CN104952096B - CPU and GPU mixed clouds rendering intent, device and system - Google Patents
CPU and GPU mixed clouds rendering intent, device and system Download PDFInfo
- Publication number
- CN104952096B CN104952096B CN201410125332.2A CN201410125332A CN104952096B CN 104952096 B CN104952096 B CN 104952096B CN 201410125332 A CN201410125332 A CN 201410125332A CN 104952096 B CN104952096 B CN 104952096B
- Authority
- CN
- China
- Prior art keywords
- cpu
- gpu
- rendering
- rendering task
- task
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Processing Or Creating Images (AREA)
Abstract
The invention discloses a kind of CPU and GPU mixed clouds rendering intent, device and system, are related to field of cloud computer technology.Rendering task is decomposed into logic continuous type rendering task in this method and parallel rendering task is sent respectively to CPU and GPU processing, passes through the primitive frame figure of the linkage output generation of CPU rendering results and GPU rendering results；Segmented fluidisation coding is carried out to the coding pass distributed in order in Unified coding treatment source pool after the primitive frame figure segmentation of sequence, the coding result that each coding pass exports is integrally formed Media Stream.This method, device and system are integrated using CPU and GPU rendering capabilities, and joint has the pattern for stressing the division of labor again, unified to be segmented fluidization treatment by CPU, GPU Unified coding processing capacity resource pool, realize that picture is rendered into the integrative solution of fluidisation.Relatively single CPU, GPU rendering intent, the program can effectively promote service quality and service efficiency.
Description
Technical field
The present invention relates to field of cloud computer technology, more particularly to a kind of CPU and GPU mixed clouds rendering intent, device and it is
System.
Background technology
It is a kind of based on high in the clouds remote rendering service to render cloud, and user is allowed independently to complete various render with web interface, entirely and is appointed
Business.It renders (Rendering) to refer in computer graphics, 2D or 3D models is transformed to the process of image.In graphics pipeline
In, rendering is last important step, and model and the final display effect of animation are obtained by it.
The render time of every width picture according to model complexity difference, differs greatly, few then several seconds, at most several days；Model
More complicated, render time is longer.With film《A Fanda》For, unit renders an average frame (one layer, 2K~4K resolution ratio)
Time is 2 hours；The film of 160 minutes, 9000 width picture (60 seconds × 30 frame × 5 layer) per minute, whole render time need
It 2880000 hours, is equivalent to 328 years.In addition 3D three-dimensional film effects, per frame picture, there are two images, rendering to need more than 600
Year could complete.
Just because of above-mentioned reason, rendering task is placed into high in the clouds, while hundreds and thousands of machines is called to carry out picture wash with watercolours
The decline of the geometry order of magnitude can be presented in dye, integral operation time, and originally unit needed the rendering task that the centuries completes, less than 1 year
Even it can be completed within several days.
Meanwhile license (License) price such as various large size Rendering software Blender, Maya and 3Ds MAX is high, and
Running environment requirement is high, is difficult to bear for small-sized animation, gaming enterprise, and exploitation can be well solved using cloud is rendered
Cost is conducive to industry development and Related product exploitation.
The key for rendering cloud is the treatment technology in high in the clouds, most of on the market at present to be rendered using CPU serial frames
Operation, such as Ali's cloud, Shanghai net wash with watercolours, Renderbus etc..Because of CPU serial frames, the logic handled one by one using process is right
In the static type picture for having precedence be applicable；But for extensive continuity such as video flowing, have 3 dimension modules
Figure application etc., does not have efficient processing capacity.
For CPU, GPU uses parallel architecture, has more computing units and memory bit wide, makes it in floating-point
Operation, parallel computation etc. have more prominent advantage.In terms of its rendering capability, for continuity Real-time graphic processing,
Such as Film Animation etc. has the advantage that CPU hardly matches.Cloud computing service-Asia based on GPU has been introduced in Amazon at present
The superpower graphics calculations example of the inferior cluster of horse, this example provides the GPU processing capacities in cloud, the memory capacity including 22GB,
33.5 EC2 computing units provide the instant access of the GPU calculated performances of hight coordinate for developer and enterprise.Domestic Zan Qike
Skill similarly constructs similar framework, using GPU advantages, handles the online real-time rendering of threedimensional model.
But, cloud renders service and is all based on single CPU or GPU architecture offer service at present, for needing static state simultaneously
Picture renders large-scale CG, the Film Animation for rendering demand online with real-time three-dimensional model, can not meet very well.
Invention content
The inventors found that above-mentioned exist in the prior art problem, and therefore at least one in described problem
A problem proposes a kind of new technical solution.
It is an object of the present invention to provide a kind of technical solutions rendered for cloud.
According to the first aspect of the invention, a kind of CPU and GPU mixed clouds rendering intent is provided, including：
Rendering task is decomposed into logic continuous type rendering task and parallel rendering task, by logic continuous type rendering task
Be sent to CPU serial process, by parallel rendering task be sent to GPU carry out concurrent operation processing, and pass through CPU rendering results with
The primitive frame figure of the linkage output generation of GPU rendering results；
The primitive frame figure for rendering generation is integrated and sorted according to fluidisation, exports the primitive frame figure of sequence；
The primitive frame figure of sequence is segmented, gives primitive frame figure segment assignments to Unified coding treatment source pool in order
In coding pass carry out segmented fluidisation coding, wherein, Unified coding treatment source pool by CPU, GPU code capacity build and
Into；
The coding result that each coding pass exports is ranked up and is integrally formed Media Stream.
