Rendering method and apparatus

Abstract

The embodiments of the application disclose a rendering method and apparatus, and relate to the technical field of image processing. The method and the apparatus herein can reduce CPU load rate, and increase bandwidth utilization rate between GPU and CPU. The method includes the following steps: in a user thread corresponding to a user, configuring a rendering process state of the rendering task data submitted by the user; transmitting at least one cache data after the rendering task data is configured with the rendering process state in a user thread to a rendering main thread; and transmitting the cache data to a graphics processor GPU through the rendering main thread for performing graphics rendering. According to the embodiments of the invention, the method herein is applied to graphics rendering.

Description

technical field [0001] Embodiments of the present application relate to the technical field of image processing, and in particular, to a rendering method and device. Background technique [0002] In the field of 3D graphics processing, the processing of graphics usually requires a complex rendering calculation process. Generally speaking, rendering is divided into real-time rendering and non-real-time rendering. Real-time rendering needs to produce several pictures within 1 second, which is mainly used for dynamic real-time display of 3D games and 3D models; Produce realistic renderings, mainly used in film and television animation, advertising planning, interior design, industrial design and other fields. [0003] Generally, during the real-time rendering process, data is first processed by a CPU (Central Processing Unit, central processing unit) and then transmitted to a GPU (Graphics Processing Unit, graphics processing unit) for processing, and finally a final rendered ...

AI Technical Summary

Problems solved by technology

like figure 1 As shown in the current 3D graphics rendering, before rendering, it is necessary to set the rendering process state of the current scene through the CPU. In the rendering process state setting, it is generally necessary to set various parameters. When performing real-time rendering on the cloud server, although The computing power of the CPU is stronger than that of an ordinary PC (personal computer, personal computer), but when implemented on a cloud server based on the C / S (Client / Server, client and server) architecture, there are thousands of When the user terminal access increases, there is a setting of the rendering process state for each user, and the setting of the rendering process state in the prior art is limited to one thread / process by the graphics API , limited by the processing method of the main thread cache FIFO (First-In First-Out, first-in-first-out), such as figure 2 As shown, the rendering process of each user's image frame may need to be processed sequentially in the main thread; and a rendering process for each user needs to include the following image 3 The state setting process of a rendering process is shown as follows: bind vertices (usually apply for memory through the bindvertex function) > set the view (usually through the setviewport function) > bind the rendering pipeline (usually through the bindpipeline function) > draw (usually through the Realized by the Draw function), and finally draw an image frame submitted by the user according to the above settings

In the above process, bindvertex and bindpipelin are usually modified in an OpenGL Contex (OpenGraphics Library Contex, open graphics library context), while the existing rendering API (Application Programming Interface, application programming interface) is limited to the main thread. An OpenGL Context modifies the state of each user's rendering process, so the calculation time spent on setting the state of the rendering process cannot be simply ignored, and CPU multi-core and multi-process cannot play a role at all

[0004] In this way, when the number of connected user terminals increases, the amount of rendering tasks increases, and the load rate of a single CPU of the cloud server is too high, and other CPUs cannot share the rendering work. Can't be used effectively, GPU is not working at its full potential

Images