CN105045658A - Method for realizing dynamic dispatching distribution of task by multi-core embedded DSP (Data Structure Processor) - Google Patents
Method for realizing dynamic dispatching distribution of task by multi-core embedded DSP (Data Structure Processor) Download PDFInfo
- Publication number
- CN105045658A CN105045658A CN201510381740.9A CN201510381740A CN105045658A CN 105045658 A CN105045658 A CN 105045658A CN 201510381740 A CN201510381740 A CN 201510381740A CN 105045658 A CN105045658 A CN 105045658A
- Authority
- CN
- China
- Prior art keywords
- core
- event
- openem
- multicore
- multinuclear
- 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.)
- Granted
Links
Abstract
The invention discloses a method for realizing the dynamic dispatching distribution of a task by a multicore embedded DSP (Data Structure Processor). A KeyStone platform brought out by TI (Texas Instruments) provides a multicore runtime system library OpenEM (Event Machine) capable of realizing the dynamic dispatching distribution of the task on the basis of Multicore Navigator, the dynamic dispatching distribution of the task can be realized through the multicore runtime system library OpenEM, and multicore load balance is realized, wherein the multicore runtime system library OpenEM is independent of an operating system. A DSP core of a multicore embedded processor based on a KeyStone architecture is divided into a main core and slave cores, wherein the main core finishes the global initialization of a programming model, and all cores finish local initialization. The programming model consists of a main core generation event, an event driver, an OpenEM dispatching distribution event and a slave core processing event. The invention provides a uniform parallel programming model of the multicore embedded DSP on the basis of the OpenEM for embedded software developers. The implementation method is high in expansibility, can be suitable for the majority of multicore or many-core embedded processors based on a KeyStone architecture and can meet the application requirement of the dispatching distribution of the task under a multicore environment and realize the dynamic load balance.
Description
Technical field
The present invention relates to the multiple programming field of multinuclear embedded system, specifically a kind of method utilizing multinuclear DSP embedded to realize dynamic task scheduling distribution.
Background technology
Along with the develop rapidly of embedded technology, embedded processing demand is also in quick growth, in today of integrated circuit technique develop rapidly, multi-core technology application is in embedded systems also more and more extensive, multi-processor cooperation is used to finish the work in embedded systems, for improving the embedded performance of system and meeting the real-time demand etc. of embedded system, have great importance.And multinuclear embedded system platform has also become current main computing platform, desktop application, Mobile solution, server or special embedded platform have all started to adopt coenocytism.
The mainstreaming of multi-core technology, all creates important impact to the exploitation of parallel computation architecture, parallel algorithm, parallel programming model and Parallel application.Traditional computing power that obviously cannot make full use of multi-core CPU based on single-threaded Programming Methodology, in multinuclear embedded system, algorithm for design must be carried out with the mode of thinking of parallel computation, the hardware superiority of multinuclear embedded platform and parallel Programming methods combining are got up, obtains higher system performance.Therefore, how effectively to utilize platform resource, scheduling multinuclear is one of the emphasis and hot issue that nowadays develop and study efficiently.Operating system under current multi-core environment can complete certain multi-core dispatching task, but, inapplicable under some situations, especially when the divided poling of task multiple little task block time, the scheduling overhead of operating system is too high, and be not suitable for allowing a thread go to perform the task of needing to spend several thousand clock period, be not suitable for too creating hundreds and thousands of threads.That is, find one not rely on operating system, the implementation method of efficient scheduler task is very necessary.
Application number is that the patent of invention of CN201510002039 discloses a kind of optimum localization tasks dispatching method based on MapReduce.This invention proposes a kind of MapReduce task scheduling algorithm that may be simultaneously operated under isomorphism and isomeric group environment, belongs to field of computer technology.This dispatching algorithm considers the handling property of each computing node in cluster, computing node and calculation task abstract be a bipartite graph, by suitably expanding this bipartite graph and forming final overall task scheduling approach in conjunction with KM cum rights Optimum Matching algorithm.But this invention is mainly for the task scheduling between cluster environment lower node, not for the task scheduling under multi-core environment between multinuclear, do not make full use of the multiple hardwares resource in embedded multi-core processor.
