What is computing power?
This is a world you are familiar with, but viewed from a different perspective, it is a new world composed of numbers. The process of processing these numbers is computation, and the capability that computation possesses is computing power. Therefore, what essentially powers this digital world is not human resources, material resources, or financial resources, but computing power.
Today, we’ve invited a senior computing power engineer from Yuanjie to explain what computing power is, its origins, leading technologies, and application areas, as well as how it can transform the world. Join us for a comprehensive and in-depth exploration of the fascinating world of computing power.
Although computers have already demonstrated astonishing computational capabilities, traditional computers are far from sufficient to meet the demands of highly complex problems in scientific research and artificial intelligence. However, when countless computers are connected to perform collaborative computing, they can unleash even more powerful computing power—this is the supercomputer, or "supercomputing" for short. Through the limitless possibilities of supercomputing, scientists and engineers can achieve higher computational capabilities and accuracy in fields such as weather forecasting, genetic research, physical simulations, and large-scale data analysis.As an outstanding achievement in science and technology, supercomputers represent a significant step forward for human progress across various fields.
Computing Power: The Force That Changes the WorldChina Claims the Supercomputing Throne
China’s supercomputer development has shown astonishing exponential growth over the past few decades. However, as supercomputers’ processing speeds gradually reached the exascale level—measured in exaflops—it became clear that simply increasing the number of CPUs was no longer sufficient to further boost computational speed. To address this challenge, Chinese scientists introduced a unique configuration: combining GPUs and CPUs into a heterogeneous hybrid system that works in tandem.
Against this backdrop, China’s Tianhe-1 supercomputer was born. Harnessing the powerful computing capabilities of 6,144 CPUs and 5,120 GPUs, it pushed computational speed to the exascale level. In 2010, it claimed the top spot on the global supercomputer rankings for the first time, making China the second country in the world—after the United States—to successfully develop an exascale supercomputer system.
The success of the Tianhe-1 supercomputer marks a significant breakthrough and a demonstration of China’s growing strength in the field of supercomputing. This achievement is not only of great significance in the realm of scientific research but also provides robust support for China’s advancements in artificial intelligence, weather forecasting, energy research, and drug development. The rapid development of China’s supercomputers underscores the nation’s leadership and technological innovation capabilities in the field of high-performance computing.
Behind this miracle lies a high cost. For instance, training the large AI model GPT-3 requires 1 million home computers running continuously for ten years; the electricity consumed for a single training session is equivalent to the annual consumption of 1,500 people. Can ordinary people access this powerful yet expensive computing power?
Fortunately, with advancements in computing technology, these seemingly powerful and expensive computing resources are gradually becoming accessible. On one hand, the rise of cloud computing technology has enabled the centralized management and allocation of vast computing resources; users can rent computing resources through cloud platforms, significantly reducing usage costs. On the other hand, the development and widespread adoption of hardware such as AI chips have continuously improved AI computing capabilities while gradually lowering prices, making them affordable for more people.
What exactly is a GPU?
As mentioned earlier, a GPU stands for Graphics Processing Unit. It is a processor specifically designed to handle graphics and image data, widely used in computer graphics, computer games, and other fields requiring image and graphics processing.
Unlike the Central Processing Unit (CPU), the design and architecture of a GPU are more focused on parallel computing and graphics processing. It employs a large number of stream processors and parallel computing units, enabling it to execute multiple threads or computational tasks simultaneously to accelerate the processing of graphics and image data. Through this parallel processing approach, GPUs can efficiently handle complex image rendering, graphics computation, and image processing operations.
In addition to its applications in computer graphics, the GPU is also widely used in scientific computing, machine learning, deep learning, and artificial intelligence. Due to its parallel processing capabilities and powerful computational performance, the GPU can accelerate complex computational tasks in these fields and improve computational efficiency. Consequently, an increasing number of researchers and developers are utilizing GPUs to accelerate various computationally intensive applications.
It is worth noting that as the demand for GPUs continues to grow, their design and functionality are constantly evolving and improving. Modern GPUs have become key components in high-performance computing and graphics processing, providing powerful computational capabilities and graphics processing performance for applications across various fields.