Optionally, this method further includes：Increase personage or cartoon character in the Media Stream for completing fluidization treatment.
Optionally, included by CPU rendering results and the primitive frame figure of the linkage output generation of GPU rendering results：
Before CPU and GPU carries out rendering task, CPU is carried out in resource coordination and hardware assigning process for rendering task
Order is registered, and is distributed and is rendered according to relevant order in render process, exports in result integration process and correspond to corresponding sequence number
Graphic result is integrated.
Optionally, logic continuous type rendering task is sent to CPU serial process to include：CPU is handled simultaneously using multinuclear
Mechanism completes serial process, and export relevant rendering result according to priority logic to rendering task；
Parallel rendering task is sent to GPU progress concurrent operation processing to include：Using GPU local memories and bus memory
Interaction, and share content by GPU and pass to each processing cores of GPU to rendering task progress parallel processing and calculating, and finally
Output is corresponding to render handling result.
Optionally, logic continuous type rendering task includes actor model rendering, action renders；Parallel rendering task includes big
Type scene rendering.
According to another aspect of the present invention, a kind of CPU and GPU mixed clouds rendering device is provided, including：
Link hybrid rending module, for rendering task to be decomposed into logic continuous type rendering task and parallel rendering times
Business, CPU serial process is sent to by logic continuous type rendering task, and parallel rendering task is sent to GPU concurrent operations processing,
And pass through CPU rendering results and the primitive frame figure of the real-time linkage output generation of GPU rendering results, to rendering the primitive frame of generation
Figure is integrated and is sorted according to fluidisation, exports the primitive frame figure of sequence；
Dynamic fluidized processing module, for the primitive frame figure of the sequence from linkage hybrid rending module to be segmented,
Primitive frame figure segment assignments are subjected to segmented fluidisation coding to the coding pass in Unified coding treatment source pool in order,
In, Unified coding treatment source pool is built-up by CPU, GPU code capacity；By the coding result that each coding pass exports into
Row sequence is simultaneously integrally formed Media Stream.
Optionally, which further includes：Character embedded module, for increasing in the Media Stream for completing fluidization treatment
Personage or cartoon character.
Optionally, linkage hybrid rending module includes：
Rendering task classification scheduling unit, for rendering task to be decomposed into logic continuous type rendering task and parallel rendering
Logic continuous type rendering task is sent to CPU serial process, parallel rendering task is sent at GPU concurrent operations by task
Reason；
Frame artwork render processing unit, for completing corresponding rendering work, CPU collection based on the arithmetic element of CPU and GPU
The middle single serial macrooperation amount rendering task of processing；GPU focuses on the macrooperation amount rendering task of parallel synchronous；It renders and completes
Primitive frame figure is respectively exported afterwards；
Graphic result integrates sequencing unit, suitable according to fluidizing for the primitive frame figure for exporting CPU and GPU hybrid rendings
Sequence is integrated and the primitive frame figure of sequence output sequence.
Optionally, dynamic fluidized processing module includes：
Primitive frame figure segmenting unit, for the primitive frame figure for rendering output to be resolved into each segmentation in sequence；
Fluidization treatment unit is segmented, it, will for building unified Video coding treatment source pool using CPU, GPU fluidisation ability
The coded treatment channel that the primitive frame figure correspondence of different fragments distributes to unified resource pond carries out fluidization treatment, after processing is completed
Coded treatment channel respectively exports segment media stream；
Result integral unit is fluidized, for integrating the segmentation fluidization treatment of each coded treatment channel output as a result, carrying out whole
Conjunction forms final Media Stream.
Optionally, logic continuous type rendering task includes actor model rendering, action renders；Parallel rendering task includes big
Type scene rendering.
According to another aspect of the invention, a kind of cloud rendering system is provided, is rendered including above-mentioned CPU and GPU mixed clouds
Device.
An advantage of the present invention is that being integrated using CPU and GPU rendering capabilities, joint has the pattern for stressing the division of labor again,
Fluidization treatment is uniformly segmented by CPU, GPU Unified coding processing capacity resource pool, realizes that picture is rendered into the integration of fluidisation
Solution can effectively promote service quality and service efficiency.
By referring to the drawings to the detailed description of exemplary embodiment of the present invention, other feature of the invention and its
Advantage will become apparent.
Description of the drawings
The attached drawing of a part for constitution instruction describes the embodiment of the present invention, and is used to solve together with the description
Release the principle of the present invention.
With reference to attached drawing, according to following detailed description, the present invention can be more clearly understood, wherein：
Fig. 1 shows the flow chart of one embodiment of CPU and GPU mixed clouds rendering intent according to the present invention；
Fig. 2 shows the overall operation frames of one embodiment of CPU and GPU mixed clouds rendering system according to the present invention
Figure；
Fig. 3 is the flow chart for another embodiment for showing CPU and GPU mixed clouds rendering intent according to the present invention；
Fig. 4 shows CPU, GPU hybrid rending procedure chart according to an embodiment of the invention；
Fig. 5 shows CPU, GPU linkage fluid mapper process figure according to an embodiment of the invention；
Fig. 6 shows the structure chart of one embodiment of CPU and GPU mixed clouds rendering device according to the present invention；
Fig. 7 shows the structure chart of another embodiment of CPU and GPU mixed clouds rendering device according to the present invention.
Specific embodiment
Carry out the various exemplary embodiments of detailed description of the present invention now with reference to attached drawing.It should be noted that：Unless in addition have
Body illustrates that the unlimited system of component and the positioned opposite of step, numerical expression and the numerical value otherwise illustrated in these embodiments is originally
The range of invention.
