Electronics

With the iterative upgrade of hardware, compared with the hardware itself, the "core competitiveness" of manufacturing equipment is now more and more reflected in the software. The problem of "stuck necks" of core components has not been solved yet, and the "stuck necks" of industrial software "The problem comes again. And it's far more serious than imagined.

Earlier, a UP host at station B, GeekLead, the captain of the geeks, wrote in "Huawei and Other Made in China, How Can the U.S. Rely on Software to Get Neck 》In this submission video, China’s shortcomings in the field of industrial software are analyzed. Compared with the core components of the “stuck neck”, the degree of restriction of industrial software is comparable to that of “stuck heads”.

Why can Chinese rockets go to the sky, but the chip will be stuck?

This is because the United States controls the foundation of the manufacturing industry. This scene is comparable to a draw from the bottom of the pan.

If it is said that domestic companies can make do with some scenarios in the production of chips, but China is far behind in terms of software for designing chips. The key software is the upstream of production, which means that the United States has not only stuck China's neck, but also stuck its head.

Industrial software is one of the biggest shortcomings in China's entire industrial system. Many people have a cognitive error about Chinese software: Many large Internet companies make software. How can Chinese software be so bad?

Here, the geek captain quoted a PPT from Academician Ni Guangnan:

It can be seen from the PPT: Although China's Internet applications are very powerful, China's large-scale industrial software and chip design software (EDA design tools) are the weakest links. As an engineer, Captain Geek has spent time in the Internet and industrial software circles. His tutor during his study abroad was an expert in the industrial software field. He also had an internship in the Siemens Industrial Software R&D Department in North America at that time. It is this internship experience that allows the geek captain to realize the difficulty of the industry and the gap in China. This popular science is also very important for public awareness.

How industrial software is stuck in the "Chinese neck"

First of all, what exactly is industrial software?

If a worker wants to do well, he must first sharpen his tools. Industrial software is actually a tool in the industrial field, like office, photoshop, etc.

Industrial software can be roughly divided into three categories: those that help design, those that help simulation, and those that help manufacture. The English abbreviation starts with CA (Computer Aided):

Design software-CAD, a simple understanding is a drawing tool. For example, to build a building or to make parts or chips, you need to draw drawings first. In the early days, they were drawn by hand. Now they are operated by a computer, and standard parts are directly combined by dragging. It can be done.

The simulation software-CAE, is a test in terms of use, that is, it helps to check whether the involved is appropriate and reasonable.

Manufacturing software-CAM. Since production depends on equipment, operating equipment depends on CAM. For example, in machine tool production, it is necessary to rely on software to control how the cutter head moves and how to cut.

Of course, these concepts have now been integrated and become a big concept, providing a complete set of services-PLM (product lifecycle management), or product lifecycle management, and the electronic field has become EDA (electronic design automation software).

At present, whether you are PLM or EDA, China is completely "stuck" by foreign countries, especially the United States.

Because software is a tool for all links, nothing can be done without industrial software, and everything is stuck if the software is stuck. Because engineers cannot draw hundreds of millions of transistors by hand. Even in the 1970s, EDA was born in Silicon Valley. At that time, Intel and Hewlett-Packard had already explored within the enterprise, and then formed separate industries-Synopsys and Cadence, and then started their business in Silicon Valley. Now they are already in the industry. Monopoly giants exist.

Why China's industrial software is not developing well

The first reason is difficult.

Many people think that industrial software belongs to the computer profession. In fact, industrial software seems to be software, but behind it is the crystallization of countless knowledge, which is an absolute interdisciplinary field. It requires you to have no shortcomings in various fields, and it is absolutely useless to have only computer professional participation.

First of all, it involves basic subjects. Here, the geek captain focused on sharing his internship experience.

At that time, the company where the geek captain was intern was the famous Siemens NX. The Chinese translation of the internship department was the animation group, and the customer was Tesla. At that time, they were developing model 3, and the shape was not finalized yet. Because the customer's design is very novel, there are huge glass or electric cars, there are many needs, and it is indeed the first at that time. Tesla gave several versions of the design, and needed to let the computer simulate the car manufacturing process to see if the design was reasonable.

At first, the geek captain thought how difficult it is to make an animation, but after the start of construction, he was dumbfounded. Although it was an animation, it was not an animation like an animation. It only needed to be able to move. The requirements of the industry are very strict and precise. The geek captain must study mechanics every day to simulate the movement process of thousands of parts and calculate the spatial displacement of each part. Moreover, different parts have a certain force directly with each other. Basic knowledge alone is not enough. of.

It can be said that if industrial software programmers do not understand dynamics, they cannot start to work at all. All animation simulations must be based on physical formulas. Industrial production is not a trifling matter, and it is not good-looking. It must be able to know the actual production.

From this, the geek captain understood: Industrial software essentially writes physical formulas into the software, allowing the software to quickly implement calculations. And if the required physics formula does not exist, then the formula must be derived first, which requires the study of the basic discipline of physics, and then the ability to research the underlying computer. Because some physics formulas are too complicated, although the theory is correct, they cannot be calculated. It is up to computer scientists to make calculations possible.

