Second in the world!, China's Computing Power Level Beijing | Imaging | Global
Not long ago, the 2022-2023 Global Computing Power Index Evaluation Report jointly prepared by International Data Corporation, Inspur Information, and Tsinghua University Global Industry Research Institute was released in Beijing. The evaluation results show that in 2022, the overall server market size in China maintained a positive growth rate of 6.9%, accounting for 25% of the global market share. The compound growth rate from 2017 to 2022 reached 48.8%. From the ranking of the national computing power index, China ranks second in the world in terms of computing power and is among the leading countries.
As a new productive force in the digital economy era, computing power plays an increasingly important role in promoting technological progress, empowering industry digital transformation, and promoting economic and social development. In recent years, China's computing power industry has been continuously developing and its innovation capabilities have been increasing. As of the end of 2022, the total scale of data center racks in use in China has exceeded 6.5 million standard racks, and the total scale of computing power has reached 18 billion times per second.
With the deepening of China's integration of computing power and application development, the empowering effect of computing power is becoming more prominent. On the one hand, computing power is deeply integrated into Internet, communication, manufacturing, scientific research and other industries, providing strong support for intelligent transformation and digital transformation; On the other hand, the new formats and models driven by computing power have become the driving force for the growth of emerging industries.
What are the new development characteristics of China's computing power industry and how will it empower high-quality development in various industries? The reporter conducted an interview.
Multiple industries require computing power support
Gradually expand from the Internet to telecommunications, manufacturing, education and other fields
Not long ago, the cloud based research and intelligent computing platform CFFF was launched at Fudan University. This platform consists of two parts: the intelligent computing cluster "Qiewen-1" for interdisciplinary integration and innovation, and the dedicated high-performance computing cluster "Jinsi-1" for high-precision research.
Based on high-speed data transmission network and cloud computing technology, the "Jinsi No.1" deployed on Fudan campus and the "Chewen No.1" hosted at the Alibaba Cloud Ulanqab data center 1500 kilometers away have connected to form a true "supercomputer" to meet the scientific intelligence research and application needs in different application scenarios.
In recent years, with the increasing abundance of scientific research data and computing resources, the application scenarios of computing power in various fields of scientific research have been expanding.
Zhao Dongyuan, an academician of the CAS Member who has studied mesoporous materials for many years and a professor of the Department of Chemistry of Fudan University, is full of expectations for the successful establishment of the cloud based scientific research intelligent computing platform of Fudan University: "There are a lot of data in experimental science. If we can find the best way to design a material through literature data, it will save a lot of time and our understanding of matter will be more in-depth."
"The intelligent computing platform based on large-scale heterogeneous computing power will become an important support force for future scientific research, promoting the construction of basic and interdisciplinary disciplines, and achieving deep integration and transformation of scientific research achievements." said Jin Li, President of Fudan University.
Nowadays, not only in universities, but also in many industries, computing power support is needed. Many places also consider promoting the improvement of computing power levels and achieving technological collaborative innovation as a strategic choice for the future. Computing power has greatly promoted the development of cutting-edge technologies such as artificial intelligence and industrial innovation, becoming a new engine for activating the potential of data elements and driving the digital transformation of the economy and society.
"Currently, China's digital infrastructure construction is accelerating, and the comprehensive capacity of computing power infrastructure is significantly improving." Nan Xinsheng, Deputy Secretary General of the China Communications Standardization Association, introduced from the three aspects of "storage, computing, and network" that as of the end of 2022, China's total data storage capacity has exceeded 1000EB, and the total computing power has reached 18 billion times per second. The one-way network latency between national hub nodes has been reduced to less than 20 milliseconds.
"Computing power has become a fundamental resource similar to water, electricity, and gas. A tiered optimization and open and diverse computing power supply system is gradually being built, rapidly developing in demand, standards, ecology, and other aspects, supporting the iteration of upper level application development, and providing the source power for industrial innovation," said Nan Xinsheng.
The White Paper on China's Computing Power Development Index released by the Chinese Academy of Information and Communications shows that China's computing power applications have gradually expanded from the Internet to telecommunications, manufacturing, education and other fields. From the perspective of the distribution of computing power applications in various industries, the demand for data processing and model training in the Internet industry continues to increase, accounting for 50% of the overall computing power, which is the industry with the largest demand for computing power; The government industry accounts for 12%, ranking second; Services, telecommunications, finance, manufacturing, education, transportation and other industries rank third to eighth respectively, with manufacturing industries with rich application scenarios having great potential to increase their demand for computing power.
