Skip to main content

Apple and Intel want to join the game, what happened to the GPU market?

Intel recently announced that it will launch Xe-LP GPU at the end of this year, officially entering the independent GPU market, and will hand over to TSMC for foundry. At the 2020 WWDC held not long ago, Apple also revealed that it is possible to abandon AMD's GPU and use a self-developed solution based on the ARM architecture. It will launch a self-developed GPU next year. What happened to the GPU market? Why are the giants entering the game?

   Massive data calls for high-performance GPU

   Why has the demand for GPUs increased so rapidly in recent years? Because we are entering an era where everything needs to be visualized. Dai Shuyu, a partner of Aiwa (Beijing) Technology Co., Ltd., told a reporter from China Electronics News that visualization requires a large amount of graphics and image computing capabilities, and a large amount of high-performance image processing capabilities are required for both the cloud and the edge.

   Aiwa (Beijing) Technology Co., Ltd. is an enterprise that provides industrial AR software solutions, and many of its customers are from the manufacturing industry. Many scenarios in smart factories now require GPUs. She gave an example of industrial Internet scenarios. A typical application scenario in smart factories is to use AI for quality inspection of production lines, which are currently used in many production lines such as automobile production lines and home appliance production lines. AI vision for inspection. "It is impossible to transmit a large amount of real-time detection data to the cloud. This requires the installation of high-performance computers and desktop computers at the production site, and the computers must have strong GPU processing capabilities. A production line uses at least one GPU and one mid-range The price of the GPU is tens of thousands of yuan, and the high-end one is more than 100,000 yuan.” Dai Shuyu said, in addition, the visualization of the digital factory involves 3D rendering, audio and video streaming, dynamic data transmission, and mining analysis. Force puts forward higher requirements.

  GPU affects "effects" and "face value", and today many parts of the world rely on "face value" for food. Tianshu Zhixin is a Chinese startup company specializing in general-purpose GPU (GPGPU). Zheng Jinshan, the chief scientist of Tianshu Zhixin, told the reporter of China Electronics News: “In fact, GPU is not just getting hot now. It has been developing. Computer/mobile phone, It is widely used in dedicated game consoles. GPU is a key unit in personal computers second only to CPU. Except for three-dimensional games, the two-dimensional acceleration of graphical interfaces is all done by GPU. And modern operating systems also have a large number of graphical interfaces Using three-dimensional technology, the look and feel of the interface without GPU is very poor. Nowadays, mobile phones also integrate GPU, and some interface effects need to be accelerated by it."

In addition to image and graphics processing, GPUs also have strong computing capabilities. Due to the powerful parallel processing capabilities and programmable pipelines of graphics processors, stream processors can process non-graphics data, especially in the face of single instruction stream and multiple data When streaming (SIMD). When the amount of data processing calculations is far greater than the needs of data scheduling and transmission, the performance of general-purpose graphics processors greatly surpasses the traditional central processing unit applications, so they are widely used in high-performance computing and AI training and analysis. It is called GPGPU (General Purpose Graphics Processing Unit).

   GPU market is booming

   We are entering a world where "parallel computing" is ubiquitous, and the powerful parallel processing capabilities of GPUs are becoming more and more useful. Chi Xuebin, deputy director of the Computer Network Information Center of the Chinese Academy of Sciences, said in an interview with a reporter from China Electronics News that we are entering the era of heterogeneous computing, and parallel computing is becoming ubiquitous. The computing system of this high-performance computer TOP500 adopts the GPU+CPU architecture, which indicates that GPU+CPU will become a trend.

   Judging from the data forecasts of various research institutions, GPUs are showing vigorous growth in the AI ​​reasoning market, server market, and data center market. Futu Securities predicts that as the demand for AI inference market continues to grow, the market size of AI inference in 2020 will be 18 billion U.S. dollars and is expected to reach 66 billion U.S. dollars by 2025. Guofu Securities predicts that the demand for GPUs from servers is showing a rapid growth trend. The demand for server GPUs in 2020 is about 9.7 million, and it is expected to grow to 25 million by 2025, with an average annual compound growth rate of about 10%. IDC predicts that the demand for GPUs in the data center market will show a rapid growth trend. In 2020, the data center GPU market will be 137 billion U.S. dollars, and it is estimated that by 2023, the data center GPU market will reach 229 billion U.S. dollars.