Application number is the resource allocation methods that the patent of invention of CN201410610548 discloses the restructural polycaryon processor of a kind of task number and performance aware.In this invention, in each operating system scheduling interval, resource allocator is first according to the number mean allocation Logic Core of task, after running a clock cycle, performance (reflection task is to the demand of resource) according to task sorts to it, find out the task little to resource requirement, reduce the granularity of shared Logic Core, and the free physical core obtained from the task there that resource requirement is little distributes to the task high to resource requirement, to increase the granularity of the Logic Core that this task high to resource requirement takies.When the load that system is current changes or task itself enters the new operation phase, resource allocator will adjust in time to make full use of resources of chip in next operation system call.But this invention is the strategy implemented in the schedule gaps of operating system, be therefore the scheduling based on operating system, separating system, just no longer possesses basic scheduling feature.
Summary of the invention
The object of the present invention is to provide a kind of method realizing Real-time Task Dispatch, distribution based on multinuclear DSP embedded, the method is based on the component N avigator on flush bonding processor, make full use of its various hardware resource, realize dynamic dispatching and the distribution of task, reach load balance, to solve the problem proposed in above-mentioned background technology.
For achieving the above object, the invention provides following technical scheme:
Utilize multinuclear DSP embedded to realize a method for dynamic task scheduling distribution, comprise following content:
1) the KeyStone platform that TI releases provides OpenEM(and OpenEventMachine during the multinuclear operation that a dynamic task scheduling based on MulticoreNavigator independent of operating system distribute), realized the dynamic dispatching distribution of task by OpenEM, realize the load balance on multinuclear;
The QMSS mainly comprising hardware management queue in MulticoreNavigator, the telecommunication media PKTDMA realizing actual data communication, load the PDSP that different firmware realizes different functions; Wherein PDSP is RSIC processor; Based on the OpenEM of MulticoreNavigator, comprise scheduler and distributor; Described scheduler operates on PDSP, gives current idle core according to certain scheduling strategy by event scheduling; Distributor operates on each core, inquires about pending event and calls corresponding process function; Customer incident is carried by QMSS, uses the encapsulation of the descriptor completing user data message in QMSS and puts into hardware queue as OpenEM event; Process function is encapsulated in EO structure;
2) the DSP core of multinuclear flush bonding processor under KeyStone framework is divided into main core and from core, described main core completes the overall initialization of programming model, this initialize flow comprises the configuration to internal memory CACHE, the configuration of PDSP, QMSS, PKTDMA and hardware queue thereof, user processes the encapsulation of function and the initialization to environmental variance each in parallel programming model; The overall initialization of described main core is shared with from core; All core completes local initialization;
3) programming model is produced event, event-driven, OpenEM scheduling Distribution Events by main core, is formed from core process event
Main core classification by function, obtains a null event, is filled in by user data in event, then event is put into the hardware queue of specifying from event pond; Scheduler by monitoring hardware queue, finds the arrival of event, by the mode of interrupting by event scheduling to the light core of load and by PKTDMA by event transmission to the object hardware queue of object core; From core once receive an event, from extract corresponding user the EO structure that hardware queue is associated and process function, and then invoke user process function realizes the process to event, then deletes processed event, discharges its space.
As the further scheme of the present invention: comprise 8192 hardware queues in QMSS altogether in step 1), use the wherein specific hardware queue for sending and the hardware queue with priority complete internuclear synchronously with communicate, realize the distribution of OpenEM event.
As the further scheme of the present invention: in step 1), PKTDMA comprises RxDMA passage, TxDMA passage and Rxflow passage, the transmission operation of descriptor is completed by the TxDMA passage in PKTDMA and RxDMA passage.
As the further scheme of the present invention: step 2) in multinuclear flush bonding processor be isomorphism or heterogeneous processor, the quantity of described DSP core is at least 8, and wherein, described main core is core 0, described from core be other cores except core 0.
As the further scheme of the present invention: in step 3), scheduler adopts the scheduling strategy of LAZY, namely by sending request from core to main core, illustrate when pronucleus is idle, main core receives request will by pending case distribution to request core.
As the further scheme of the present invention: in step 3), the process of event handling is the pattern of run-to-complete.
Compared with prior art, the invention has the beneficial effects as follows:
1, the invention provides and unify parallel programming model to a kind of multinuclear DSP embedded based on OpenEM of embedded software developing personnel;
2, the parallel programming model implementation method extendability of the present invention's proposition is very strong, go for most of multinuclear based on KeyStone framework or many core flush bonding processors at present, the scheduling distribution of task under multi-core environment can be met, the application demand of balancing dynamic load.
3, fully employ the hardware resource of flush bonding processor in the present invention, efficiently dispatch distributed tasks, realize the load balance between multinuclear.