What has driven the rapid development of computing power?
The Rise of Big Data: With the widespread adoption of the internet and sensor technology, vast amounts of data are being generated and collected, requiring powerful computing capabilities to process and analyze these massive datasets. The rise of big data has driven the demand for computing power and spurred the development of more efficient and powerful computing devices and computing networks.
The Rise of Artificial Intelligence: The development of AI technology requires robust computing power. The training and inference processes of AI algorithms, such as deep learning and machine learning, demand substantial computational resources. Therefore, to accomplish more complex AI tasks—including image and speech recognition, as well as natural language processing—enhancing computing power is indispensable.
Computing Power Has Now Acquired "Energy" Value
In today’s era of explosive growth in computing power demand, even a powerhouse like cloud computing requires a comprehensive upgrade—specifically, the “computing power network,” a brand-new concept proposed by China Mobile in 2021. Just as the power grid can flexibly allocate different types of electricity, the computing power network can aggregate diverse computing resources through channels such as 5G and fiber optics. Supercomputing, intelligent computing, general-purpose computing, and even quantum computing can all be integrated into a single network, ready to be summoned.
Against the backdrop of today’s explosive growth in computing power demand, cloud computing—as a powerful computing capability—also requires a comprehensive upgrade. However, in 2021, China Mobile introduced a brand-new concept—the Computing Power Network—bringing revolutionary changes to cloud computing.
Just as the power grid can flexibly allocate different types of electricity, the computing power network leverages communication technologies such as 5G and fiber optics to aggregate diverse computing resources. Various forms of computing power—including supercomputing, intelligent computing, general-purpose computing, and even quantum computing—can be integrated into a unified network, ready to be called upon and utilized.
When we take a broader view, we discover even more similarities between computing power and electricity. Both have a massive impact on the economy. According to statistical data, a one-point increase in the computing power index can drive 3.6‰ growth in the digital economy and 1.7‰ growth in GDP.
However, much like energy, computing power is distributed extremely unevenly. The eastern regions are economically developed and densely populated, but land and electricity costs are correspondingly high, leading to a shortage of computing power. The western regions are the exact opposite: while demand for computing power is low, conditions such as land, electricity, and climate are highly suitable. If certain computing tasks could be migrated to the western regions, they could alleviate the computing pressure on the east through low-cost, low-energy-consumption, and low-emission methods.
However, much like energy, computing resources are distributed extremely unevenly.The eastern regions are economically developed and densely populated, but the high costs of land and electricity have led to a shortage of computing power. In contrast, the western regions have lower demand for computing power, yet their conditions—including land, electricity, and climate—are highly suitable. By shifting some computing tasks to the western regions, we can alleviate the computing pressure in the east through low-cost, low-energy-consumption, and low-emission methods. Just as you might watch a TV drama produced in Hengdian, Zhejiang, its rendering might actually be completed in a city in the west.
China’s Computing Power Strategy: East Data, West Computing
The digital world urgently needs a mega-project comparable to the South-to-North Water Diversion Project and the West-to-East Power Transmission Project—and that is "East Data, West Computing."Officially launched in 2022, the “East Data, West Computing” initiative marks the practical implementation of China’s computing power network. The project aims to transform the entire network into a massive supercomputer by integrating ten national-level data center clusters and eight computing hubs, leveraging infrastructure built on an all-fiber backbone network to provide lower-cost computing services to all industries.
In the fiercely competitive global computing power arena, the Chinese government is actively striving to transition from catching up to keeping pace, and even taking the lead. In the future, China will further increase investment, strengthen talent development, and promote the growth of the computing power industry. We are committed to gaining an advantage in global computing power competition and maintaining our leading position.
Key Technologies for Computing Power Development:
Key technologies for computing power development span multiple areas; the following are a few examples:
1. Integrated Circuit Technology: Integrated circuits are the core of computing devices. Composed of billions of transistors, they enable high levels of integration and scalability in computing power. As integrated circuit technology advances, chips are becoming smaller in size yet more powerful in performance, thereby driving rapid improvements in computing capabilities.
