CN115827258B - Distributed task operation method, main control server, simulation computing node and system - Google Patents
Distributed task operation method, main control server, simulation computing node and system Download PDFInfo
- Publication number
- CN115827258B CN115827258B CN202310161094.XA CN202310161094A CN115827258B CN 115827258 B CN115827258 B CN 115827258B CN 202310161094 A CN202310161094 A CN 202310161094A CN 115827258 B CN115827258 B CN 115827258B
- Authority
- CN
- China
- Prior art keywords
- information
- app
- task
- subtask
- scheduling
- 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
- 238000004088 simulation Methods 0.000 title claims abstract description 145
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000004364 calculation method Methods 0.000 claims description 33
- 230000001360 synchronised effect Effects 0.000 claims description 32
- 238000004422 calculation algorithm Methods 0.000 claims description 13
- 238000012216 screening Methods 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 abstract description 3
- 230000008878 coupling Effects 0.000 description 7
- 238000010168 coupling process Methods 0.000 description 7
- 238000005859 coupling reaction Methods 0.000 description 7
- 238000012544 monitoring process Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Multi Processors (AREA)
- Computer And Data Communications (AREA)
Abstract
The invention relates to the technical field of distributed technology, and in particular discloses a distributed task operation method, a main control server, a simulation computing node and a system, wherein the distributed task operation method comprises the following steps: receiving and analyzing APP related information sent by a client, and obtaining APP information and task information; generating APP directed graph information and subtask execution information according to the APP information and the task information; carrying out distributed computation scheduling according to the sub-task execution information to obtain scheduling information; the scheduling information, the APP information and the subtask execution information are all sent to a simulation computing node; receiving a subtask execution result fed back by the simulation computing node, and repeating the process of obtaining the subtask execution result to complete all subtask execution information; and feeding back the subtask execution results of all the subtask execution information to the client. The distributed task operation method provided by the invention effectively solves the problem of sharing expert simulation experience.
Description
Technical Field
The present invention relates to the field of distributed technologies, and in particular, to a distributed task operation method, a master control server, a simulation computing node, and a distributed task operation system.
Background
In the traditional simulation, the expert has the limitation that the own simulation experience is difficult to share, so that the expert's simulation experience can be shared by an application program (APP) in a logical sequence and data flow sequence mode, and the expert's experience is solidified. However, this approach is only directed to single-level software, where institutions may have abundant computing resources, have high-performance computing clusters, and may require cooperation between two different systems for one simulation, where the expert's APP is unable to share the core computing program due to intellectual property issues.
Therefore, how to share expert simulation experience is a technical problem to be solved by those skilled in the art.
Disclosure of Invention
The invention provides a distributed task operation method, a main control server, a simulation computing node and a distributed task operation system, which solve the problem that expert simulation experience sharing cannot be realized in the related technology.
As a first aspect of the present invention, there is provided a distributed task running method, including:
receiving and analyzing APP related information sent by a client, and obtaining APP information and task information;
generating APP directed graph information and subtask execution information according to the APP information and the task information;
performing distributed computation scheduling according to the sub-task execution information to obtain scheduling information;
the scheduling information, the APP information and the subtask execution information are all sent to a simulation computing node, wherein the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP information and the task information and obtain a subtask execution result;
receiving a subtask execution result fed back by the simulation computing node, and repeating the process of obtaining the subtask execution result to complete all subtask execution information in the APP directed graph information;
and feeding back the subtask execution results of all the subtask execution information to the client.
Further, generating APP directed graph information and subtask execution information according to the APP application information and the task information, including:
generating APP directed graph information according to the APP information, wherein the APP directed graph information comprises APP attribute information and related information of an APP component;
and generating sub-task execution information according to execution logic according to the task information and the related information of the APP component.
Further, performing distributed computation scheduling according to the sub-task execution information to obtain scheduling information, including:
screening the subtask execution information to obtain an available simulation calculation node list capable of carrying out component subtask calculation;
the subtask execution information and the available simulation calculation node list are sent to a scheduling system, wherein the scheduling system can determine simulation calculation nodes corresponding to the subtask execution information according to a scheduling algorithm;
and obtaining the simulation computing node corresponding to the subtask execution information fed back by the scheduling system.
Further, the scheduling system can determine the simulation computing node corresponding to the subtask execution information according to a scheduling algorithm, including:
the scheduling system can schedule sub-task execution information according to a scheduling algorithm; and
the state score of each simulation computing node can be determined according to the current state of each simulation computing node in the available simulation computing node list, and the simulation computing node with the highest state score is used as the simulation computing node corresponding to the current subtask execution information;
the current state of the simulation computing node comprises CPU utilization rate, memory use condition, current running task number and priority level information.
