Artificial Intelligence Promotes Computing to Smart Computing


   Our reporter Huang Xin

  At the 2021 Artificial Intelligence Computing Conference held a few days ago, the "2021-2022 China Artificial Intelligence Computing Power Development Evaluation Report" jointly released by the International Data Corporation IDC and Inspur Information showed that compared with 2020, artificial intelligence will be used in finance, manufacturing, energy, Industries such as public utilities and transportation have played a particularly significant role in promoting.

At the same time, the computing power infrastructure represented by the intelligent computing center provides public computing power, data and algorithm services to make computing power services easy to use and solve the problem of computing power service supply.

  "In the past four years, we have found that artificial intelligence computing power has received more and more attention, and the applications in this area have become more and more mature. Whether it is the diversification of chips or the computing power and computing capacity of artificial intelligence servers, they have become more important than four years ago. Great improvement.” Zhou Zhengang, Assistant Vice President of IDC Enterprise Research, said in an interview with a reporter from the Economic Daily.

  Zhou Zhengang said that compared with last year, the penetration of artificial intelligence in various industries is increasing, especially in the Internet industry and the financial industry.

In addition, the application of artificial intelligence in the manufacturing, transportation and energy industries has also become more in-depth.

  It is understood that more than 60 countries and regions around the world have issued artificial intelligence policies and issued national artificial intelligence strategies.

IDC predicts that in 2021, the total investment of global enterprises in artificial intelligence software, hardware and services will exceed 85 billion US dollars, and it is expected to increase to 204.5 billion US dollars in 2025, with a five-year compound growth rate of 24.5%.

  However, Wang Endong, academician of the Chinese Academy of Engineering and chief scientist of Inspur, believes that artificial intelligence has also brought about exponential growth in computing power demand, and the computing industry is facing the trend and challenge of diversification, huge quantification, and ecological discretization.

On the one hand, diversified intelligent scenarios require diversified computing power, and computing power has become the top priority for the continued development of artificial intelligence; on the other hand, there is still a huge gap in the transformation from chips to computing power, and the value of multiple computing power is not Not fully released.

How to quickly complete the innovation from multiple chips to computing systems has become a key link in promoting the development of the artificial intelligence industry.

  The demand for computing power in the industrialization of artificial intelligence is soaring. Liu Jun, vice president of Inspur Information, said that the development of algorithm models will also be more complicated. Massive models will be the basis for large-scale innovation. The emergence of massive models such as "Source 1.0", It has become a development trend to build large models and improve artificial intelligence processing performance.

  At present, the world's well-known artificial intelligence companies have invested heavily in huge models. Companies such as Google, Microsoft, Nvidia, Inspur, Zhiyuan Artificial Intelligence Research Institute, Baidu, and Ali have successively launched their own huge models.

"One of the core characteristics of megaquantization is that there are many model parameters and a large amount of training data." Liu Jun introduced the Chinese artificial intelligence massive model "Source 1.0" developed by Inspur Artificial Intelligence Research Institute as an example. The number is as high as 245.7 billion. The data set size reaches 5000GB.

"Our pursuit of computing power has no limits." Liu Jun said.

  Artificial intelligence chips are showing a diversified development trend. The diversification of chips provides an important industrial foundation and richer choices for the acceleration of artificial intelligence industrialization.

However, there is still a huge industrial gap between the creation of chips and their large-scale use.

Take the development of an artificial intelligence server as an example. The entire system needs to go through more than 30 development processes, use more than 150 manufacturing processes, and strictly control the quality of more than 280 key process control points. Issues such as optimization and adaptation of artificial intelligence applications.

  "To release the value of multiple computing power and promote artificial intelligence innovation, one must pay attention to the innovation of intelligent computing systems, increase the construction of new artificial intelligence infrastructure, and design the chain from technology to application, from system structure, chip design, System design, system software, development environment and other fields have formed a situation of clear division of labor and collaborative innovation; the second is to accelerate the construction of open standards, and use unified and standardized standards to transform diversified computing power into schedulable resources. The computing power is easy to use and easy to use." Wang Endong said.

  The scale of China's artificial intelligence infrastructure market has maintained rapid growth, and Chinese server vendors have become the backbone of the global server market.

IDC predicts that the artificial intelligence accelerated server market will reach 5.69 billion U.S. dollars in 2021, an increase of 61.6% compared to 2020. By 2025, the artificial intelligence accelerated server market in China will reach 10.86 billion U.S. dollars.

  my country clearly proposes to deploy the national hub nodes of the computing power network in the country, and at the same time actively promote the construction of intelligent computing centers, and strive to build a universal, inclusive, safe and reliable modern infrastructure system and ecology.

Intelligent computing centers have been regarded by more and more local governments as the key information infrastructure to support and lead the development of digital economy, smart industry, smart city, and smart society, and provide a platform for computing power, data, ecology, and industrial development support.

Our reporter Huang Xin