Accompanying drawing explanation
Fig. 1 is OpenEM general frame figure of the present invention;
Fig. 2 is that in the present invention, user processes function encapsulation process figure;
Fig. 3 is state-event transition diagram in the present invention;
Fig. 4 is serial parallel event model figure in the present invention;
Fig. 5 is image procossing applicating flow chart in the present invention;
Fig. 6 is image procossing application testing result figure in the present invention, each core process event number;
Fig. 7 is image procossing application testing result figure in the present invention, each core average treatment cycle number.
Fig. 8 is process flow diagram of the present invention.
Embodiment
Below in conjunction with the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making other embodiments all obtained under creative work prerequisite, belong to the scope of protection of the invention.
Refer to Fig. 8, in the present invention, OpenEM is a multinuclear runtime system storehouse realizing task scheduling distribution, and OpenEM does not rely on operating system, can naked race on chip, can run on operating system yet.Provide the scheduling of the task for a large amount of ratio processes, thread smaller particle size beyond operating system.Can efficient scheduling distributed tasks on multinuclear, realize dynamic load balance on multinuclear by task being distributed to the light core of load.
MulticoreNavigator(is called for short Navigator) be a kind of hardware mechanisms, help the collaborative work realized between the movement of data and multinuclear.Mainly provide internuclear, communication between network and between peripheral hardware, comprise data and message, after sending message, the operation that reprocessing is not relevant to this message, that is after a message has been sent, whether transmit leg is received regardless of message.Briefly, be exactly only need to be loaded into data, all the other operations have all been responsible for by Navigator, voluntarily without the need to the intervention of CPU.Navigator is primarily of QueueManagerSubsystem(QMSS) and multiple PacketDMA(PKTDMA) and two PDSP(PackedDataStructureProcessors) composition.
QueueManager is a hardware module of being responsible for carrying out hardware queue acceleration management.Comprising 8192 hardware queues, different hardware queue purposes is different.Write descriptor (Descriptor) address of 32 by specific mapped position in module, a bag can be added in hardware queue.On the contrary, from then on the same position of hardware queue reads value, then can complete contracting out team.Descriptor (Descriptor) is the message with information and data in internuclear movement.The data carry of user just can be encapsulated into an event on descriptor.Namely PKTDMA is the movement of a kind of DMA, primary responsibility data.PKTDMA comprises many RxDMA passages, many TxDMA passages and many Rxflow passages.
PDSP is a kind of RSIC processor, can realize accumulation, monitoring or the function of service quality by loading firmware.Can monitor hardware queue and the descriptor in hardware queue is accumulated, can also send to core or peripheral hardware and interrupt.In OpenEM, PDSP upper loading scheduler firmware, the descriptor in hardware queue by the monitoring to hardware queue, and is dispatched to the light core of load by certain scheduling strategy by PDSP.Two PDSP are also incomplete same, need to load different firmwares and realize identical function.
Multinuclear flush bonding processor is isomorphism or the heterogeneous processor adopting KeyStone framework generally more, and the quantity of DSP core is at least 8, and wherein, using core 0 as main core, other core except core 0 is as from core.Core 0 as main core completes the overall initialization of OpenEM environment, comprising the configuration to buffer memory, to the initialization of event pond and event, to the initialization of each component of MulticoreNavigator, user is processed to the encapsulation of function and the initialization to environmental variance each in parallel programming model.Of overall importance also with from core the sharing during initialization of core 0.Only initialized for main core environmental variance copy portion need be placed on local use from core initialization.
First main core be configured buffer memory, can improve operational efficiency by the switch controlling caching function.Then according to the functional requirement of user, event pond is set up to different demands, as the deposit space of each function.In event pond, mark off event, when a needs production event, from event pond, take out a null event fill.Main core also needs initialization QMSS and opens the hardware queue of configuration quantity, creates and the communication port opened in PKTDMA and be PDSP loading scheduler firmware.Final step, will process function by user and be encapsulated in EO(Executionobject) in structure, and associate with EMQueue.
Wait for that main core completes the initialization of the overall situation from core, then copy initialized for main core environmental variance to this locality.
For the initialization of environmental variance in parallel programming model, mainly initialization is carried out to the various global variables used in model.All tasks carrying core is all needed to the variable shared, carry out initialization by main core, other core directly shares this variable, and the global variable in all core is consistent, and in order to realize zero-copy thus improve execution efficiency, global variable is stored in the sharable region of memory of multinuclear.
Described main core completes generation to the process encapsulation of function and event, encapsulation and triggering.Describedly obtain pending event from core, extract the process that corresponding process function completes event.The present invention is by event-driven, and event experience is from being created to the process being scheduled for and being processed to release.The process of event handling is the pattern of run-to-complete, and namely the process of event can not be suspended, until an event complete the event that can not process other.
