"Shenwei Taihu Light" supercomputer. Visual China Information
He Liang
Recently, China's self-developed supercomputer "Shenwei Taihu Light" once again won the World Supercomputer Championship at the 2017 International High Performance Computing Conference, winning the first three consecutive championships of the Chinese domestic supercomputer system in the world computer championship. , coupled with the glorious history of the previous Tianhe No. 2 winning six times, China's super-calculator is very dazzling on the world stage.
However, the super-calculator under the spotlight has a short board in its application although it has "fast" performance and good performance.
"Boating out to the sea" is not always a winning strategy
Prior to the birth of “The Light of Shenwei·Taihu Lakeâ€, China’s supercomputer Tianhe 2 was once the “overlord†of high-performance computing in the world. It has won the world's Top 500 six times and is widely used in research fields such as weather, climate, biomedicine, and new materials. But it has also suffered "criticism", the Tianhe series of processors from the United States - Tianhe No. 1 is Intel's CPU plus Nvidia's GPU; Tianhe No. 2 is Intel's GPU plus Intel's many-core processor.
“The matured chip technology and the rapid development of its role have been regarded as a development strategy, but the autonomous and controllable technology cannot be realized.†Dr. Gan Lin from the Department of Computer Science at Tsinghua University said in an interview with a reporter from Science and Technology Daily on July 28th. A series of measures announced by the United States on the embargo imposed on China's core processors also led directly to the obstruction of development in the Tianhe system.
“Tianhe 2 surpasses the US Titan and other supercomputers through 'borrowing to the sea', but in key areas such as national defense security, the country has made clear requirements for autonomous control.†Gan Lin said that key technologies are subject to humans, although they can "Long-term skills," but at the application level greatly reduced.
China Core Lets It Count on Jedi to "Counterattack"
Why does China Super Calculate get so excited after the “Light of Shenwei·Taihu†reaches the world? The reason behind this is that in the face of the unilateral “sanctions†banned by the United States, China Supercomputer has completely adopted the self-developed Chinese core, the “Shenwei 26010†heterogeneous heteronuclear processor, to achieve the Jedi “counterattack†and the public’s understanding "Completely winning with speed" translates into "a complete breakthrough in terms of autonomous controllability, peak speed, continuous performance, green indicators, and applications."
The National Institute of Computational Fluid Dynamics (CCM) completed the massively parallel computing of the multi-sets of Tiangong-1 aircraft's fall flight status using the “Aerospace Aircraft Unified Algorithm Numerical Simulation†software developed based on the “Light of Shenwei·Taihu Lakeâ€. Using 16384 processors (1/8 of the entire Taihu light machine) completed the conventional 12-month calculation task in 20 days, and the calculated results are in good agreement with the wind tunnel experiment results. Moreover, the materials science closely related to the aerospace field also needs to simulate what elements are added to achieve the most effect.
Now, the peak computing performance of Taihu Lake's light reaches 12.54 billion times per second. This means that the computational power of 1 minute of Taihu Lake's light is equivalent to the global population of 7.2 billion people who has used the calculator continuously for 32 years. Obviously, China’s supercomputer hardware capabilities are not “without a name,†then what is the power of such superpowers?
"Intermediate Interruption" Limits Application Amplification
In 2016, an application project “running†in the light of Taihu Lake won the “Gordon Bell†award, achieving a breakthrough of 29 years in the world of high-performance applications in China.
The project "Ten Thousand Nuclear Extensible Global Atmospheric Dynamics Full Implicit Simulation" is the use of more than 10 million cores to complete a numerical simulation of the climate. Its special feature is the algorithm - a new "full implicit" calculation method. The computational efficiency of this algorithm is improved by nearly one order of magnitude compared to the highest level of the current international problem. Yang Chao, head of the award-winning team at “Gordon Bell†and a researcher at the Institute of Software of the Chinese Academy of Sciences, stated that algorithms and performance are the main reasons for winning the team.
The implementation of high-performance computing to achieve the application is not a single supercomputer can complete. It is more like a chain, with software from the middle tier connecting upstream application scenarios and downstream machine hardware. Professor Yang Guangwen, a professor at Tsinghua University and director of the National Supercomputing Wuxi Center, said in an interview with a reporter from Science and Technology Daily that people engaged in computer research cannot design better algorithms and application software because they are unfamiliar with the application. However, software application users lack computers because they are not familiar with applications. The ability to produce high-quality software... Even if the technology is mature and the demand is strong, as long as the software is lacking, the ability of super-calculation cannot be applied.