Further, receiving the process of repeating the sub-task execution result to complete all sub-task execution information in the APP directed graph information, including:
searching the next APP component according to the APP directed graph information, generating sub-task execution information of the next APP component, and repeating the process of obtaining the sub-task execution result to obtain the sub-task execution result of the next APP component;
and sequentially completing the subtask execution information of all APP components in the APP directed graph information, and obtaining subtask execution results corresponding to all the subtask execution information.
As another aspect of the present invention, there is provided a distributed task running method, including:
receiving scheduling information, APP information and task information sent by a main control server, wherein the main control server can receive and analyze APP related information sent by a client, obtain APP information and task information, generate APP directed graph information and subtask execution information according to the APP information and the task information, and perform distributed computation scheduling according to the subtask execution information to obtain scheduling information;
calling a matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining a subtask execution result;
and sending the sub-task execution result to a main control server, wherein the main control server can feed back the sub-task execution results of all the sub-task execution information to the client.
Further, calling the matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining a subtask execution result, including:
calling a local simulation node instance, and opening a local deployment APP according to the APP information;
local task creation is carried out according to the subtask execution information, and task parameter information of the subtask execution information is synchronized;
and driving the APP component in the APP information to execute the simulation task according to the subtask execution information synchronized to the local, and obtaining a subtask execution result.
As another aspect of the present invention, there is provided a master server for implementing the distributed task running method described above, including:
the first receiving module is used for receiving and analyzing the APP related information sent by the client and obtaining APP information and task information;
the generation module is used for generating APP directed graph information and subtask execution information according to the APP information and the task information;
the scheduling module is used for performing distributed computation scheduling according to the sub-task execution information to obtain scheduling information;
the first sending module is used for sending the scheduling information, the APP information and the subtask execution information to a simulation computing node, wherein the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP information and the task information and obtain a subtask execution result;
the second receiving module is used for receiving the subtask execution result fed back by the simulation computing node and repeating the process of obtaining the subtask execution result so as to complete all the subtask execution information in the APP directed graph information;
and the feedback module is used for feeding back the subtask execution results of all the subtask execution information to the client.
As another aspect of the present invention, there is provided a simulated computing node for implementing the distributed task running method described above, including:
the third receiving module is used for receiving scheduling information, APP information and task information sent by the main control server, wherein the main control server can receive and analyze APP related information sent by the client, obtain APP information and task information, generate APP directed graph information and subtask execution information according to the APP information and the task information, and perform distributed computation scheduling according to the subtask execution information to obtain scheduling information;
the calling module is used for calling the matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining a subtask execution result;
the second sending module is used for sending the sub-task execution result to the main control server, wherein the main control server can feed back the sub-task execution results of all the sub-task execution information to the client.
As another aspect of the present invention, there is provided a distributed task execution system, including: the simulation system comprises a main control server and a plurality of simulation calculation nodes, wherein each simulation calculation node is in communication connection with the main control server, the main control server comprises the main control server, and each simulation calculation node comprises the simulation calculation node.
According to the distributed task operation method provided by the invention, the main control server can analyze the APP related information sent by the client, generate APP directed graph information and subtask execution information according to the analyzed APP information and task information, send the subtask execution information to the simulation computing node to realize simulation after computing and scheduling the subtask execution information, obtain the subtask execution result, and feed back all final subtask execution results to the client.
Drawings
The accompanying drawings are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate the invention and together with the description serve to explain, without limitation, the invention.
FIG. 1 is a flow chart of an embodiment of a method for distributed task execution provided by the present invention.
Fig. 2 is a flowchart of generating APP directed graph information and subtask execution information provided by the present invention.
Fig. 3 is a flowchart for obtaining scheduling information according to the present invention.
Fig. 4 is a flowchart of completing all sub-task execution information provided by the present invention.
Fig. 5 is a block diagram of a master server according to the present invention.
FIG. 6 is a flowchart of another embodiment of a distributed task execution method provided by the present invention.
Fig. 7 is a flowchart of an embodiment of obtaining a subtask execution result provided by the present invention.
FIG. 8 is a block diagram of a simulated computing node provided by the present invention.
Detailed Description
It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other. The invention will be described in detail below with reference to the drawings in connection with embodiments.