Simultaneously, it should be appreciated that for ease of description, the size of the various pieces shown in attached drawing is not according to reality
Proportionate relationship draw.
It is illustrative to the description only actually of at least one exemplary embodiment below, is never used as to the present invention
And its application or any restrictions that use.
Technology, method and apparatus known to person of ordinary skill in the relevant may be not discussed in detail, but suitable
In the case of, the technology, method and apparatus should be considered as authorizing part of specification.
In shown here and discussion all examples, any occurrence should be construed as merely illustrative, without
It is as limitation.Therefore, the other examples of exemplary embodiment can have different values.
It should be noted that：Similar label and letter represents similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, then in subsequent attached drawing does not need to that it is further discussed.
GPU computing units are more, equally have prominent ability for rendering to calculate and handle, still, due to being based on CPU framves
The solution of structure and mainboard inbuilt display is more easy to dispose, and monokaryon processing capacity is stronger etc., the CPU architecture in actual application
Cloud rendering solution become mainstream.
Demand is rendered in order to more preferably meet CG and the big batch of Film Animation, mass data, the present inventors have noted that：Needing will
CPU serially and GPU concurrent operation ability unified plannings and scheduling, realizes that GPU, CPU mixing linkage render, can just accomplish hardware money
Source optimization scheduling, graph data efficient process and real-time online render, the seamless connection of fluidisation.
Fig. 1 shows the flow chart of one embodiment of CPU and GPU mixed clouds rendering intent according to the present invention.
As shown in Figure 1, step 101, logic continuous type rendering task and parallel rendering task are decomposed by rendering task, it will
Logic continuous type rendering task is sent to CPU serial process, and parallel rendering task is sent to GPU carries out concurrent operation processing,
And pass through CPU rendering results and the primitive frame figure of the real-time linkage output generation of GPU rendering results.
Step 102, the primitive frame figure for rendering generation is integrated and is sorted according to fluidisation, export the primitive frame of sequence
Figure.
Step 103, the primitive frame figure of sequence is segmented, in order by primitive frame figure segment assignments to Unified coding at
It manages the coding pass in resource pool and carries out segmented fluidisation coding, wherein, Unified coding treatment source pool encodes energy by CPU, GPU
Power is built-up.
Step 104, the coding result that each coding pass exports is ranked up and is integrally formed Media Stream.
Fig. 2 shows the overall operation frames of one embodiment of CPU and GPU mixed clouds rendering system according to the present invention
Figure.It will illustrate how to realize on the basis of bottom software and hardware mainly in combination with Fig. 2 below and include CPU, GPU hardware integration capability and base
Plinth Rendering software ability realizes CPU, GPU linkage hybrid rending function.
Wherein, bottom software and hardware resources are the bases for realizing linkage hybrid rending, and hardware resource mainly includes CPU and GPU
Refer to render function tool software, such as including Maya, 3DS Max etc. to provide basis Deng, software.
1) link hybrid rending management.It is main to include structure bottom hardware (CPU, GPU) resource and capability distribution, task tune
Function is spent, realization makes full use of hardware capabilities, realizes CPU, GPU hybrid rending function.The part can be further subdivided into：Task
It decomposes and scheduling of resource, graphic result is integrated and sorted, fluidizes result merging and output etc..Task-decomposing and scheduling of resource, will
Rendering task decomposes, and comprehensive transfer CPU, GPU resource, is taken jointly by CPU, GPU resource and completes related rendering task；Figure
Shape result is integrated and sequence, is completed for graphically rendering processing as a result, primitive frame figure is integrated and sorted, for rear afterflow
Change processing lays the foundation；It fluidizes result to merge and export, the Media Stream segment for referring to for CPU, GPU to be separately encoded output is incorporated into
Together, animation or CG etc. that output needs are edited and.
2) linkage renders process part.Mainly include：Primitive frame figure renders, dynamic fluidized processing and character embed three
A part.The rendering of primitive frame figure is the core of CPU, GPU hybrid rending, is referred mainly to various raw data, such as person model,
Pixel and scene original painting etc. are totally dispatched by CPU, flexibly use CPU serial and the parallel advantages of GPU, corresponding distribution are relevant
Rendering task or step, realize the hybrid rending of fusion CPU, GPU ability, and final output can be directly used for the original of fluidization treatment
Frame figure.Dynamic fluidized processing refers to according to design requirement, and the primitive frame figure for rendering acquisition is converted into the mistake of animation or CG fluidisations
Journey.Fluidisation task should be given to both CPU and GPU, synthesis Video coding energy respectively in the process using segmentation fluidization treatment mechanism
Power, relatively single CPU or GPU patterns, can more fast complete fluid mapper process.Character embeds, and refers to and completes to fluidize
In the Media Stream of processing, increase corresponding personage or cartoon character according to plot design, be finally completed the system of film CG or animation
Make process.
It is low to break through single CPU, GPU rendering efficiencies for GPU+CPU mixed cloud rendering intents in above-described embodiment, it is difficult to full
The task that foot high-volume, extensive compatible static and dynamic render, such as (Computer Graphics are calculated by 3D films, CG
Machine figure) etc., using the integration of CPU and GPU rendering capabilities, joint has the pattern for stressing the division of labor again, and CPU emphasis handles merely big
Operand serial task；The parallel macrooperation task of GPU emphasis processing, finally, passes through CPU, GPU Unified coding processing capacity resource
Pond, it is unified to be segmented fluidization treatment, realize that picture is rendered into the integrative solution of fluidisation.Relatively single CPU, GPU is rendered
Method, the program can effectively promote service quality and service efficiency.