Moreover, the bigger problem China faces is that not only software is backward, but also there is a certain gap between software ecology and foreign countries. For example, Gan Lin said: The prosperous processor of Taihu Lake is equivalent to 10 million ants in coordination and pulling heavy objects at the same time. If an ant “sacrificesâ€, the ant’s task load will need to be allocated to other software. Ants work to ensure that calculations are conducted properly. This seemingly simple step requires common software to control and form a mechanism for fault tolerance and error correction. Although China also has, it is not as mature as foreign software. Most of the programs that run on the light of Taihu Lake have previously required code modifications to optimize the hardware characteristics of Taihu Lake's light. The tuning process depends on software tools.
“Although we hold the 'Gordon Bell' award, we still have a long way to go to fully improve our country’s super-computing application level.†Yang Guangwen made a very direct statement. This is similar to that just after leaving the laboratory. The efficacy is predictable and the implementation remains to be tested. Yang Guangwen believes that in order to do a good job of super-calculation, researchers need to have “artisan spiritâ€.
From hardware to software, China has been catching up and constantly making efforts to chase and run from the running. At the national level, several projects in the national major scientific and technological special projects are being used in the field of funding under the support of super-computing, and investment has continued to increase. Yang Guangwen said that the system's policies, projects, and talent support will prompt the development of ultra-numerical R&D and nurturing industries and promote the national industrial transformation and upgrading.
There is no competitive relationship between quantum computing and hypercomputing
At the "2017 International Big Data and E-class Computing Symposium", Yang Guangwen once disclosed that the development of the new generation of tens of billions of billions of supercomputers in the "Light of Shenwei and Taihu Lake" has been included in the national "13th Five-Year Plan" and is currently high. The performance calculation focuses on the development of the three prototypes of “Shenwei,†“Tianhe,†and “Dawnâ€. It is expected that the first domestically-produced 100 billion-scale supercomputer will be launched around 2020.
From the scale of billions to billions of billions of dollars, we face enormous challenges. Yang Guangwen told the reporter of the Science and Technology Daily that the increase in computing power has led to increasing power consumption. Adopting low-power chip design, efficient cooling technology, and optimized management of the entire system management are problems that are currently being tackled.
Compared with emerging quantum computers, Ms. Lin believes that the next generation of supercomputers does not compete with them for decades. Compared to the traditional computer, 1 bit can only be represented by two types of 0 or 1. The quantum computer relies on the uncertainty of quantum to realize the polymorphic expression of the quantum representation. For example, the qubit may have multiple states, which may make the quantum Computers are much more powerful than traditional computers at the same scale, but the uncertainty of molding time adds more uncertainty to this possibility.
"To be mature, quantum computers must have a mature algorithm, and this algorithm must be able to simulate the application and match the quantum uncertainty in order to produce the desired results." said Gan Lin, which also leads to only specific application areas. It is suitable for quantum computers.
"In the next generation of super-computing development and competition, China must not only win the battle of speed, but more importantly win the battle of application!" Yang Guangwen said that the lag in domestic application software and the lack of talented people have seriously hampered the development of super-computing. . With the ultimate goal of “applicationâ€, the “Light of Shenwei·Taihu Lake†team will further organize domestic and foreign strengths to jointly create a domestic chip eco-environment and strengthen application software development, especially in weather and climate services, advanced manufacturing, and life and health. Major application areas such as big data analysis do their homework, break through the application bottlenecks, and make up for the shortcomings of Chinese supercomputer applications.
Meet safety requirement such as GB31241, UL1642, IEC62133 etc; Targeted R&D, support customization. high safety battery, rechargable battery, li battery, quick charge battery, high capacity battery, ultra thin battery, steel cylilndrical battery.
A lithium polymer battery, or more correctly lithium-ion polymer battery is a rechargeable battery of lithium-ion technology using a polymer electrolyte instead of a liquid electrolyte. High conductivity semisolid (gel) polymers form this electrolyte.
12V Lipo Battery,Lithium Ion Polymer Battery,Lithium Polymer Rechargeable Battery,1800Mah Lithium Polymer Batteries
Shenzhen Glida Electronics Co., Ltd. , https://www.szglida.com