Embodiment 1
In the embodiment of the present invention, choose image procossing application realize model and test, simultaneously, the TMS320C6678(using TI to release is called for short C6678) processor is as implementation platform, C6678 is the high-performance multi-core DSP based on KeyStone framework, Embedded 8 C66x cores, merge fixed point and floating-point processing capacity, comprise component N avigator.
As shown in Figure 1, the scheduler of OpenEM operates on risc processor, and distributor operates on each core.Core is to scheduler dispatches event request, the queue of scheduler monitoring hardware, takes out pending event from hardware queue, gives distributor by the event of scheduling, give corresponding EO structure by distributor again, utilize and wherein encapsulate volume user and process function event is processed.EO structure is wherein shared by all core by unified establishment of main core.An EO structure can only encapsulate a user and process function.
As shown in Figure 2, user processes the encapsulation process of function, first will create an EO structure, and in EO structure, indicates user process function.Then hardware queue is created, here hardware queue is not the hardware queue in QMSS, neither software construct queue, its Main Function is mapped event and EO structure, thus the user in the EO structure that the event be sent in a hardware queue just can be mapped processes function process.Therefore, the 3rd step needs created EO structure and hardware queue to associate.Finally start EO structure.
As shown in Figure 3, whole model is by event-driven, null event is by initialization in event pond, therefore event is in the state of free, user calls em_alloc and namely distributes a new events, then system can be taken out first null event and be returned to user from null event queue, and at this moment event is in preparing state.User obtains a new events, and fill user data, be then sent in hardware queue by em_send by event, now event transition is ready state.Now, if having core to scheduler dispatches request, namely scheduling can be in the event delivery of ready state to request core, and request core calls corresponding user and processes function and process, event and then become running state.When completing the process of event, user needs release event space, calls em_free, thus event returns to again the state of free, can be multiplexing in next event.
As shown in Figure 4, when user's algorithm is larger, multiple process function may be divided into, the corresponding a kind of event type of each process function.Wherein process between function and may there is certain front and back dependence, now the process function of Existence dependency relationship can be defined as series process, and without the function of dependence, being then defined as can independent parallel.User processes the relation between function, programming model is passed to by the description definition of user structure body, first all process functions encapsulate by model, secondly when triggering, first event in all parallel events and all series processes is all put into hardware queue, thus these event concurrency process.When an event handling is near completion time, judge whether this event has successor, namely whether have the event of connecting with it to need serial processing.If nothing, then terminate the process of current event, if having, then in current event, trigger the next event of serial.The event of serial can be carried out scheduling distribution processor by system as the new events that is tantamount to other event.That is, the event in series process can't be bound to same core, but may by different core process.
As shown in Figure 5, the availability of test model and performance should be used for by realizing image procossing.Main core reads original image, and carries out initialization to user structure body.Then user structure body is given the model initialization that model carries out OpenEM.Then directly model initialization is carried out from core.All users are processed function at initial phase by main core, and namely split, dehaze, combine are encapsulated in EO structure.And according to the parameter in user structure body, be user's dispense event, and by user data carry in the middle of event, then trigger event.After all core all completes initialization, just enter model processing stage.All core participating in process all can to scheduler dispatches event request, and scheduler obtains pending event, is distributed to each core.Each core receives event, and the process function called in EO structure carries out the work of split, dehaze and combine successively.Namely three process functions in this test routine are the relations of serial, namely first will calling split segmentation to original image, then call the image of dehaze to a slice and process, finally calling combine by a picture of process from being newly tied.Therefore, only need trigger all split events first, when split closes to an end time, then trigger dehaze, same, trigger combine by dehaze.This wherein, all core all completes the process of split, dehaze, combine.But all events are all divide scheduling to be distributed to the light core of load by scheduler, that is, as the dehaze that triggered by main core split might not by main core processing, and may by other from core process.Equally, the dehaze triggered by core 7dehaze is processed by core 7, and may by other core process.
As Figure 6-Figure 7, the present invention is application image process when testing, and gives model, carry out split, dehaze, combine by model by original image.In test, choose the pending image of 512384, and Iamge Segmentation is 12 pieces and carries out parallel processing by setting.As shown in Figure 6, can find out that number that each core completes event accounts for the number percent of total event number, Fig. 7 describes the average treatment event of each core, as can be seen from histogram, the average treatment event of each core remains basically stable, difference is little, thus the balancing dynamic load on the multinuclear reflecting course of event scheduling distribution and reach.