Fig. 3 is the flow chart for another embodiment for showing CPU and GPU mixed clouds rendering intent according to the present invention.
As shown in figure 3, step 301, high in the clouds rendering server receives the rendering server request of user, responds user's wash with watercolours
Contaminate service request.It is, for example, the remote rendering service request from remote user to render service request.
Step 302, in response to the rendering service request of user, high in the clouds rendering server carries out rendering hardware environment prison
It surveys, judges whether hardware environment condition meets service requestIf so, continuing step 303, otherwise, the service of return can not carry
For (step 310).
Step 303, high in the clouds rendering server configuration hardware environment, calls bottom hardware resource, starts Rendering software, structure
Render hardware environment.
Step 304, user prepares original rendering material, sets rendering demand.
Step 305, high in the clouds rendering server synthesis CPU, GPU ability, mixing call CPU, GPU to carry out primitive frame figure fusion
It renders, exports rendering result primitive frame figure.In one embodiment, high in the clouds is according to rendering setting requirements, unified co-allocation
CPU, GPU rendering task by resource consolidations such as archetype, scene and pixels, are mainly undertaken serially, simultaneously respectively by CPU, GPU
Row rendering task, joint carry out hybrid rendering together；Various pictures are exported by rendering, it is defeated respectively according to fluidisation frame number order
Go out and preserve.
Step 306, it is material using the primitive frame figure for rendering generation, CPU, GPU is called to carry out segment encoding fluidisation.It will render
The primitive frame figure of completion calls CPU, GPU hybrid coding resource pool, using segmentation fluidization treatment.
Step 307, the integration of each segment encoding stream is summarized, forms the output of standard cartoon medium stream.
Step 308, in standard media stream, increase corresponding personage/cartoon character image, complete Film Animation or CG knots
Fruit.In large-scale production of film and TV or game, need after first corresponding scene is completed, be just embedded in personage or cartoon character
In animated video stream, final film effect is formed.
GPU, CPU mixed architecture of the disclosure based on innovation, comprehensive GPU, CPU processing capacity, structure are suitble to draw from static state
Face renders, to real time 3-D image output and CPU, GPU hybrid rending integration solution of fluidization treatment.
Fig. 4 shows CPU, GPU hybrid rending procedure chart according to an embodiment of the invention.As shown in figure 4, in this implementation
In example, CPU, GPU linkage hybrid rending include：Scheduling of resource renders processing and output, figure knot with hardware coordination, primitive frame figure
Fruit integration and sequence etc..The serial advantages of CPU are wherein made full use of, scheduling of resource and hardware coordination are carried out by CPU first.Then,
CPU, GPU carry out primitive frame figure and render processing and output respectively according to the task of coordination.Wherein CPU is called using bus memory
Multinuclear ability realizes linear logic processing processing；GPU is interacted using shared video memory with bus memory, and each processing core passes through GPU
Sharing Memory Realization parallel processing；Rendered finally by cpu logic operation and GPU figures, graphic result is uniformly given CPU into
Row is integrated, sequence, realizes the overall logicization series connection of rendering result, overall to complete CPU, GPU linkage hybrid rending link.Specifically
Including：
(1) scheduling of resource and hardware coordination
CPU operational efficiency and core processing ability, rendering is made full use of to distribute scheduling of resource, i.e. rendering task before starting
It is dominated by it, for original render picture and task arrangement.By bus memory, carry out related data and store and transmit, it need to
Calculation process is serialized, such as CPU is distributed in actor model rendering, action rendering etc.；It will need parallelization calculation process
, such as large scene rendering etc., distribute to GPU.
(2) primitive frame figure renders processing and output
It is distributed according to unified task, CPU is because its processing core is relatively fewer, but single arithmetic element processing capacity is strong
Feature, emphasis processing individual task needs a large amount of operations work such as rendering actor model, action/special efficacy renders, using more
Core while treatment mechanism complete serial process, and export relevant rendering result according to priority logic to rendering；GPU is according to it
The characteristics of processing core is more, comprehensive operational capability far super CPU, emphasis processing such as large scene, need to complete big spirogram parallel
The situation elements such as element, object calculate, and are interacted using GPU local memories and bus memory, and share content by GPU and pass to
Each processing cores of GPU carry out parallel processing and calculating to rendering task, and final output renders handling result accordingly.It is handling
In the process, CPU with GPU handling principles are similar, are all based on respective arithmetic element and complete corresponding rendering work, difference just exists
It is few in CPU calculating core, but single core computing capability is strong；GPU calculating core is more, but single core ability is weaker, comprehensive superposition
Ability is superpower.It is based on above-mentioned basis and the single serial a large amount of operand tasks of CPU centralized processings just occurs, at GPU concentrations
Manage the macrooperation amount task of parallel synchronous.It, can also be according to CPU, GPU idle degrees each other, orderly point in actual mechanical process
With respective task, without being confined to mentioned above principle.
(3) graphic result is integrated, sorts
It is that rendering task is registered into row order in resource coordination and hardware assigning process before CPU and GPU is rendered, wash with watercolours
It distributes and renders according to relevant order during dye, in final output result integration process, correspond to corresponding sequence number by figure
As a result it is integrated and is sorted, complete to export according to fluidisation sequence primitive frame figure.
Fig. 5 shows CPU, GPU linkage fluid mapper process figure according to an embodiment of the invention.As shown in figure 5, in this implementation
In example, CPU, GPU linkage fluidization treatment include：The segmentation of primitive frame figure, segmentation fluidization treatment, fluidisation result integration etc..
