In digital twin projects, cloud rendering technology has become an important tool. It allows users to use it on different terminals, including computers, mobile phones, tablets, smart screens, TVs and TV boxes. For users, using cloud rendering is as simple and convenient as opening a web page.There is no need to learn or install additional programs, just click to use .
Moreover, when network conditions are good, cloud rendering can also break through the limitations of time and space. Since the previous epidemic, more and more digital twin project owners have required projects to use real-time cloud rendering technology and support access from multiple terminals. This trend makes digital twin projects more flexible and convenient, able to meet the needs of different users.
The emergence of cloud rendering technology gradually emerged with the continuous development of 5G network and computer technology. It is a relatively new technology. Although UE4 officially provides a plug-in for the pixel flow solution, in actual projects, many developers need to go through in-depth exploration before they can apply cloud rendering technology into practice. In the process of long-term contact with many project personnel, we found that one of their common questions about cloud rendering technology is what type of graphics card the server needs to be equipped with.
Real-time rendering and offline rendering are different. Real-time rendering pays more attention to real-time interactivity, unlike offline rendering which has high requirements on the CPU. Real-time rendering uses the computing power of the server-side GPU to complete rendering and encoding, and transmits real-time images to the terminal through the network. This technology enables real-time interaction on the terminal. Users can perform real-time rendering through the cloud server and instantly transmit the rendering results to the terminal for display. Compared with offline rendering, real-time rendering pays more attention to response speed and interactivity, providing users with a better real-time experience.
What are the requirements for server graphics cards in the digital twin project? This depends on two factors:
- How resource intensive is the digital twin program itself?
- How many concurrencies the server should support.
In order to evaluate the resource consumption of the digital twin program without considering the number of concurrencies, we can run the program on a general work computer and find the parts with a larger number of faces in the program to test the GPU and CPU usage of the computer. Condition. In this way, you can determine whether the program can run smoothly under the corresponding graphics card and CPU configuration.
You can refer to this configuration when preparing a server to use the cloud rendering system. If you need to support multiple concurrencies, you can open several more programs to run on the basis of the above, and monitor the computer's GPU and CPU usage. When the occupancy rate reaches 80%, it is the limit of the maximum number of concurrencies that the server can achieve under the same configuration.
If you want to support more concurrent requirements of digital twins, you can consider adding servers or adding graphics card resources. By using point cloud rendering technology to achieve load balancing, multiple concurrent tasks can be better supported. However, it should be noted that there is no fixed reference number for the specific number of concurrent support. Even for the same model of 3070 graphics server, the differences between different programs are huge.
For example, a large-scale smart city or smart factory digital twin model developed based on UE4 may only support 1-2 concurrent tasks, while a virtual simulation digital twin program based on Unity may support more than a dozen concurrent tasks. Therefore, specific judgment and evaluation are required for different procedures.
Real-time rendering has been widely used in digital twin smart cities, smart factories, 3D games, industrial design, virtual reality, disaster simulation and product display and other fields. Corresponding to real-time rendering is offline rendering. Offline rendering is generally used in special effects production such as movies and animations. It has lower requirements for real-time performance but higher requirements for picture quality. It usually uses the CPU. To render.