Through above-mentioned realization, image procossing application can be realized by parallel programming model and operate in embedded multi-core DSP platform, not only can pass through the use improving performance of hardware resource, and this implementation method have very strong extensibility.
To those skilled in the art, obviously the invention is not restricted to the details of above-mentioned one exemplary embodiment, and when not deviating from spirit of the present invention or essential characteristic, the present invention can be realized in other specific forms.Therefore, no matter from which point, all should embodiment be regarded as exemplary, and be nonrestrictive, scope of the present invention is limited by claims instead of above-mentioned explanation, and all changes be therefore intended in the implication of the equivalency by dropping on claim and scope are included in the present invention.
In addition, be to be understood that, although this instructions is described according to embodiment, but not each embodiment only comprises an independently technical scheme, this narrating mode of instructions is only for clarity sake, those skilled in the art should by instructions integrally, and the technical scheme in each embodiment also through appropriately combined, can form other embodiment that it will be appreciated by those skilled in the art that.
Claims (6)
1. utilize multinuclear DSP embedded to realize a method for dynamic task scheduling distribution, it is characterized in that, comprise following content:
1) the KeyStone platform that TI releases provides the OpenEM during multinuclear operation that a dynamic task scheduling based on MulticoreNavigator independent of operating system distribute, realized the dynamic dispatching distribution of task by OpenEM, realize the load balance on multinuclear;
The QMSS mainly comprising hardware management queue in MulticoreNavigator, the telecommunication media PKTDMA realizing actual data communication, load the PDSP that different firmware realizes different functions; Wherein PDSP is RSIC processor; Based on the OpenEM of MulticoreNavigator, comprise scheduler and distributor; Described scheduler operates on PDSP, gives current idle core according to certain scheduling strategy by event scheduling; Distributor operates on each core, inquires about pending event and calls corresponding process function; Customer incident is carried by QMSS, uses the encapsulation of the descriptor completing user data message in QMSS and puts into hardware queue as OpenEM event; Process function is encapsulated in EO structure;
2) the DSP core of multinuclear flush bonding processor under KeyStone framework is divided into main core and from core, described main core completes the overall initialization of programming model, this initialize flow comprises the configuration to internal memory CACHE, the configuration of PDSP, QMSS, PKTDMA and hardware queue thereof, user processes the encapsulation of function and the initialization to environmental variance each in parallel programming model; The overall initialization of described main core is shared with from core; All core completes local initialization;
3) programming model produces event by event-driven, main core, OpenEM dispatches Distribution Events, forms from core process event
Main core classification by function, obtains a null event, is filled in by user data in event, then event is put into the hardware queue of specifying from event pond; Scheduler by monitoring hardware queue, finds the arrival of event, by the mode of interrupting by event scheduling to the light core of load and by PKTDMA by event transmission to the object hardware queue of object core; From core once receive an event, from extract corresponding user the EO structure that hardware queue is associated and process function, and then invoke user process function realizes the process to event, then deletes processed event, discharges its space.
2. the method utilizing multinuclear DSP embedded to realize dynamic task scheduling distribution according to claim 1, it is characterized in that, 8192 hardware queues are comprised in QMSS altogether in step 1), use the wherein specific hardware queue for sending and the hardware queue with priority complete internuclear synchronously with communicate, realize the distribution of OpenEM event.
3. the method utilizing multinuclear DSP embedded to realize dynamic task scheduling distribution according to claim 1, it is characterized in that, in step 1), PKTDMA comprises RxDMA passage, TxDMA passage and Rxflow passage, and the transmission operation of descriptor is completed by the TxDMA passage in PKTDMA and RxDMA passage.
4. the method utilizing multinuclear DSP embedded to realize dynamic task scheduling distribution according to claim 1, it is characterized in that, step 2) in multinuclear flush bonding processor be isomorphism or heterogeneous processor, the quantity of described DSP core is at least 8, wherein, described main core is core 0, described from core be other cores except core 0.
5. the method utilizing multinuclear DSP embedded to realize dynamic task scheduling distribution according to claim 1, it is characterized in that, in step 3), scheduler adopts the scheduling strategy of LAZY, namely by sending request from core to main core, illustrate when pronucleus is idle, main core receives request will by pending case distribution to request core.