(1) primitive frame figure is segmented
Using cpu logic processing capacity, the primitive frame figure for rendering output is resolved into different sections in sequence, unification will
CPU, GPU, which merge, becomes unified fluidization treatment resource pool, and overall scheduling task unites different fragments mixed allocation to CPU, GPU
Video coding treatment channel in one resource pool.
(2) it is segmented fluidization treatment
Corresponding CPU, GPU mixed processing resource pool, the primitive frame figure that each processing unit alignment processing is respectively segmented, into
Row segmentation fluidization treatment.
(3) fluidisation result is integrated
The result for being segmented fluidization treatment completion is integrated, forms final Media Stream result.
After process is processed as above in completion, you can the corresponding personage of addition or cartoon character become 3D films or CG results.
Fig. 6 shows the structure chart of one embodiment of CPU and GPU mixed clouds rendering device according to the present invention.Such as Fig. 6 institutes
Show, which includes：The hybrid rending module of linking 61, for rendering task to be decomposed into logic continuous type rendering task and parallel
Logic continuous type rendering task is sent to CPU serial process by rendering task, and parallel rendering task is sent to GPU transports parallel
Calculation is handled, and passes through CPU rendering results and the primitive frame figure of the real-time linkage output generation of GPU rendering results, is generated to rendering
Primitive frame figure integrated and sorted according to fluidisation, export the primitive frame figure of sequence；Dynamic fluidized processing module 62, for inciting somebody to action
The primitive frame figure of sequence from linkage hybrid rending module is segmented, and in order compiles primitive frame figure segment assignments to unified
Coding pass in code treatment source pool carries out segmented fluidisation coding, wherein, Unified coding treatment source pool is compiled by CPU, GPU
Code ability is built-up；The coding result that each coding pass exports is ranked up and is integrally formed Media Stream.Logic is continuous
Type rendering task is for example rendered including actor model, action renders；Parallel rendering task is for example rendered including large scene.
Fig. 7 shows the structure chart of another embodiment of CPU and GPU mixed clouds rendering device according to the present invention.Such as Fig. 7
Shown, in one embodiment, cloud rendering device includes：Link hybrid rending module 71, dynamic fluidized processing module 72 and people
Object angle color embedded module 73.Wherein, character embedded module 73 is used to increase personage in the Media Stream for completing fluidization treatment
Or cartoon character.
In one embodiment, linkage hybrid rending module 71 includes：Rendering task classification scheduling unit 711, for inciting somebody to action
Rendering task is decomposed into logic continuous type rendering task and parallel rendering task, and logic continuous type rendering task is sent to CPU
Parallel rendering task is sent to GPU concurrent operations processing by serial process；Frame artwork render processing unit 712, for being based on
The arithmetic element of CPU and GPU completes corresponding rendering work, and CPU focuses on single serial macrooperation amount rendering task；GPU
Focus on the macrooperation amount rendering task of parallel synchronous；Primitive frame figure is respectively exported after the completion of rendering；Graphic result integrates row
Sequence unit 713 for the primitive frame figure for exporting CPU and GPU hybrid rendings, sorts according to fluidisation sequence integration and sequence output
Primitive frame figure.
In one embodiment, dynamic fluidized processing module 72 includes：Primitive frame figure segmenting unit 721, for that will render
The primitive frame figure of output resolves into each segmentation in sequence；Fluidization treatment unit 722 is segmented, for being fluidized using CPU, GPU
The primitive frame figure correspondence of different fragments is distributed to the coding in unified resource pond by the unified Video coding treatment source pool of ability structure
Treatment channel carries out fluidization treatment, and coded treatment channel respectively exports segment media stream after processing is completed；Result is fluidized to integrate
Unit 723, for integrating the segmentation fluidization treatment of each coded treatment channel output as a result, carrying out being integrally formed final media
Stream.
The treatment mechanism of disclosed technique scheme renders processing procedure using CPU as high in the clouds arithmetic core, in entire cloud and increases
Add GPU concurrent operation advantages, set up the new technique scheme for being more suitable for rendering development.Wherein CPU and GPU joint-action mechanisms, group
It is one of technological core and the basis for realizing CPU and GPU hybrid rendings to build fusion and render integration.With 3D Film Animation wash with watercolours
For dye, it is as follows that GPU+CPU mixed clouds render integrity service main process：
(1) software and hardware resources respond：Hardware resource is called in high in the clouds, starts large-scale rendering basic software, CPU overall coordinations money
Source renders demand according to setting such as rendering quality, rendering duration (start frame -- ending frame), rendering dimension of picture,
It is responsible for overall instruction scheduling, software and hardware resources basis is got out to render.
(2) CPU, GPU linkage hybrid rending：It is one of core of the technical program, the scene based on art designing's modeling
Picture, pixel and person model, according to the rendering requirements set, the process of output and production serial picture, these pictures are made
Primitive frame for final animated video stream.Its key is based on CPU overall tasks distribution mechanisms, and logic continuous type is rendered
Task gives CPU serial process；GPU concurrent operations is given to handle the not strong task of a large amount of logicalities.It is formed using CPU as totally
Coordinate, divide the work according to rendering task and give serial order task to CPU, a large amount of parallel rendering tasks are given into GPU, and pass through
CPU and the real-time linkage of GPU rendering results, more efficiently export corresponding rendering result；
(3) dynamic fluidized processing：By CPU, GPU linkage hybrid rending render generation as a result, will render generate original
Beginning frame transforms into the process of animation streams, and the method for realization is that frame figure is gathered and edited according to corresponding format as animated video stream.Cause
Newest CPU and GPU has the function of more powerful Video coding, by segmenting video stream editor, and the side finally integrated
Formula can effective raising efficiency, reach hardware performance maximization.
