube.jpg
 
Last edited:
Top