6. the method utilizing multinuclear DSP embedded to realize dynamic task scheduling distribution according to claim 1, it is characterized in that, in step 3), the process of event handling is the pattern of run-to-complete.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510381740.9A CN105045658B (en) | 2015-07-02 | 2015-07-02 | A method of realizing that dynamic task scheduling is distributed using multinuclear DSP embedded |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510381740.9A CN105045658B (en) | 2015-07-02 | 2015-07-02 | A method of realizing that dynamic task scheduling is distributed using multinuclear DSP embedded |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105045658A true CN105045658A (en) | 2015-11-11 |
| CN105045658B CN105045658B (en) | 2018-10-23 |
Family
ID=54452222
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510381740.9A Active CN105045658B (en) | 2015-07-02 | 2015-07-02 | A method of realizing that dynamic task scheduling is distributed using multinuclear DSP embedded |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105045658B (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106851296A (en) * | 2015-12-04 | 2017-06-13 | 宁波舜宇光电信息有限公司 | Image processing system and image processing method based on embedded platform |
| CN107302570A (en) * | 2017-06-09 | 2017-10-27 | 东华大学 | A kind of monitoring of tools cloud component design method based on Priority Queues and Canvas technologies |
| CN107357666A (en) * | 2017-06-26 | 2017-11-17 | 西安微电子技术研究所 | A kind of multi-core parallel concurrent system processing method based on hardware protection |
| CN107608784A (en) * | 2017-06-28 | 2018-01-19 | 西安微电子技术研究所 | A kind of multi-modal dispatching method of mass data flow under multi-core DSP |
| CN107832129A (en) * | 2017-10-24 | 2018-03-23 | 华中科技大学 | A kind of dynamic task scheduling optimization method of Based on Distributed stream calculation system |
| CN108958905A (en) * | 2017-05-25 | 2018-12-07 | 北京忆恒创源科技有限公司 | The lightweight operating system of embedded multi-core central processing unit |
| CN108958904A (en) * | 2017-05-25 | 2018-12-07 | 北京忆恒创源科技有限公司 | The driver frame of the lightweight operating system of embedded multi-core central processing unit |
| CN109144691A (en) * | 2018-07-13 | 2019-01-04 | 哈尔滨工程大学 | A kind of task scheduling and allocating method of multi-core processor oriented |
| CN109508231A (en) * | 2018-11-17 | 2019-03-22 | 中国人民解放军战略支援部队信息工程大学 | Synchronous method and device between the equivalents of isomery multi-mode processor |
| CN109558226A (en) * | 2018-11-05 | 2019-04-02 | 上海无线通信研究中心 | A kind of DSP multi-core parallel concurrent calculating dispatching method based on internuclear interruption |
| CN109905898A (en) * | 2017-12-07 | 2019-06-18 | 北京中科晶上科技股份有限公司 | Baseband processing resource distribution method |
| CN110347504A (en) * | 2019-06-28 | 2019-10-18 | 中国科学院空间应用工程与技术中心 | Many-core computing resource dispatching method and device |
| CN112243266A (en) * | 2019-07-18 | 2021-01-19 | 大唐联仪科技有限公司 | Data packaging method and device |
| CN112491426A (en) * | 2020-11-17 | 2021-03-12 | 中国人民解放军战略支援部队信息工程大学 | Service assembly communication architecture and task scheduling and data interaction method facing multi-core DSP |
| CN112486681A (en) * | 2020-11-26 | 2021-03-12 | 迈普通信技术股份有限公司 | Communication method and network equipment |
| CN112859753A (en) * | 2021-01-19 | 2021-05-28 | 深圳市汇川技术股份有限公司 | Secondary development method, device and equipment for numerical control system and readable storage medium |
| WO2022141297A1 (en) * | 2020-12-30 | 2022-07-07 | 华为技术有限公司 | Event processing method and apparatus |
| CN114741137A (en) * | 2022-05-09 | 2022-07-12 | 潍柴动力股份有限公司 | Software starting method, device, equipment and storage medium based on multi-core microcontroller |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010244332A (en) * | 2009-04-07 | 2010-10-28 | Nec Corp | Means of task assignment for multi-core system, method of the same, and program of the same |
| US20140245308A1 (en) * | 2013-02-25 | 2014-08-28 | Texas Instruments Incorporated | System and method for scheduling jobs in a multi-core processor |
| CN104331331A (en) * | 2014-11-02 | 2015-02-04 | 中国科学技术大学 | Resource distribution method for reconfigurable chip multiprocessor with task number and performance sensing functions |