(4) character embeds：In finally formed 3D Film Animations, according to plot design requirement, increase personage master
Angle image or cartoon character etc. form final mute.The process, in contrast handle it is fairly simple, without being answered
Miscellaneous CPU, GPU mixing is called, and only need to be based primarily upon CPU according to actual requirement addition and interior according to special-purpose software call-by mechanism
It is embedding.
Above-mentioned (2), (3) step are required to CPU, GPU coordination, complete rendering and fluid mapper process using respective advantage, are structures
It builds CPU, GPU mixed cloud and renders main process.
The disclosure abandons existing single CPU, GPU high in the clouds render mode, makes full use of concentrating type CPU serially and GPU is parallel
Operation advantage with CPU overall coordination resources, by the rendering task with logical relation, gives CPU serial process；It will be extensive
Rendering task without logical relation gives GPU parallel processings.Overall building is suitble to render from tableaux, real-time three-dimensional model
It is rendered into CPU, GPU hybrid rending integration solution of dynamic fluidized output.
So far, CPU and GPU mixed clouds rendering device and method and system according to the present invention is described in detail.For
The design of the masking present invention is avoided, does not describe some details known in the field.Those skilled in the art are according to above
Description, completely it can be appreciated how implementing technical solution disclosed herein.
The method and system of the present invention may be achieved in many ways.For example, can by software, hardware, firmware or
Software, hardware, firmware any combinations come realize the present invention method and system.The said sequence of the step of for the method
Merely to illustrate, the step of method of the invention, is not limited to sequence described in detail above, special unless otherwise
It does not mentionlet alone bright.In addition, in some embodiments, the present invention can be also embodied as recording program in the recording medium, these programs
Including being used to implement machine readable instructions according to the method for the present invention.Thus, the present invention also covering stores to perform basis
The recording medium of the program of the method for the present invention.
Although some specific embodiments of the present invention are described in detail by example, the skill of this field
Art personnel it should be understood that above example merely to illustrating, the range being not intended to be limiting of the invention.The skill of this field
Art personnel are it should be understood that can without departing from the scope and spirit of the present invention modify to above example.This hair
Bright range is defined by the following claims.
Claims (11)
1. a kind of CPU and GPU mixed clouds rendering intent, which is characterized in that including：
Rendering task is decomposed into logic continuous type rendering task and parallel rendering task, by the logic continuous type rendering task
CPU serial process is sent to, the parallel rendering task is sent to GPU carries out concurrent operation processing, and render and tie by CPU
Fruit and the primitive frame figure of the linkage output generation of GPU rendering results；
The primitive frame figure of the generation is integrated and sorted according to fluidisation, exports the primitive frame figure of sequence；
The primitive frame figure of the sequence is segmented, gives primitive frame figure segment assignments to Unified coding treatment source pool in order
In coding pass carry out segmented fluidisation coding, wherein, the Unified coding treatment source pool is by CPU, GPU code capacity structure
It builds；
The coding result that each coding pass exports is ranked up and is integrally formed Media Stream.
2. it according to the method described in claim 1, it is characterized in that, further includes：
Increase personage or cartoon character in the Media Stream for completing fluidization treatment.
3. the according to the method described in claim 1, it is characterized in that, connection by CPU rendering results and GPU rendering results
The primitive frame figure of dynamic output generation includes：
Before CPU and GPU carry out rendering task, CPU is rendering task in resource coordination and hardware assigning process into row order
It registers, distributes and render according to relevant order in render process, export and corresponding sequence number is corresponded in result integration process by figure
Shape result is integrated.
4. according to the method described in claim 1, it is characterized in that, described be sent to the logic continuous type rendering task
CPU serial process includes：
Using multinuclear, treatment mechanism completes serial process, and export relevant wash with watercolours to CPU according to priority logic to rendering task simultaneously
Contaminate result；
The parallel rendering task is sent to GPU progress concurrent operation processing to include：
It is interacted using GPU local memories and bus memory, and shares content by GPU and pass to each processing cores of GPU to rendering
Task carries out parallel processing and calculating, and final output renders handling result accordingly.
5. method as claimed in any of claims 1 to 4, which is characterized in that the logic continuous type rendering task
It is rendered including actor model, action renders；The parallel rendering task is rendered including large scene.
6. a kind of CPU and GPU mixed clouds rendering device, which is characterized in that including：
Link hybrid rending module, will for rendering task to be decomposed into logic continuous type rendering task and parallel rendering task
The logic continuous type rendering task is sent to CPU serial process, and the parallel rendering task is sent at GPU concurrent operations
Reason, and pass through CPU rendering results and the primitive frame figure of the real-time linkage output generation of GPU rendering results, to the original of the generation
Beginning frame figure is integrated and is sorted according to fluidisation, exports the primitive frame figure of sequence；
Dynamic fluidized processing module, for the primitive frame figure of the sequence from the linkage hybrid rending module to be segmented,
Primitive frame figure segment assignments are subjected to segmented fluidisation coding to the coding pass in Unified coding treatment source pool in order,
In, the Unified coding treatment source pool is built-up by CPU, GPU code capacity；The coding knot that each coding pass is exported
Fruit is ranked up and is integrally formed Media Stream.
7. device according to claim 6, which is characterized in that further include：
Character embedded module, for increasing personage or cartoon character in the Media Stream for completing fluidization treatment.
