"In the era of digital economy, computing power has become a new productive force and a new focus of scientific and technological competition in various countries. Ni Guangnan, academician of the Chinese Academy of Engineering, said in his speech at the 2023 Digital Ecology Industry Summit and the 16th Annual Conference of Information Executives, "The computing power industry has the characteristics of facing new fields and high content of innovative technologies, and is a knowledge-intensive emerging industry and future industry. The construction of a collaborative computing power industrial system for computing, storage, and transportation is the driving force for promoting the collaborative innovation of the three key core technologies of advanced computing technology, data storage technology, and network communication technology, the engine for integrating and prospering the three industrial ecosystems to enhance international competitiveness, and the guarantee for accelerating the formation of new quality productivity and building a strong country in science and technology."
Efficient coordination of computing power, storage capacity, and transportation capacity
Storage power is the foundation of digital infrastructure. It is about the security, stability and efficient storage of data, and is the premise that data can be effectively used. The strong storage power can ensure that massive data is effectively managed and utilized, and provide support for subsequent data processing and analysis. Computing power is at the heart of digital infrastructure. It is directly related to the efficiency and quality of data processing, and is the cornerstone of technologies such as artificial intelligence, cloud computing, and big data. With the support of computing power, complex data processing tasks can be completed in a short time, thereby accelerating technological innovation and industrial upgrading. Capacity is the guarantee of digital infrastructure. It is about the fast and secure transmission of data between different places, and is an important link for the development of the digital economy.
At present, the scissors gap between model scale and computing power growth continues to expand, and "storage power" and "computing power" have become bottlenecks restricting the development of China's AI industry. Only the efficient coordination of computing power, storage capacity, and transportation capacity can support the high-quality development of AI computing power centers. Ni Guangnan said: "Storage power, computing power, and transportation capacity together constitute the foundation of the new information infrastructure of the digital economy. For AI intelligent computing centers, it is necessary to balance the allocation of storage capacity, computing power, and transportation capacity, and must pay attention to the proportions and not be out of balance, so as to achieve the greatest economic and social benefits. ”
Promote the development of the computing power industry from four aspects
How to accelerate the technological innovation of the computing power industry system?Ni Guangnan suggested that efforts should be made in the following four aspects.
First, industrial development, standards first. It is recommended to give the appropriate proportion range of "computing power", "storage capacity" and "transportation capacity" through the "industry standard", "group standard" or the issuance of "computing center construction guide".It is proposed that SSD (Solid State Drive Solid State Drive) advanced storage capacity is used in the appropriate proportion range of "storage power";Propose the parameters that "capacity" should have;It is proposed that "computing power" should support the requirements of local AI models. "To avoid the one-sided pursuit of a certain indicator, we should make balanced deployment, balanced development, and seize the commanding heights of a new round of scientific and technological revolution and industrial transformation."
Second, policy guidance and strengthening innovation. SSD should be given priority and should not be evaluated based on a single ** indexSupport domestic independent and controllable network equipment to promote independent innovation of core network technologies. "AI computing power should give priority to supporting local AI platforms and driving the development of domestic related technology platforms, software and hardware. It is necessary to pay attention to the evaluation of technical products and services related to computing, storage and transportation, and the current domestic technology has strong competitiveness, and it should support localization through fair, open and fair evaluation. Each link of the AI computing center must be based on scientific and objective evaluation to achieve independent innovation and ensure the security of the first chain."
Third, talent comes first. The storage, communication and computing industries are all technology-intensive industries, and there is a shortage of talents in this area in China. Taking the storage field as an example, there are more than 60 universities in North America to carry out research, with more than 80 leading professors and more than 500 doctoral teamsThere are more than 20 universities in Western Europe to carry out research, with more than 30 leading professors and more than 200 doctoral teamsAt present, there are only more than 10 universities in China to carry out research, with more than 10 leading professors, more than 100 doctoral teams, and less than 300 master's degrees. Ni Guangnan suggested that colleges and universities should add majors, courses, and laboratories related to computing power, storage capacity, and transportation capacity, expand the scale of talent training, and attract and gather outstanding talents at home and abroad through incentive mechanisms and public services.
Fourth, gather industries and work together to tackle key problems. It is suggested that the computing power, storage capacity, and transportation capacity of the AI center should be incorporated into the chain length system of the industrial chain to better coordinate resources, implement upstream and downstream industry collaboration, and promote scientific and technological innovation and industrial ecological development. "In this way, we can give full play to the advantages of China's new first-class system, the advantages of the super-large-scale market, the advantages of rich application scenarios, and the advantages of talents. The industrial chain of computing power, storage capacity and transportation capacity can also be integrated into a unified AI computing power industry chain, promote the efficient coordination of computing, storage and transportation, form new quality productivity as soon as possible, and enhance the international competitiveness of China's AI computing power center."
In view of the important role played by the computing industry system in accelerating the construction of the digital economy, Ni Guangnan said: "We should plan and deploy the construction of the computing power industry system from a global perspective, through policy guidance, highlight systematic development, formulate a standard system, attach importance to talent training, and prosper the industrial ecology, so as to realize the efficient collaboration of cutting-edge innovative technologies in computing, data storage, and network communication, form a new quality productivity as soon as possible, and provide solid support for the construction of a modern computing power industry system." ”
Author: Yang Guang.
Editor: Gao Shanshan.
Producer: Liu Jing.