  From the perspective of GPU suppliers, Nvidia and AMD occupy the second place in the GPU market. In the past five years, Nvidia's stock price has soared by 2268%. After the market closed on July 8 this year, Nvidia's market value surpassed Intel for the first time, and it once became a chip company with market value in the US. AMD's market value has also soared, and its market value has approached the $100 billion mark in recent times.

   Whether it is customer demand or market data, or changes in the market value of suppliers, it is proving the fact that the popularity of GPU is getting more and more prosperous.

   Many manufacturers have entered the market

   The prosperous GPU market attracts more manufacturers to enter the game.

Intel is one of the entrants. The rapid soaring of Nvidia and AMD in recent years and the rapid growth of the GPU market are impossible for Intel to fail to see. Therefore, since last year, Intel no longer emphasizes that it is a CPU manufacturer, but an XPU manufacturer. Including CPU, GPU, NPU, VPU, FPGA, etc.

   Prior to this, Intel had a GPU, but the GPU function was integrated in the CPU, which was an integrated graphics card. "The GPU integrated in the CPU can't make the performance very strong. If you want to make it stronger, the area will become larger and the power consumption will increase." Li Yang, an analyst at CCID Consulting Integrated Circuit Information Industry Research Center Said in an interview with a reporter from China Electronics News. It is impossible to use integrated graphics in high-performance computing, AI training and analysis. With the popularization of 5G, the advancement of autonomous driving, the rapid development of AI, and the explosive growth of data volume, it has become a must for Intel to do discrete graphics and launch discrete GPUs. selected.

   Two years ago, Intel announced that it would launch a dedicated GPU chip and formulated a plan for a single GPU architecture Xe, which consists of three different micro-architectures: Xe-LP, Xe-HP and Xe-HPC. Respectively, from ultra-mobile devices to gaming consoles, to workstations and data centers. At the recent Intel 2020 Annual Architecture Day, Intel spent a lot of time talking about its GPU architecture plans. It can be seen from Intel’s Xe-LP GPU architecture that it is preparing to provide discrete GPUs for all products from data centers, HPCs, high-end game consoles to notebooks, which means that Intel will not release any of them in almost all markets where the exclusive display is used. Over. The news that Intel GPU manufacturing will be handed over to TSMC for foundry has also recently spread, so Intel’s independent GPUs are getting closer and closer.

   Another company eager to test the GPU market is Apple. At the WWDC 2020 held not long ago, Apple revealed that it is possible to abandon AMD's GPU and use a self-developed solution based on the ARM architecture. There is news in the industry that Apple will cooperate with Apple Silicon to launch a self-developed GPU chip in the second half of 2021. The chip is code-named lifuka, using TSMC's 5nm process, and will be carried in iMaC to be launched next year.

   Apple’s GPU was originally a license purchased from Imagination. Now Apple hopes to design its own GPU with stronger capabilities to achieve differentiation. All kinds of information flowed out, further confirming the news that Apple wants to develop its own GPU.

  GPU market pattern may change

   The entry of Intel and Apple will definitely have an impact on the landscape of the GPU market. Apple's research and development of GPU is mainly "self-sufficient", forming a differentiated advantage and will not have a huge impact on the supplier camp, and whether Intel can stir the entire GPU market has become a highlight. At present, the independent GPU market is almost dominated by Nvidia and AMD, so will Intel's entry make the GPU market supply pattern "change two to three"?

   Zheng Jinshan believes that the difficulty of GPU lies in the fact that 3D rendering itself is a relatively difficult problem, especially the real-time rendering algorithm of light and shadow. Corresponding hardware needs to be designed with accelerators, and meet a variety of algorithms and software interfaces. And in history, there are still many APIs that must be compatible, and they need to match a variety of games in performance, and the overall requirements are relatively high. GPU has now entered the realization of ray tracing algorithms, and the gaming industry requires products with more realistic rendering and higher resolution that never stop. With the rise of the VR market, the algorithm/computing power demand needs further development.

   From the current information given by Intel, its Xe-LP will meet the various needs of users in modern games, including fiber tracking and 4K enthusiast performance. Some analysts say that it is expected to compete with Nvidia's Ampere series graphics cards and AMD's RDNA2 graphics cards this year. Other Xe-HPCs are currently confirmed to be used in the U.S. Department of Energy's exascale supercomputing Aurora. Nvidia founder Huang Renxun once said about Intel’s entry: “Intel has finally realized the importance of GPUs, and welcome to the GPU world.” The original GPU concept was proposed by Nvidia, so Nvidia is not a “dry”. From the blueprint to the technical realization, whether Intel can deliver satisfactory answers on schedule depends on how the product actually comes out.

