I believe you have already felt that we are in an era of artificial intelligence. If you want to ask what is the most important thing in the era of artificial intelligence? It must be: computing power! computing power! computing power (important things are said three times)
As one of the three major elements that promote the development of artificial intelligence, computing power is known as the "engine" and core driving force of artificial intelligence.
Computing power is very important, but what is computing power?
01
What is hashrate
"Computing power", which literally means computing power (Computational Power), from the ancient abacus to the current supercomputer, are the bearers of computing power.
Now we talk about computing power refers to the ability of computing systems (such as computers, servers, data centers, etc.) to process information and perform calculations. The higher the computing power of the computing system, the faster it can process data and the more complex the tasks it can complete.
If the computing power of a computer is compared to the speed of a student's problem solving, then a computer with strong computing power is like a high school student who solves problems very quickly, and a computer with weak computing power is like a primary school student with a slow problem solving speed.
The same arithmetic problem may only take a minute for high school students to solve, while it may take more time for elementary school students. Therefore, the amount of computing power directly affects the speed and efficiency of computer data processing.
The units of computing power are as follows:
- FLOPS
Floating Point Operations Per Second，浮点运算次数/秒。 表示计算系统每秒钟能执行多少次浮点运算，是最常用的衡量算力的单位。
- IPS
Instructions Per Second，指令/秒。 表示计算系统每秒钟能执行多少条指令，更多地用于衡量CPU的处理速度。
- TOPS
Trillion Operations Per Second，万亿次/秒。 表示处理器每秒钟可进行多少万亿次（10^12）操作。
TOPS/W can also be used as a performance indicator to evaluate computing power, indicating how many trillion operations the processor can perform under the condition of 1W power consumption.
02
The development history of computing power
After years of development, computing power has achieved a huge leap from abacus to cloud, which can be divided into the following stages:
01
The Age of Machines: Abacus and Differential Machines
- In ancient times, human computing power was limited to ten fingers and a few simple counting tools.
- The emergence of the abacus has ushered in the first leap in computing power.
- In 1642, Bryce Pascal invented the Pascal adder, the first real computer in human history, capable of performing addition and subtraction operations.
- At the beginning of the 19th century, Charles Babbage conceived the differential machine capable of complex calculations, which is considered the earliest general-purpose computer concept.
02
The dawn of the electronic age: ENIAC and the microprocessor
- In 1946, the world's first electronic computer, ENIAC, was born, which opened the electronic era of computing power development.
- In 1971, Intel introduced the 4004 microprocessor, marking the dawn of the era of personal computing. The computing power of this chip seems to be insignificant now, but it laid the groundwork for the technological revolution that followed.
03
The Big Bang of Personal Computing: The PC Revolution
- In the 1980s, personal computers (PCs) became popular, and computing power was no longer reserved for large institutions, allowing ordinary people to enjoy the convenience of digital technology.
- In the 1990s, the emergence and popularization of the Internet accelerated the growth of computing demand and promoted the further development of computing power.
04
Computing power in the cloud: Cloud computing and AI
- In the 21st century, mobile computing and cloud computing are booming. With the advent of cloud computing, computing power can "flow" through the network to every corner where it is needed, just like water and electricity.
- The rise of artificial intelligence has put forward higher requirements for computing power, and the emergence of special hardware such as GPU (Graphics Processing Unit) and TPU (Tensor Processing Unit) has greatly improved processing efficiency and provided strong support for the training and inference of machine learning models.
03
The importance of computing power
As mentioned at the beginning of this article, the most important thing in the new era of artificial intelligence is computing power. Computing power is essential to AI, and it affects almost every aspect of the AI field.
Implement complex models
Modern artificial intelligence, especially deep learning techniques, relies on complex neural network models that contain millions or even billions of parameters.
The high computing power makes it possible to train these large models, and only enough computing power can enable the model to complete the learning of massive data within a reasonable time frame.
Shorten your training time
Powerful computing power can greatly shorten the training time of the model. Under low computing power, training a complex deep learning model can take weeks or even months. With the increase in computing power, the same training task can be completed in days, hours, or even less.
Enable real-time analytics and decision-making
In many application scenarios, such as autonomous vehicles, financial transaction analysis, and intelligent medical systems, AI models need to process data and make decisions in real time. Powerful computing power ensures that these systems can quickly process the incoming data and react in a timely manner.
Promote innovation in AI technology
As computing power increases, AI researchers are able to try more experiments and explore new algorithms and model architectures. This not only accelerates the improvement of existing technologies, but may also lead to breakthrough new technologies.
For example, significant advances in natural language processing (NLP) in recent years have been largely due to significant increases in available computing power.
In addition to artificial intelligence, computing power has a significant impact on climate simulation, astrophysics, financial analysis, healthcare, transportation, information security, scientific research, and other aspects.
It can be said that the improvement of computing power is the foundation and guarantee for the realization of these applications, and it is also the cornerstone of the digital world.
04
Conclusion
That's all for today's computing power, knock on the blackboard to summarize the key points~~
It is believed that computing power will continue to play an important role in the future development process. Whether it is in games, artificial intelligence or other fields, the development of computing power will bring us more innovation and possibilities.
Source: ZTE Docs
Original title: Computing Power Series Basics - Computing Power 101: Understanding Computing Power from Scratch
Edit: Fish not me
The reprinted content represents the author's views only
It does not represent the position of the Institute of Physics of the Chinese Academy of Sciences
If you need to reprint, please contact the original official account