2. Parallel Computing: Parallel computing is a vital technology that enhances computing power by utilizing multiple processing units to perform calculations simultaneously. It allows multiple computational tasks to be executed concurrently or in an overlapping manner, thereby accelerating the computational process. Through parallel computing, large-scale problems can be divided into multiple sub-problems and processed in parallel across multiple processing units, improving computational efficiency.
3. High-Performance Computing (HPC) Technology: High-performance computing is a computing technology optimized for large-scale computation and data processing problems. HPC technology utilizes methods such as parallel computing, task scheduling, and data transfer to improve computational speed and efficiency. High-performance computing systems typically consist of multiple processing nodes, high-speed interconnect networks, and high-capacity storage devices to provide exceptional computational performance and storage capabilities.
4. Quantum Computing: Quantum computing is a computational technology based on the principles of quantum mechanics, utilizing the parallelism and superposition of quantum bits (qubits) to perform computations. It is characterized by the ability to perform large-scale parallel computations and solve certain problems with exponential speedup. Although quantum computing is still in its early stages of development, it has already demonstrated immense potential for specific problems.
5. Artificial Intelligence (AI) Technology: Artificial intelligence relies on powerful computing power for model training and inference. AI technologies such as deep learning and machine learning require substantial computational resources for model training and optimization, thereby enabling the learning and decision-making processes for complex data. Through technologies such as high-performance computing and parallel computing, sufficient computing power can be provided to support the development of artificial intelligence.
In addition to the key technologies listed above, many other technologies also play a crucial role in the advancement of computing power, such as high-speed communication, efficient data storage and management, optimized algorithms, and computational models. The continuous innovation and development of these technologies not only drive improvements in the performance of computing devices but also spur rapid advancements in computing power across various fields.
Application Areas of Computing Power
Computing power has a wide range of applications, primarily encompassing the following areas:
1. Scientific Research: Computing power plays a vital role in the field of scientific research. Through technologies such as high-performance computing, data analysis, and simulation modeling, scientists can leverage powerful computing resources to solve complex scientific problems in areas such as astrophysics, climate modeling, quantum chemistry, and bioinformatics.
2. Artificial Intelligence: Computing power has played a decisive role in the development of artificial intelligence. AI technologies rely on the ability to process massive amounts of data and execute complex algorithms. With powerful computing capabilities, AI can be applied to fields such as image recognition, speech recognition, natural language processing, and data mining, leading to numerous applications such as smart assistants, autonomous driving, and intelligent recommendation systems.
3. Healthcare: Computing power has extensive applications in the healthcare sector. By leveraging high-performance computing and big data analytics, it can accelerate genomics research, improve the accuracy of disease diagnosis and treatment, and advance the development of personalized medicine. Additionally, computing power supports medical imaging diagnostics, drug discovery, and disease prediction.
4. Finance and Business: In the finance and business sectors, the application of computing power is both extensive and critical. High-performance computing and big data analytics capabilities can be applied to financial risk management, quantitative investing, market analysis and forecasting, and supply chain management, providing precise data modeling and decision support.
5. Energy and Environmental Research: Computing power also plays a critical role in energy and environmental research. Through simulation and model analysis, high-performance computing is used to optimize energy systems, improve energy efficiency, and address issues related to environmental protection and sustainable development.
6. Manufacturing and Engineering: In the manufacturing and engineering sectors, the application of computing power can enhance production efficiency and quality. Through simulation and optimization algorithms, the design process can be accelerated, and product performance and reliability improved. Additionally, the application of computing power in technologies such as digital twins, virtual reality, and augmented reality has brought significant innovation and benefits to the manufacturing and engineering sectors.
In addition to the fields listed above, computing power is also applied in education, entertainment, transportation, security, and many other sectors. Its widespread application and continuous innovation have brought humanity a more convenient, efficient, and intelligent lifestyle.
China’s Current Computing Power Challenges:
1. Technological Dependency and Innovation Capacity: Although China has achieved significant accomplishments in manufacturing and hardware production, there remain relative weaknesses in the critical areas of chips and high-performance computing. China’s computing power development still relies on imported chips and lacks independent innovation capabilities. In particular, recent U.S. export controls on chips to China have significantly constrained the independent development of China’s computing power industry.