| CN104572483A (en) * | 2015-01-04 | 2015-04-29 | 华为技术有限公司 | Device and method for management of dynamic memory |
-
2015
- 2015-07-02 CN CN201510381740.9A patent/CN105045658B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010244332A (en) * | 2009-04-07 | 2010-10-28 | Nec Corp | Means of task assignment for multi-core system, method of the same, and program of the same |
| US20140245308A1 (en) * | 2013-02-25 | 2014-08-28 | Texas Instruments Incorporated | System and method for scheduling jobs in a multi-core processor |
| CN104331331A (en) * | 2014-11-02 | 2015-02-04 | 中国科学技术大学 | Resource distribution method for reconfigurable chip multiprocessor with task number and performance sensing functions |
| CN104572483A (en) * | 2015-01-04 | 2015-04-29 | 华为技术有限公司 | Device and method for management of dynamic memory |
Non-Patent Citations (2)
| Title |
|---|
| FILIP MOERMAN: "Open event machine:A multi-core run-time designed for performance", 《PROCEEDINGS OF THE 6TH EUROPEAN EMBEDDED DESIGN IN EDUCATION AND RESEARCH》 * |
| 周梦: "基于多核嵌入式DSP的并行编程模型研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106851296A (en) * | 2015-12-04 | 2017-06-13 | 宁波舜宇光电信息有限公司 | Image processing system and image processing method based on embedded platform |
| CN108958905A (en) * | 2017-05-25 | 2018-12-07 | 北京忆恒创源科技有限公司 | The lightweight operating system of embedded multi-core central processing unit |
| CN108958905B (en) * | 2017-05-25 | 2024-04-05 | 北京忆恒创源科技股份有限公司 | Lightweight operating system of embedded multi-core central processing unit |
| CN108958904B (en) * | 2017-05-25 | 2024-04-05 | 北京忆恒创源科技股份有限公司 | Driver framework of lightweight operating system of embedded multi-core central processing unit |
| CN108958904A (en) * | 2017-05-25 | 2018-12-07 | 北京忆恒创源科技有限公司 | The driver frame of the lightweight operating system of embedded multi-core central processing unit |
| CN107302570B (en) * | 2017-06-09 | 2020-05-26 | 东华大学 | Equipment monitoring cloud component design method based on priority queue and Canvas technology |
| CN107302570A (en) * | 2017-06-09 | 2017-10-27 | 东华大学 | A kind of monitoring of tools cloud component design method based on Priority Queues and Canvas technologies |
| CN107357666A (en) * | 2017-06-26 | 2017-11-17 | 西安微电子技术研究所 | A kind of multi-core parallel concurrent system processing method based on hardware protection |
| CN107608784A (en) * | 2017-06-28 | 2018-01-19 | 西安微电子技术研究所 | A kind of multi-modal dispatching method of mass data flow under multi-core DSP |
| CN107608784B (en) * | 2017-06-28 | 2020-06-09 | 西安微电子技术研究所 | Multi-mode scheduling method for mass data stream under multi-core DSP |
| CN107832129A (en) * | 2017-10-24 | 2018-03-23 | 华中科技大学 | A kind of dynamic task scheduling optimization method of Based on Distributed stream calculation system |
| CN109905898A (en) * | 2017-12-07 | 2019-06-18 | 北京中科晶上科技股份有限公司 | Baseband processing resource distribution method |
| CN109144691A (en) * | 2018-07-13 | 2019-01-04 | 哈尔滨工程大学 | A kind of task scheduling and allocating method of multi-core processor oriented |
| CN109144691B (en) * | 2018-07-13 | 2021-08-20 | 哈尔滨工程大学 | Task scheduling and distributing method for multi-core processor |
| CN109558226A (en) * | 2018-11-05 | 2019-04-02 | 上海无线通信研究中心 | A kind of DSP multi-core parallel concurrent calculating dispatching method based on internuclear interruption |
| CN109558226B (en) * | 2018-11-05 | 2021-03-30 | 上海无线通信研究中心 | DSP multi-core parallel computing scheduling method based on inter-core interruption |
| CN109508231B (en) * | 2018-11-17 | 2020-09-18 | 中国人民解放军战略支援部队信息工程大学 | Synchronization method and device between equivalents of heterogeneous multimode processors |
| CN109508231A (en) * | 2018-11-17 | 2019-03-22 | 中国人民解放军战略支援部队信息工程大学 | Synchronous method and device between the equivalents of isomery multi-mode processor |
| CN110347504A (en) * | 2019-06-28 | 2019-10-18 | 中国科学院空间应用工程与技术中心 | Many-core computing resource dispatching method and device |
| CN112243266A (en) * | 2019-07-18 | 2021-01-19 | 大唐联仪科技有限公司 | Data packaging method and device |
| CN112243266B (en) * | 2019-07-18 | 2024-04-19 | 大唐联仪科技有限公司 | Data packet method and device |
| CN112491426A (en) * | 2020-11-17 | 2021-03-12 | 中国人民解放军战略支援部队信息工程大学 | Service assembly communication architecture and task scheduling and data interaction method facing multi-core DSP |