   "In addition to the hardware part, the focus of competition is the software ecosystem of the entire game industry, that is, software products with excellent user experience." Zheng Jinshan said.

Li Yang expressed the same view. He believes that Nvidia is not only strong at the hardware level. In recent years, it has accumulated a large number of ecological partner resources in games, graphics processing manufacturers, academic institutions, developers, etc., from hardware to development platforms to The barriers established by the partner ecosystem are the gulf that Intel is insurmountable in the short term.

   New entrants have yet to be tested by the market

Regarding software, developers, and application ecology, Intel, which is preparing to enter the independent display GPU market, could not have thought of it. Therefore, at the architecture day held not long ago, Intel Chief Architect Raja Koduri emphasized that Intel’s Xe architecture is the main design of Xe There are three points of thinking—software first, expansion, and new applications. Xe is not an isolated product, but will be designed in collaboration with Intel's complete hardware product portfolio and a unified software development environment. Raja has repeatedly emphasized that he hopes to use the same software stack and the same micro-architecture design to achieve rapid expansion of multiple Xe products.

   At present, the main architecture of Xe is based on slices. There are 6 subSlices, up to 96 EU (Execution Unit) in total, which can provide 1536 floating-point operations per cycle. Compared with Intel's previous generation of integrated displays, Xe-LP has redesigned its own micro-architecture. Xe is a brand-new GPU product series, and one of its main optimization focuses is the design of the software system.

   On the Architecture Day, everyone raised many questions about GPUs, but Intel did not answer them all. But in any case, Intel entered the independent display market and brought a redesigned architecture, bringing more choices and imagination to this market. The ability of its products still needs to be tested by the market. So Intel is here, will those high-end GPUs that are sold at the price of a car in the GPU sector drop in price? This is an important question that many netizens care about. See you after the mass production of the product.


  1. I generally check this kind of article and I found your article which is related to my interest.Wholesale RTX, AMD Graphics Cards Genuinely it is good and instructive information. Thankful to you for sharing an article like this.

  2. I heard yesterday that Netflix is looking for some developers to usher in a new era of video games. Is Netflix likely to merge with Apple and Intel in the future? Dissertation Editing Services

  3. I just wanted to say this is an elegantly composed article as we have seen here. I got some knowledge from your article and also it is a significant article for us. Thanks for sharing an article like this.Macbook Retina 12 Parts Supplier


Post a Comment

Popular posts from this blog

AMD's GPU technology enters the mobile phone chip market for the first time

In addition to the release of the Exynos2100 processor, Samsung also confirmed a major event at this Exynos event, that is, the custom GPU that they have worked with AMD for many years will soon appear and will be used on the next flagship machine. The current Exynos2100 processor uses ARM’s Mali-G78GPU core with a total of 14 cores, so the GPU architecture developed by Samsung will be the next Exynos processor, and the GPU will be the focus. This is probably the meaning of Exynos2100’s GPU stacking. The key reason. Dr. InyupKang, president of Samsung’s LSI business, confirmed that the next-generation mobile GPU in cooperation with AMD will be used in the next flagship product, but he did not specify which product. Samsung is not talking about the next-generation flagship but the next one, so it is very likely that a new Exynos processor will be available this year, either for the GalaxyNote21 series or the new generation of folding screen GalaxyZFold3. In 2019, AMD and Samsung reached

NVIDIA officially launches RTX 30 series mobile graphics cards

In the early morning of January 13, NVIDIA officially launched the RTX30 series of mobile graphics cards at the CES2021 exhibition. Ampere-based GPUs have also reached the mobile terminal, mainly including RTX3080, RTX3070 and RTX3060 models. In addition to improving game performance, the RTX30 series of mobile graphics cards have twice the energy efficiency of the previous generation, and support the third-generation Max-Q technology, mainly supporting DynamicBoost2.0 dynamic acceleration technology, WisperMode2.0 noise control, ResizableBAR (similar to AMD’s SAM technology) and DLSS. The third-generation Max-Q technology uses AI and new system optimization to make high-performance gaming laptops faster and more powerful than ever. These technologies include: ·DynamicBoost2.0: The CPU and GPU powers of traditional gaming notebooks are fixed, while games and creative applications are dynamic, and the requirements for the system will vary with the number of frames. With DynamicBoost2.0,