No longer pursuing the accumulation of computing power
Realizing greater value in the collaboration of various digital elements
After receiving the order, the intelligent warehouse will automatically sort the designated goods out of the warehouse, transfer them to the self driving trucks through AGV, and the "green wave belt" signal coordination technology will ensure the priority of the trucks... In the Chongqing Lianglu Guoyuan Port Comprehensive Bonded Zone, the new generation of information technology represented by edge computing will complete the large-scale deployment and create a new mode of intelligent logistics.
"Use roadside sensing devices and edge computing platforms to achieve intelligent monitoring at the end of the car and at the roadside. Distribute Internet of Vehicles data to edge computing nodes through 5G networks. Combine cloud control platforms supported by private clouds to conduct fusion analysis and data distribution, and finally achieve vehicle road collaboration." Li He, director of 5G intelligent transportation solutions of Lenovo Group, introduced.
In addition to port logistics, computing power empowerment is extending to other tracks and fields. The Yellow River Delta National Nature Reserve is an important migration hub, wintering area, and breeding ground for migratory birds, and is therefore known as the "international airport for birds". In recent years, a smart monitoring system has been built here, allowing senior engineer Shan Kai, who has worked in the protected area for more than 20 years, to completely bid farewell to the piles of "field investigation record forms for animal sampling methods".
Spotted duck, identified a total of 2714; The Oriental White Stork has been identified 333 times in total... Rich data is clearly displayed in the intelligent monitoring system. "This system greatly saves our work time and improves the efficiency of bird monitoring data statistics. While reducing the waste of imaging resources such as aerial photography, it also achieves deep processing of real-time and historical data," said Dan Kai.
"At first, experts in the field of wildlife did not recognize computing power, and many wild animals underwent significant changes in appearance from cubs to adults, making tracking and recognition difficult." Wei Biao, Executive Director of Chuangshi Intelligent Technology Co., Ltd., told reporters that after continuous optimization, the system can now more accurately identify birds and better meet the application needs of protected areas.
Experts believe that with the advent of the intelligent era, computing power has become ubiquitous and ubiquitous. Both the "end" in the hands of users and the widely covered "cloud" can generate computing power.
"With the exponential growth of data volume, the foundation of computing power will continue to be strengthened, which puts higher requirements on the development of computing power. The infrastructure of the digital economy and the core driving force of general artificial intelligence are the two major characteristics of future computing power." Chen Zhenkuan, Vice President of Lenovo Group, said that from the perspective of digital transformation needs in various industries, "computing power" cannot be simply stacked, but needs to develop towards "intelligence". More widely applied and smarter computing power will eventually become a basic resource like water and electricity, realizing greater value in the synergy of various digital elements.
The development of computing power drives the transformation of service models
Diversified application scenarios require diverse computing power support
In terms of classification, current computing power can be mainly divided into three types: basic computing power, intelligent computing power, and high-performance computing power. Chen Zhenkuan explained, "Basic computing power is mainly based on servers with CPU chips, intelligent computing power is mainly based on accelerated computing platforms with GPUs and other chips, and high-performance computing power is mainly based on the fusion of CPU chips and GPU chips to create computing clusters, which are respectively targeted at application scenarios such as basic general-purpose computing, artificial intelligence computing, and scientific engineering computing."
Diversified application scenarios require the support of diverse computing power. The demand for massive and complex data processing is driving a significant increase in computing power, and intelligent upgrading of computing power has become a trend. According to the calculations of the China Academy of Information and Communications Technology, intelligent computing power has become the main driving force for the growth of computing power in China. Duan Xiaodong, Vice President of China Mobile Research Institute, believes that the accelerated integration of the computing industry into the real economy is facing scenario demands and experiential challenges from various industries. Both from the perspective of the development of the computing industry itself and the need to empower industrial upgrading, it is necessary to promote the development of diverse computing power.
Zheng Weimin, an academician of the CAE Member and a professor of Tsinghua University, believes that the key to the scale deployment of diversified computing power lies in two aspects: first, do a good job in software and hardware collaboration and integration innovation from the supply side to release the potential of diversified computing power; The second is to lead the construction of a consistency evaluation system by the demand side, driving rapid product iteration and guiding users to use diverse computing power products boldly and extensively.
The development of computing power is also driving the transformation of service models. In April this year, the Ministry of Science and Technology launched the national supercomputing Internet deployment, and realized the overall scheduling of computing resources by building a market-oriented operation and service system. In the view of Zhang Yongjian, CEO of Broad Data, computing resource services will become the focus of future data center development. The transition from the traditional model of selling cabinets to a new model of providing computing power services is beneficial for the computing power industry to reach more users and expand into a broader market.