8. device according to claim 6, which is characterized in that the linkage hybrid rending module includes：
Rendering task classification scheduling unit, for rendering task to be decomposed into logic continuous type rendering task and parallel rendering times
Business, is sent to CPU serial process, it is parallel that the parallel rendering task is sent to GPU by the logic continuous type rendering task
Calculation process；
Frame artwork render processing unit, for completing to render work accordingly based on the arithmetic element of CPU and GPU, at CPU concentrations
Manage single serial macrooperation amount rendering task；GPU focuses on the macrooperation amount rendering task of parallel synchronous；It is each after the completion of rendering
From output primitive frame figure；
Graphic result integrates sequencing unit, whole according to fluidisation sequence for the primitive frame figure for exporting CPU and GPU hybrid rendings
It closes and sequence exports the primitive frame figure to sort.
9. device according to claim 6, which is characterized in that the dynamic fluidized processing module includes：
Primitive frame figure segmenting unit, for the primitive frame figure for rendering output to be resolved into each segmentation in sequence；
Fluidization treatment unit is segmented, it, will be different for building unified Video coding treatment source pool using CPU, GPU fluidisation ability
The coded treatment channel that the primitive frame figure correspondence of segment distributes to unified resource pond carries out fluidization treatment, and processing encodes after completing
Treatment channel respectively exports segment media stream；
Result integral unit is fluidized, for integrating the segmentation fluidization treatment of each coded treatment channel output as a result, carrying out integration shape
Into final Media Stream.
10. the device according to any one in claim 6 to 9, which is characterized in that the logic continuous type rendering task
It is rendered including actor model, action renders；The parallel rendering task is rendered including large scene.
11. a kind of cloud rendering system, which is characterized in that mixed including the CPU described in any one in claim 6 to 10 and GPU
Close cloud rendering device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410125332.2A CN104952096B (en) | 2014-03-31 | 2014-03-31 | CPU and GPU mixed clouds rendering intent, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410125332.2A CN104952096B (en) | 2014-03-31 | 2014-03-31 | CPU and GPU mixed clouds rendering intent, device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104952096A CN104952096A (en) | 2015-09-30 |
| CN104952096B true CN104952096B (en) | 2018-06-08 |
Family
ID=54166722
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410125332.2A Active CN104952096B (en) | 2014-03-31 | 2014-03-31 | CPU and GPU mixed clouds rendering intent, device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104952096B (en) |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105427366B (en) * | 2015-11-11 | 2018-07-27 | 广州华多网络科技有限公司 | A kind of image rendering method and image rendering system |
| CN105827700A (en) * | 2016-03-15 | 2016-08-03 | 北京金山安全软件有限公司 | Dynamic file transmission method and device and electronic equipment |
| CN105916023B (en) * | 2016-05-16 | 2019-04-23 | 武汉斗鱼网络科技有限公司 | A kind of Video Rendering method and system based on GPU |
| CN106447595A (en) * | 2016-09-23 | 2017-02-22 | 努比亚技术有限公司 | Distributed image and graph rendering method and terminal |
| CN106534667B (en) * | 2016-10-31 | 2020-02-11 | 努比亚技术有限公司 | Distributed collaborative rendering method and terminal |
| CN106600521A (en) * | 2016-11-30 | 2017-04-26 | 宇龙计算机通信科技(深圳)有限公司 | Image processing method and terminal device |
| CN108289185B (en) * | 2017-01-09 | 2021-08-13 | 腾讯科技（深圳）有限公司 | Video communication method, device and terminal equipment |
| GB2577642B (en) | 2017-04-21 | 2021-05-19 | Zenimax Media Inc | Systems and methods for rendering & pre-encoded load estimation based encoder hinting |
| CN108961380B (en) * | 2017-05-26 | 2022-06-14 | 创新先进技术有限公司 | Graph rendering method and device |
| CN108170526B (en) * | 2017-12-06 | 2021-04-20 | 北京像素软件科技股份有限公司 | Load capacity optimization method and device, server and readable storage medium |
| CN112445605A (en) * | 2019-08-30 | 2021-03-05 | 中兴通讯股份有限公司 | Media data processing method and device and media server |
| CN111190735B (en) * | 2019-12-30 | 2024-02-23 | 湖南大学 | On-chip CPU/GPU pipelining calculation method based on Linux and computer system |
| CN112153411A (en) * | 2020-08-18 | 2020-12-29 | 南斗六星系统集成有限公司 | Method and system for improving performance of chat room streaming media server |
| CN112163986A (en) * | 2020-09-28 | 2021-01-01 | 山东黄金矿业(莱州)有限公司三山岛金矿 | Distributed processing method for underground logging and mining three-dimensional data of metal mine |
| CN112218117B (en) * | 2020-09-29 | 2023-01-03 | 北京字跳网络技术有限公司 | Video processing method and device |
| CN113244614B (en) * | 2021-06-07 | 2021-10-26 | 腾讯科技（深圳）有限公司 | Image picture display method, device, equipment and storage medium |
| CN114140312A (en) * | 2021-12-08 | 2022-03-04 | 稿定（厦门）科技有限公司 | Double-thread rendering method and device and storage medium |
| CN114529444B (en) * | 2022-04-22 | 2023-08-11 | 南京砺算科技有限公司 | Graphics processing module, graphics processor, and graphics processing method |