Here, the geek captain enumerated a typical application: CFD (computational fluid dynamics). The effect of calculating fluids (such as air, water, etc.) on products is one of the important applications of CFD. This type of simulation is used in the design of aircraft wings and car shapes. All have to be done, and now more areas have been designed, such as the heat dissipation of the chip.

In fact, the physical equations related to CFD were proposed as early as the 19th century, but it is very difficult to solve without a computer, so it basically relied on experiments before, such as building a wind tunnel on the wing to prevent the wing, and then using a fan to simulate the airflow. Re-use sensors to measure data, this process is costly and time-consuming.

After an IBM computer solved this equation, it has since become possible to use a computer to solve the physical equation. Under the influence of the continuous iteration of technology, CFD has become more and more powerful, and both the amount of calculation and the speed of calculation continue to rise.

In the field of CFD, China can be said to be squeezed to death by the United States. In 2019, the US Huawei ban has just begun, and the US Ansys has been banned. As the industry's crown-level industrial software, TSMC and MediaTek are all Ansys customers. , And its business card-level product Fluent in the CFD field is banned.

Professor Philip.L.Roe, a leader in the field of CFD at the University of Michigan in the United States, commented that: Now CFD analysis has completely replaced experiments to design industrial products. If the United States and the European Union are serious about implementing relevant policies, China's car and aircraft design will have to go back to blow holes.

At present, the industrial simulation software is mainly controlled by the United States, Germany and France, such as Ansys of the United States, Siemens of Germany, and Dassault of France. The key is that the European Union and the United States cannot separate their relations. Ansys has acquired many European companies. To put it bluntly, in the field of industrial software, the United States still controls the audience.

The second reason is not enough money.

To be honest, you can’t rely on love to generate electricity for anything. When the geek captain was doing an internship at Siemens, in a very small city, the prices were very low. During a summer vacation for two to three months, the after-tax salary was 15,000 knives, almost more than 100,000 yuan. . Compared with the same period last year, many domestic industrial software friends, the wage gap of the geek captain is very wide.

There are market reasons for low wages, because the head effect of the industrial software industry is very obvious. In order not to affect their own development cycle, customers will only pay for the most advanced and reliable software. However, domestic software has no advantage to high-end customers under fair competition. Why should customers bet on their own future to give warmth to domestic software? Foreign goods are cheap and cheap, so it’s not easy to use them directly?

Moreover, foreign industrial software has actually penetrated into the entire training process of Chinese engineers. At present, most of the engineering majors in Chinese universities have to learn foreign software as soon as they enter the university. The first 3D design software they learn is SolidWorks, and its parent company is Dassault, France. To control the university is to control the engineers of China's future, and to stifle Chinese talents from the source.

The third reason is policy.

The U.S. Department of Defense took the lead in launching the Electronic Renaissance Program (ERI). The "renaissance" here is not due to backwardness, but the United States found that its leading advantage is not obvious enough, and it must rely on policies to further expand its advantages. ERI plans to invest US$1.5 billion in 5 years, and EDA software giant Cadence has received US$24 million in funding for this plan. And this is just one of the various support policies of the United States in recent years.

In China, statistics show that during the 15th and 11th Five-Year Plan, only about 50 million yuan was invested in various policy investments in related fields, and no more than 200 million yuan during the 12th Five-Year Plan period. The gap is evident. In other words, the intensity of a support policy in the United States is equal to 50 times that of China's investment in 15 years. There is another reason for this, that is, the position of industrial software is very embarrassing. If it belongs to software, it belongs to industrial applications according to the Chinese classification; but as mentioned before, the most important thing about industrial software is interdisciplinary basic research, not just industrial Can be covered. The typical Chinese moment has appeared again. According to experience, multi-content integration, the result of both ends is to play the ball, regardless of both ends.

And there is another problem. Conceptually, scientific research investment is “softer than harder”. For example, in the Huawei incident, many popular science videos talked about lithography machines. Therefore, most of the above policy investments like to see money to build experiments. In the laboratory, buying high-end experimental devices, and investing in software, the results are not good, how can the leaders show them when they visit? Looking at a bunch of computers? On this point, Academician Ni Guangnan has also called for a change in the idea that software is not important. It is necessary to know that industrial software is an industrial strength as important as lithography machines and machine tools.

Fortunately, the country has begun to attach great importance to it. The "14th Five-Year Plan" focuses on solving the national chocking project. 35 key chocking technologies such as lithography machines, chips, operating systems, and tactile sensors will be the focus of the future.

Conclusion: The development of China's industrial software still relies on the vast number of industry practitioners, young talents in China, after the demographic dividend, the next step is the engineer's dividend brought by the improvement of population quality. China has the largest group of college students in the world, and in the future will also have the largest group of engineers in the world. I hope to take this opportunity to arouse the interest of the younger generation in industrial software, and it is worth the effort.