| CN112491426B (en) * | 2020-11-17 | 2022-05-10 | 中国人民解放军战略支援部队信息工程大学 | Service assembly communication architecture and task scheduling and data interaction method facing multi-core DSP |
| CN112486681A (en) * | 2020-11-26 | 2021-03-12 | 迈普通信技术股份有限公司 | Communication method and network equipment |
| WO2022141297A1 (en) * | 2020-12-30 | 2022-07-07 | 华为技术有限公司 | Event processing method and apparatus |
| CN112859753A (en) * | 2021-01-19 | 2021-05-28 | 深圳市汇川技术股份有限公司 | Secondary development method, device and equipment for numerical control system and readable storage medium |
| CN112859753B (en) * | 2021-01-19 | 2022-03-25 | 深圳市汇川技术股份有限公司 | Secondary development method, device and equipment for numerical control system and readable storage medium |
| CN114741137A (en) * | 2022-05-09 | 2022-07-12 | 潍柴动力股份有限公司 | Software starting method, device, equipment and storage medium based on multi-core microcontroller |
| CN114741137B (en) * | 2022-05-09 | 2024-02-20 | 潍柴动力股份有限公司 | Software starting method, device, equipment and storage medium based on multi-core microcontroller |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105045658B (en) | 2018-10-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105045658A (en) | Method for realizing dynamic dispatching distribution of task by multi-core embedded DSP (Data Structure Processor) | |
| CN103761139B (en) | General purpose computation virtualization implementation method based on dynamic library interception | |
| CN102890643B (en) | Resource scheduling system based on immediate feedback of application effect under display card virtualization | |
| CN102360309B (en) | Scheduling system and scheduling execution method of multi-core heterogeneous system on chip | |
| CN103999051B (en) | Strategy for tinter resource allocation in the minds of shader core | |
| US20050149908A1 (en) | Graphical development of fully executable transactional workflow applications with adaptive high-performance capacity | |
| WO2016155335A1 (en) | Task scheduling method and device on heterogeneous multi-core reconfigurable computing platform | |
| CN102469126B (en) | Application scheduling system, method thereof and related device | |
| Sengupta et al. | Scheduling multi-tenant cloud workloads on accelerator-based systems | |
| CN102271145A (en) | Virtual computer cluster and enforcement method thereof | |
| CN104503832B (en) | A kind of scheduling virtual machine system and method for fair and efficiency balance | |
| CN103500119B (en) | A kind of method for allocating tasks based on pre-scheduling | |
| CN104050043A (en) | Share cache perception-based virtual machine scheduling method and device | |
| CN105550040A (en) | KVM platform based virtual machine CPU resource reservation algorithm | |
| CN113535362B (en) | Distributed scheduling system architecture and micro-service workflow scheduling method | |
| CN112905297A (en) | Container cluster resource scheduling method and device | |
| CN111459622A (en) | Method and device for scheduling virtual CPU, computer equipment and storage medium | |
| CN105100180A (en) | Cluster node dynamic loading method, device and system | |
| Nguyen et al. | Resource allocation for heterogeneous cloud computing | |
| CN115658278B (en) | Micro-task scheduler supporting high concurrency protocol interaction | |
| CN115964128A (en) | Heterogeneous GPU resource management and scheduling method and system | |
| CN113254143B (en) | Virtualized network function network element arrangement scheduling method, device and system | |
| CN114281529A (en) | Distributed virtualized client operating system scheduling optimization method, system and terminal | |
| De Munck et al. | Design and performance evaluation of a conservative parallel discrete event core for GES | |
| Lufei et al. | vStarCloud: An operating system architecture for Cloud computing |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210107 Address after: 311200 room b1-3-034, No. 198, Qidi Road, economic and Technological Development Zone, Xiaoshan District, Hangzhou City, Zhejiang Province Patentee after: Hangzhou purevision Technology Co.,Ltd. Address before: No.2, Taibai South Road, Yanta District, Xi'an City, Shaanxi Province Patentee before: XIDIAN University |
|
| TR01 | Transfer of patent right |