Traditional Culture Encyclopedia - Traditional festivals - What is the development status of gpu chips?
What is the development status of gpu chips?
According to statistics, in 2020, the global gpu industry market will be 254 1 billion dollars, and it is expected to reach 1.853 1 billion dollars in 2027, with an average annual growth rate of 32.82%, maintaining a high-speed growth state.
Driven by market demand, policies and capital, gpu is widely sought after in China. Under the background of geopolitics and big country game, compared with the domestic cpu that has accumulated to a certain extent, it is more urgent and necessary to break the domestic gpu. The talent reserve level and exit mechanism provided by science and technology innovation board also create a good opportunity for the development of domestic gpu.
As a result, gpu has become the sexiest chip investment track after AI. The influx of capital has attracted many elites from big factories to start businesses. According to incomplete statistics, there are nearly 20 start-ups in this field, all of which are involved by first-line investment institutions.
Second, what is the development trend of gpu chips?
There are two main development directions of gpu application in the future. From the demand side, the advantage of gpu lies in parallel computing, and the field that needs large-scale application in parallel computing is the development direction of gpu in the future.
1, a more realistic graphic presentation.
Graphic display is the initial function of gpu chip. With the development of vision technology and virtual reality technology, more realistic graphics display effect will put forward higher requirements for gpu's parallel computing ability. Therefore, graphic display is an important development direction of gpu chips in the future.
2. High performance computing
High performance computing mainly includes general computing and artificial intelligence computing. Generally, gpu is used to handle some operations that cpu can handle, but it is more suitable for gpu operations with strong floating-point computing ability, such as face recognition.
Artificial intelligence computing is another kind of high-performance computing, which is different from the traditional gpu based on stream processor. In most cases, the gpu used for AI computing has lower requirements for floating-point computing accuracy, but higher requirements for computing throughput.
- Previous article:Scheme of incorruptible culture entering primary school campus activities
- Next article:Introduction of Guangxi Special Snacks
- Related articles
- Is it a liar to gradually cross the platform?
- I am a southerner who came to Beijing for a trip and saw the special mutton soup in Beijing. What is mutton soup? Thank you, everyone.
- How to pronounce the six-character Buddhist mantra? Pinyin
- What fonts are used to write English for middle school students
- How to inherit and carry forward China's excellent traditional culture?
- What are the main categories of commonly used Chinese medicines? What are the characteristics?
- Where is Qi Baishi from?
- Is tofu milk moldy tofu
- Do you think traditional festivals like Chinese New Year will gradually disappear?
- Marketing strategy of Ipoh mineral water