2. Distribution and Utilization of Computing Resources: Although China possesses a vast population and massive data resources, the distribution and utilization of computing resources are uneven. While some advanced cities and research institutions have large-scale computing clusters and high-performance computing equipment, computing capacity remains severely insufficient in less developed regions. This leads to an imbalance in resource utilization and may result in a “digital divide” in computing development.
3. Energy Consumption and Environmental Pressure: As computing power increases, the demand for energy is experiencing explosive growth. High-performance computing equipment and large-scale data centers consume vast amounts of electricity, placing enormous pressure on energy supply and the environment. At the same time, energy consumption exposes China’s computing industry to issues such as carbon emissions and environmental pollution, necessitating the search for low-energy, high-efficiency solutions.
4. Data Privacy and Security: As computing power advances, public concerns regarding data privacy and security are growing. Computing technologies require access to vast amounts of sensitive information—such as personal identities and financial data—during data processing and analysis. Protecting data privacy and security has become an urgent priority, necessitating the establishment of stricter regulations and security standards, alongside enhanced technical capabilities to prevent data breaches and cyberattacks.
5. Talent Development and Innovative Research: The advancement of high-level computing power requires substantial support from highly qualified talent and the drive of innovative research. Although China possesses a vast talent pool in the fields of artificial intelligence and computing, there remains a shortage of specialized professionals in areas such as computing infrastructure, chip design, and optimization. Cultivating and attracting talent, as well as enhancing innovative research capabilities, are critical tasks for the future development of China’s computing power industry.
In summary, China faces challenges in the computing power sector, including technological dependence, resource distribution, energy consumption, data security, and talent development and innovation. Addressing these challenges requires joint efforts from the government, enterprises, and research institutions to strengthen cooperation, drive technological innovation, and enhance independent R&D capabilities, thereby ensuring China’s long-term sustainable development in the computing power sector.
What Can Computing Power Bring to the World?
In this new world built on data, computing power has become the core force driving its operation. The future of this digital world will profoundly impact our real lives through technologies such as artificial intelligence, big data, and cloud computing.
First, the development of computing power will significantly drive technological innovation. In the field of artificial intelligence, through technologies such as deep learning and neural networks, computers can continuously self-learn and optimize, thereby achieving remarkable results in areas such as image recognition, speech recognition, and natural language processing. These technologies will be widely applied in fields such as autonomous driving, smart homes, and medical diagnostics, greatly improving production efficiency and reducing living costs.
Second, big data technology enables us to extract valuable insights from massive amounts of data, thereby allowing us to make predictions about the future. For example, in sectors such as finance, meteorology, and transportation, collecting and analyzing data can effectively help prevent risks, facilitate sound decision-making, and improve the efficiency of resource allocation. At the same time, big data technology also offers vast opportunities for development in areas like business intelligence and personalized recommendations.
Furthermore, cloud computing technology facilitates the sharing and optimized allocation of computing resources, providing convenient, low-cost computing services to both businesses and individuals. Building on this foundation, emerging technologies such as edge computing and container technology will further drive improvements in computing power and the efficient utilization of resources.
Moreover, the enhancement of computing power will accelerate the development of emerging technologies such as the Internet of Things (IoT) and 5G. In the future, a world where everything is interconnected will be filled with sensors and smart devices. Through real-time data transmission and analysis, functions such as remote control and intelligent scheduling will be realized, significantly improving people’s quality of life.
In this process, the Chinese government has placed high priority on the development of computing power, introducing a series of policy measures to promote infrastructure construction, talent development, and industrial growth. In the fiercely competitive global race for computing power, China is striving to transition from a follower to a co-leader, and ultimately to a leader.
The continuous improvement of computing power will lead us into a brand-new digital world, transforming the way humans produce and live. In this process, we must keep pace with technological advancements, fully tap into the potential of the digital world, and contribute to building a better future.
Therefore, when examining the transformations brought about by computing power, we can see that it has not only changed the world but has even created an entirely new world composed of digital elements.