| CN115375530B (en) * | 2022-07-13 | 2024-07-09 | 北京松应科技有限公司 | Multi-GPU collaborative rendering method, system, device and storage medium |
| CN116433818B (en) * | 2023-03-22 | 2024-04-16 | 宝钢工程技术集团有限公司 | Cloud CPU and GPU parallel rendering method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101587583A (en) * | 2009-06-23 | 2009-11-25 | 长春理工大学 | The farm of playing up based on the GPU cluster |
| CN102609971A (en) * | 2012-01-11 | 2012-07-25 | 南京大学 | Quick rendering system using embedded GPU (Graphics Processing Unit) for realizing 3D-GIS (Three Dimensional-Geographic Information System) |
| CN102663208A (en) * | 2012-04-28 | 2012-09-12 | 浪潮电子信息产业股份有限公司 | Method for carrying out large-eddy simulation real-time rendering by utilizing graphic processing unit (GPU) |
| CN102959517A (en) * | 2010-06-10 | 2013-03-06 | Otoy公司 | Allocation of gpu resources accross multiple clients |
| CN103268220A (en) * | 2012-02-24 | 2013-08-28 | 苏州蓝海彤翔系统科技有限公司 | Software architecture suitable for large-scale animation rendering service cloud platform |
| CN103631634A (en) * | 2012-08-24 | 2014-03-12 | 中国电信股份有限公司 | Graphics processor virtualization achieving method and device |
-
2014
- 2014-03-31 CN CN201410125332.2A patent/CN104952096B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101587583A (en) * | 2009-06-23 | 2009-11-25 | 长春理工大学 | The farm of playing up based on the GPU cluster |
| CN102959517A (en) * | 2010-06-10 | 2013-03-06 | Otoy公司 | Allocation of gpu resources accross multiple clients |
| CN102609971A (en) * | 2012-01-11 | 2012-07-25 | 南京大学 | Quick rendering system using embedded GPU (Graphics Processing Unit) for realizing 3D-GIS (Three Dimensional-Geographic Information System) |
| CN103268220A (en) * | 2012-02-24 | 2013-08-28 | 苏州蓝海彤翔系统科技有限公司 | Software architecture suitable for large-scale animation rendering service cloud platform |
| CN102663208A (en) * | 2012-04-28 | 2012-09-12 | 浪潮电子信息产业股份有限公司 | Method for carrying out large-eddy simulation real-time rendering by utilizing graphic processing unit (GPU) |
| CN103631634A (en) * | 2012-08-24 | 2014-03-12 | 中国电信股份有限公司 | Graphics processor virtualization achieving method and device |
Non-Patent Citations (3)
| Title |
|---|
| GPU异构平台下的深度包检测技术;张首斌等;《天津理工大学学报》;20130228;第29卷(第1期);第41-44页 * |
| 关于CPU+GPU异构计算的研究与分析;许帧;《科技信息》;20101231(第17期);第613、530页 * |
| 基于GPU加速的SDOG并行可视化实验;薄海光等;《地理与地理信息科技》;20130731;第29卷(第4期);第72-76页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104952096A (en) | 2015-09-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104952096B (en) | CPU and GPU mixed clouds rendering intent, device and system | |
| US8497865B2 (en) | Parallel graphics system employing multiple graphics processing pipelines with multiple graphics processing units (GPUS) and supporting an object division mode of parallel graphics processing using programmable pixel or vertex processing resources provided with the GPUS | |
| KR101231291B1 (en) | Fast reconfiguration of graphics pipeline state | |
| DE102020124932A1 (en) | Apparatus and method for real-time graphics processing using local and cloud-based graphics processing resources | |
| US7233331B2 (en) | Parallel object task engine and processing method | |
| CN106575431A (en) | Method and apparatus for highly efficient graphics processing unit (GPU) execution model | |
| DE112017001703T5 (en) | Method and apparatus for more efficiently ray tracing instantiated geometry | |
| CN108734644A (en) | Enhanced using the HDR of time-multiplexed progress | |
| CN110060320A (en) | Animation producing method and device based on WEBGL | |
| CN107077833B (en) | The device and method that efficient frame for the framework that finally sorts is utilized to frame coherence | |
| DE102019110027A1 (en) | TILE-BASED TRANSMITTERS FOR SEVERAL RESOLUTIONS OF PICTURES | |
| CN108694376A (en) | Including static scene determination, occlusion detection, frame rate conversion and the video motion processing for adjusting compression ratio | |
| CN109644260A (en) | The mixing solid rendering of Depth Expansion is carried out in dynamic light field display | |
| CN102819855A (en) | Method and device for generating two-dimensional images | |
| EP1152331B1 (en) | Parallel task processing system and method | |
| Fussell et al. | A vlsi-oriented architecture for real-time raster display of shaded polygons | |
| US20230316626A1 (en) | Image rendering method and apparatus, computer device, and computer-readable storage medium | |
| Liu et al. | Design and implementation of distributed rendering system | |
| DE102019128017A1 (en) | WORKLOAD MANAGEMENT FOR DISTRIBUTED GEOMETRY PROCESSING | |
| Hu et al. | Research on 3d interactive model selection and customization of ceramic products based on big data cloud service platform | |
| Levus et al. | Using software agents in a distributed computing system for procedural planetoid terrain generation | |
| Wang et al. | LVDIF: a framework for real-time interaction with large volume data | |
| KR20180008168A (en) | Holographic Content Creation System and Method using a Parallel Processing Environment | |
| Streit et al. | Understanding next-generation VR: classifying commodity clusters for immersive virtual reality | |
| Sharmila et al. | Neural Graphics Primitive–Meta Human in Unreal Engine Using NeRF [Neural Radiance Field] |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |