WO2022048097A1

WO2022048097A1 - Single-frame picture real-time rendering method based on multiple graphics cards

Info

Publication number: WO2022048097A1
Application number: PCT/CN2021/070641
Authority: WO
Inventors: 郭建君; 吴越昕; 孙华庆
Original assignee: 北京蔚领时代科技有限公司
Priority date: 2020-09-02
Filing date: 2021-01-07
Publication date: 2022-03-10
Also published as: CN112057851A; CN113521735A

Abstract

Disclosed in the present invention is a single-frame picture real-time rendering method based on multiple graphics cards, the method comprising: acquiring a picture to be rendered; on the basis of the picture, determining target GPU, the target GPU comprising at least one rendering GPU; dividing the picture into at least one rendering part, the number of rendering parts being less than or equal to the number of rendering GPU; by means of the at least one rendering GPU, rendering the at least one rendering part; synthesising the rendering result of the at least one rendering GPU into a picture frame and outputting same, and implementing rendering of the next picture frame. Also disclosed in the present invention are a single-frame picture real-time rendering system based on multiple graphics cards and an electronic device. The beneficial effects of the present invention are: implementing rendering of the same picture frame by means of multiple GPU, so that the multiple GPU resources can be used more effectively to implement rendering of complex 3D scenes, maintaining a high frame rate, ensuring the smoothness of the picture, and using fewer resources to serve more users.

Description

A real-time rendering method of single-frame image based on multiple graphics cards

technical field

The invention relates to the technical field of cloud games, and in particular, to a real-time rendering method of a single frame image based on multiple graphics cards.

Background technique

Real-time rendering is a GPU-based graphics rendering technology, which is mainly used in games, applications and other fields with real-time graphical interfaces. The current real-time rendering is basically completed by one GPU. Because the performance is sufficient, the general user will not need multi-GPU rendering support. Of course, when the user is in the hardware upgrade stage, when the new and old hardware is used together, the multi-GPU rendering can also make full use of the hardware. Improve user experience. In the field of cloud game cloud hosts, high-performance CPUs and multi-GPUs are standard. It is necessary to make full use of multi-GPU resources to serve user interface and game interface rendering to meet real-time rendering requirements.

SUMMARY OF THE INVENTION

In order to solve the above problem, the purpose of the present invention is to provide a mode that is different from the existing single GPU rendering mode, and can more efficiently use multiple GPU resources to complete the synchronous rendering of the same frame.

In order to achieve the above object, the present invention provides a real-time rendering method for a single frame based on multiple graphics cards, the method comprising the following steps:

obtaining a picture to be rendered, wherein the picture to be rendered is a frame of picture;

Determine a target GPU according to the to-be-rendered picture, wherein the target GPU includes at least one rendering GPU for rendering the to-be-rendered picture;

dividing the picture to be rendered into at least one rendering part, wherein the number of rendering parts is less than or equal to the number of rendering GPUs;

rendering the at least one rendering portion by the at least one rendering GPU;

The rendering result of the at least one rendering GPU is synthesized into one frame of picture and output, and the next frame of picture is rendered.

As a further improvement of the present invention, including multiple GPUs,

The determining of the target GPU according to the to-be-rendered picture includes:

According to the size of the picture to be rendered, at least one rendering GPU that renders the picture to be rendered is determined from the plurality of GPUs.

As a further improvement of the present invention, in step 103, the picture to be rendered is divided into at least one rendering part, including:

The to-be-rendered image is divided into at least one rendering part according to the size of the to-be-rendered image and the number of rendering GPUs.

As a further improvement of the present invention, the picture to be rendered is a vector graphic represented by coordinates,

Divide the to-be-rendered picture into at least one rendering part, including:

generating a bitmap after the vector graphic is rendered through geometric operations and pixel coloring;

According to the size of the bitmap and the number of rendering GPUs, the bitmap is divided to obtain at least one rendering part.

As a further improvement of the present invention, a swap buffer is allocated to each rendering GPU,

The rendering of the at least one rendering part by the at least one rendering GPU includes:

Render a rendering part through a rendering GPU;

The rendering results of each of the rendering GPUs are imported into the corresponding swap buffers.

As a further improvement of the present invention, the target GPU further includes a synthesis GPU for synthesizing rendering results,

Wherein, the outputting after synthesizing the rendering result of the at least one rendering GPU into one frame of picture includes:

Copy the rendering result of each rendering part to memory from the swap buffer corresponding to each of the rendering GPUs;

The rendering results of each of the rendering parts are synthesized by the synthesizing GPU.

As a further improvement of the present invention, a corresponding swap buffer is allocated to the synthetic GPU,

The composite rendering result of the composite GPU is imported into a swap buffer corresponding to the composite GPU and then output.

The present invention also provides a real-time rendering system for single-frame images based on multiple graphics cards, the system comprising:

an obtaining unit, configured to obtain a to-be-rendered picture, wherein the to-be-rendered picture is a frame of picture;

A target GPU determination unit for determining a target GPU according to the picture to be rendered, wherein the target GPU includes at least one rendering GPU for rendering the picture to be rendered;

a splitting unit, configured to split the to-be-rendered picture into at least one rendering part, wherein the number of rendering parts is less than or equal to the number of rendering GPUs;

a rendering unit, configured to execute a rendering program, and the at least one rendering GPU renders the at least one rendering part;

An output unit, which combines the rendering results of the at least one rendering GPU into a frame of pictures and outputs them.

The present invention also provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to achieve the The method of any of claims 1-7.

As a further improvement of the present invention, the computer program is executed by a processor to implement the method according to any one of claims 1-7.

The beneficial effects of the present invention are: the rendering of the same frame picture is completed by multiple GPUs, the rendering of complex 3D scenes can be completed by using multiple GPU resources more efficiently, the frame rate can be maintained high, the smoothness of the picture can be guaranteed, and the use of less Resources serve more users.

Description of drawings

FIG. 1 is a schematic diagram of an implementation flowchart of a method for real-time rendering of a single-frame image based on multiple graphics cards according to an embodiment of the present invention;

2 is a multi-GPU rendering structure diagram of a method for real-time rendering of single-frame images based on multiple graphics cards according to an embodiment of the present invention;

3 is a flowchart of multi-GPU rendering of a method for real-time rendering of single-frame images based on multiple graphics cards according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a system structure of a real-time rendering system for a single frame image based on multiple graphics cards according to an embodiment of the present invention.

detailed description

The present invention will be further described in detail below through specific embodiments and in conjunction with the accompanying drawings.

A real-time rendering method for a single-frame image based on multiple graphics cards according to an embodiment of the present invention, as shown in Figures 1-3, the method includes the following steps:

A target GPU is determined according to the to-be-rendered picture, wherein the target GPU includes at least one rendering GPU for rendering the to-be-rendered picture; the target GPU is an configurable GPU among multiple GPUs.

rendering the at least one rendering portion by the at least one rendering GPU;

Wherein, for example, N rendering GPUs participating in rendering are acquired, and N≥1; then the frame to be rendered is divided into M rendering parts, and 1≤M≤N;

The picture to be rendered can be, for example, a vector graphic represented by coordinates,

Divide the to-be-rendered picture into at least one rendering part, including:

An optional implementation comprising multiple GPUs,

In an optional implementation manner, the to-be-rendered picture is divided into at least one rendering part, including:

For example, as shown in Figure 2, for a frame to be rendered, the frame to be rendered is divided into four rendering parts, and four rendering GPUs, namely GPU1, GPU2, GPU3, and GPU4, are used for rendering. When rendering a task, record the identification of the rendering part and the identification of the corresponding rendering GPU. GPU1, GPU2, GPU3 and GPU4 execute the rendering program in parallel. The rendering program provides the resources required for rendering for each rendering part, and the rendering resources include the current rendering. Some models, textures and other data that need to be used. After all the rendering GPUs complete the rendering and output the rendering results to be synthesized, the synthesis GPU of the synthesizer synthesizes the four rendering parts according to the identification of the rendering part, and after the synthesis GPU completes the synthesis rendering, it outputs the rendered image of one frame, repeating The above rendering operation renders the scene image in real time, and outputs a smooth frame image.

In an optional implementation manner, the target GPU further includes a synthetic GPU for synthesizing rendering results,

Wherein, synthesizing the rendering results of at least one rendering GPU into one frame of picture and outputting it includes: copying the rendering results of each rendering part from the swap buffer corresponding to each rendering GPU to the memory; synthesizing the rendering results of each rendering part by synthesizing the GPU .

An optional implementation manner, the number of rendering GPUs is the same as the number of rendering parts obtained by division, that is, N=M, one rendering GPU corresponds to rendering one rendering part, and the composite GPU is shared with one of the rendering GPUs with a smaller rendering workload. A GPU resource. In this way, different GPUs are allocated to improve the utilization rate of the GPU of the whole machine.

In an optional implementation manner, a swap buffer is allocated to each of the rendering GPUs, and each of the rendering GPUs includes completing image rendering and importing the rendering result into the swap buffer.

Further, allocate a corresponding swap buffer for the composite GPU, wherein the rendering result of at least one rendering GPU is synthesized into one frame of picture and then output, including: importing the composite rendering result of the composite GPU into the swap buffer corresponding to the composite GPU. output.

As shown in Figure 3, a rendering GPU corresponds to rendering a rendering part, each rendering part is associated with a rendering GPU, and the composite GPU is an independent GPU, but generally does not monopolize a GPU, because the workload is small, you can choose to use a dedicated GPU The GPU for display output does the compositing, or one GPU can do both rendering and compositing at the same time. Each rendering process includes rendering, swapping buffers, and copying GPU rendering results to memory. The composite process includes sending the copied rendering results of each rendering GPU to the composite GPU to complete composite rendering, and outputting the composite rendering results to the swap buffer, and the swap buffer outputs the completed rendering results of one frame to the display device. Or an encoding device, completes the rendering of one frame of picture, repeatedly performs the above-mentioned rendering, renders the scene picture in real time, and outputs a smooth frame picture.

In an optional implementation manner, the method is applied to a game process of a cloud game, and the game scenes of multiple users are rendered in real time. Since multiple GPUs are simultaneously rendering at the same time, smooth 3D game scene rendering can be achieved. Effect. When in use, the method of the present invention can be encapsulated into a graphics API library with the ability to call multiple graphics cards, and developers can use one or more graphics cards to perform graphics rendering in a single process by using the provided API.

As shown in FIG. 4 , the present invention also provides a real-time rendering system for single-frame images based on multiple graphics cards, the system includes:

a target GPU determination unit, configured to determine a target GPU according to the to-be-rendered picture, wherein the target GPU includes at least one rendering GPU for rendering the to-be-rendered picture;

in,

The acquisition unit acquires a frame to be rendered, and the target GPU determination unit determines a target GPU according to the to-be-rendered image, and the target GPU includes at least one rendering GPU for rendering the to-be-rendered image; determined by the size of the one-frame image The number of rendering GPUs involved in rendering; a frame to be rendered is usually a vector graphic represented by coordinates, and a bitmap is first generated through geometric operations and pixel coloring; the segmentation unit is based on the size of the bitmap and the rendering GPU. number of divisions. The number of rendering GPUs is the same as the number of divided rendering parts. One rendering GPU renders one rendering part, and provides resources required for rendering for each rendering part. The rendering resources include the models and textures that the current rendering part needs to use. data. The rendering unit executes rendering programs, and each rendering process includes rendering, swapping buffers, and copying GPU rendering results to memory. The output synthesis process includes sending the copied rendering results of each rendering GPU to the synthesis GPU to complete the synthesis rendering, outputting the synthesized rendering results to the exchange buffer, and outputting the completed rendering results of one frame to the display from the exchange buffer. The device or encoding device completes the rendering of a frame. By continuously executing frame stream data, the scene image is rendered in real time, and a smooth video stream image is output.

In an optional implementation manner, in a game process, the game scenes of multiple users are rendered in real time, and the identification of the rendering part, the identification of the corresponding rendering GPU, and the user identification are recorded when the screen is split. After the GPUs complete the rendering respectively, according to the rendering part ID, rendering GPU ID and user ID, the current game scene is synthesized for the user, and the synthesized and rendered game scene is sent to the corresponding user terminal to serve multiple users simultaneously in one game process. , which can ensure the smoothness of the game screen and improve the user experience while reducing the consumption of server resources and the number of servers.

The present disclosure also relates to an electronic device, including a server, a terminal, and the like. The electronic device includes: at least one processor; a memory communicatively connected to the at least one processor; and a communication component communicatively connected to the storage medium, the communication component receiving and transmitting data under the control of the processor; wherein the memory stores Instructions executable by at least one processor to implement the methods in the above-described embodiments.

In an optional implementation manner, the memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs and modules. The processor executes various functional applications and data processing of the device by running the non-volatile software programs, instructions and modules stored in the memory, that is, implementing the method.

The memory may include a stored program area and a stored data area, wherein the stored program area can store an operating system, an application program required by at least one function; the stored data area can store an option list and the like. Additionally, the memory may include high speed random access memory, and may also include nonvolatile memory, such as at least one magnetic disk storage device, flash memory device, or other nonvolatile solid state storage device. In some embodiments, the memory may optionally include memory located remotely from the processor, such remote memory being connectable to an external device via a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

One or more modules are stored in memory, and when executed by one or more processors, perform the method in any of the above method embodiments.

The above product can execute the methods provided by the embodiments of the present application, and have functional modules and beneficial effects corresponding to the execution methods. For technical details not described in detail in the present embodiments, reference may be made to the methods provided by the embodiments of the present application.

The present disclosure also relates to a computer-readable storage medium for storing a computer-readable program, the computer-readable program being used for a computer to execute some or all of the above method embodiments.

That is, those skilled in the art can understand that all or part of the steps in the method for implementing the above embodiments can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium and includes several instructions to make a device ( It may be a single chip microcomputer, a chip, etc.) or a processor (processor) to execute all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .

In the description provided herein, numerous specific details are set forth. It will be understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Furthermore, it will be understood by those of ordinary skill in the art that although some of the embodiments described herein include certain features, but not others, included in other embodiments, that combinations of features of different embodiments are intended to be within the scope of the invention within and form different embodiments. For example, in the claims, any of the claimed embodiments may be used in any combination.

It will be understood by those skilled in the art that although the invention has been described with reference to exemplary embodiments, various changes may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, the inventions are not to be limited to the particular embodiments disclosed, but the inventions are to include all embodiments falling within the scope of the appended claims.

Claims

A method for real-time rendering of single-frame images based on multiple graphics cards, characterized in that the method comprises the following steps:

obtaining a picture to be rendered, wherein the picture to be rendered is a frame of picture;

Determine a target GPU according to the to-be-rendered picture, wherein the target GPU includes at least one rendering GPU for rendering the to-be-rendered picture;

dividing the picture to be rendered into at least one rendering part, wherein the number of rendering parts is less than or equal to the number of rendering GPUs;

rendering the at least one rendering portion by the at least one rendering GPU;

The rendering result of the at least one rendering GPU is synthesized into one frame of picture and output, and the next frame of picture is rendered.
The method for real-time rendering of a single-frame picture based on multiple graphics cards according to claim 1, wherein the method comprises a plurality of GPUs,

The determining of the target GPU according to the to-be-rendered picture includes:

According to the size of the picture to be rendered, at least one rendering GPU that renders the picture to be rendered is determined from the plurality of GPUs.
The real-time rendering method for a single-frame picture based on multiple graphics cards according to claim 1, wherein dividing the to-be-rendered picture into at least one rendering part comprises:

The to-be-rendered image is divided into at least one rendering part according to the size of the to-be-rendered image and the number of rendering GPUs.
The method for rendering a single-frame image in real time based on multiple graphics cards according to claim 3, wherein the image to be rendered is a vector graphic represented by coordinates,

Divide the to-be-rendered picture into at least one rendering part, including:

generating a bitmap after the vector graphic is rendered through geometric operations and pixel coloring;

According to the size of the bitmap and the number of rendering GPUs, the bitmap is divided to obtain at least one rendering part.
The method for real-time rendering of a single-frame picture based on multiple graphics cards according to claim 1, wherein a swap buffer is allocated to each rendering GPU,

Wherein, the rendering of the at least one rendering part by the at least one rendering GPU includes:

Render a rendering part through a rendering GPU;

The rendering results of each of the rendering GPUs are imported into the corresponding swap buffers.
The real-time rendering method for a single-frame picture based on multiple graphics cards according to claim 5, wherein the target GPU further comprises a composite GPU for synthesizing rendering results,

Wherein, the outputting after synthesizing the rendering result of the at least one rendering GPU into one frame of picture includes:

Copy the rendering result of each rendering part to memory from the swap buffer corresponding to each of the rendering GPUs;

The rendering results of each of the rendering parts are synthesized by the synthesizing GPU.
The real-time rendering method for a single-frame picture based on multiple graphics cards according to claim 6, wherein a swap buffer is allocated correspondingly to the composite GPU,

Wherein, the outputting after synthesizing the rendering result of the at least one rendering GPU into one frame of picture includes:

The composite rendering result of the composite GPU is imported into the swap buffer corresponding to the composite GPU and then output.
A single-frame picture real-time rendering system based on multiple graphics cards, characterized in that the system includes:

an obtaining unit, configured to obtain a to-be-rendered picture, wherein the to-be-rendered picture is a frame of picture;

a target GPU determination unit, configured to determine a target GPU according to the to-be-rendered picture, wherein the target GPU includes at least one rendering GPU for rendering the to-be-rendered picture;

a splitting unit, configured to split the to-be-rendered picture into at least one rendering part, wherein the number of rendering parts is less than or equal to the number of rendering GPUs;

a rendering unit, configured to execute a rendering program, and the at least one rendering GPU renders the at least one rendering part;

An output unit that combines the rendering results of the at least one rendering GPU into a frame of pictures and outputs them.
An electronic device comprising a memory and a processor, characterized in that the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement claims 1-7 The method of any of the above.
A computer-readable storage medium on which a computer program is stored, wherein the computer program is executed by a processor to implement the method according to any one of claims 1-7.