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate in order to describe the embodiments of the invention herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In this embodiment, a distributed task operation method is provided, and fig. 1 is a flowchart of the distributed task operation method provided according to an embodiment of the present invention, as shown in fig. 1, including:
s100, receiving and analyzing APP related information sent by a client, and obtaining APP information and task information;
in the embodiment of the invention, taking the quick prediction of the ship propulsion power for APP operation as an example, after receiving the APP related information sent by the client, analyzing the APP related information to obtain the APP information and the task information. The APP application information in the embodiment of the present invention may specifically include information such as an APP name, a security class, and the task information may specifically include information such as a task name, a security class, a priority level, and a task parameter. In addition, the user information can be analyzed, and the user information can comprise information such as user names, departments, security classes and the like.
S200, generating APP directed graph information and subtask execution information according to the APP information and the task information;
after the APP information and the task information are obtained through analysis, the directed graph information of the APP is rapidly forecast on the ship propulsion power according to the analyzed information, and subtask execution information is generated according to the first APP component to be executed.
Specifically, in the embodiment of the present invention, generating APP directed graph information and subtask execution information according to the APP application information and task information, as shown in fig. 2, includes:
s210, generating APP directed graph information according to the APP information, wherein the APP directed graph information comprises APP attribute information and related information of an APP component;
it should be noted that, in the embodiment of the present invention, the APP directed graph information is specifically generated according to the description of APP application information, where the APP directed graph information may specifically include specific attribute information of the APP, component information included in APP execution, parameter information of the APP, and parameter information of the component.
S220, generating sub-task execution information according to execution logic according to the task information and the related information of the APP component.
In the embodiment of the invention, the current APP adopts a parallel coupling assembly, and the assembly has two branches: the total duration of the Simulink and Adams branches, the total duration of the parallel coupling components is set to 2, the step length of the Adams branches is 0.0002, the step length of the Simulink branches is 0.0001, the minimum step length of the two branches is 0.0001, the total parallel times are 2/0.0001=2000, the two components are sequentially executed when the time is 0 in the first running time, the Simulink branches are executed once when the time is 0.0001 in the second running time, the Adams branches are not executed, the Simulink branches and Adams branches are respectively executed once when the time is 0.0002 in the third running time, the Simulink branches are executed once when the time is 0.0003 in the fourth running time, the Adams branches are not executed, and then the next analogies are sequentially executed, and when the current time is a common multiple of the step lengths of the two branches, the components in the two branches are subjected to data mapping once.
It should be noted that the parallel coupling design principle is: in parallel coupling circulation operation, a synchronous control method combining multithreading, a synchronous point array, a synchronous queue and a waiting queue is adopted to control the simulation operation. And respectively creating a thread for each loop branch in the simulation block to schedule a calculation model in the loop, and in order to reduce the number of synchronous control, carrying out synchronous processing only when the model which is completed in the current operation is a synchronous point model (namely the last model of the loop branch).
First, a synchronization point array is created, the array element of which is a multiple of the second smallest step size (set to the next smallest value) of the loop branches in the simulation block to the end time of the simulation block (for example, assuming that there are 6 loop branches in the simulation block, the step sizes are 3,1,3,6, 12,1, respectively; the end time of all loop branches in the same simulation block is the same, set to 120, the next smallest value is 3, and the element of the synchronization point array is (3, 6,9, 12, …, 120)). Each array element of the synchronization point array is a synchronization point moment, and at each synchronization point, there are some circulation branches to be synchronized. Secondly, a synchronous queue is established, and the elements of the synchronous queue are the current simulation time plus a cycle branch number with the cycle step length equal to the moment of the current synchronous point. The synchronous queue is used for controlling the synchronization of each circulating branch in the simulation block, and when the synchronization point advances forward, the synchronous threads corresponding to all circulating branches in the synchronous queue are awakened to continue to execute. Then, a waiting queue is established, the elements of which are the current simulation time plus the loop branch number at which the loop step length is greater than the current synchronization point. The waiting queue is used for controlling each circulating branch in the simulation block to wait for operation, and when the simulation time advances, and the simulation time plus the step length of a certain circulating branch in the waiting queue is equal to the synchronous point, the circulating branch is transferred into the synchronous queue. And finally, establishing synchronous threads, wherein the synchronous threads corresponding to each loop branch of the simulation block are used for processing the synchronization of the corresponding loop branch. When the loop branch is in the synchronous queue or the waiting queue, the synchronous thread is in a waiting state; when all the circulating branches are in the synchronous queue or the waiting queue, the synchronous threads corresponding to all the circulating branches in the synchronous queue are awakened.
Therefore, the distributed joint simulation environment adopts a synchronous control method combining multithreading (which refers to threads of professional model computing software, which run simultaneously), a synchronous point array, a synchronous queue and a waiting queue, so that each professional computing model runs orderly and cooperatively to realize joint simulation time sequence control among each computing model.
In addition, in the embodiment of the invention, the implementation can also be realized by adopting a serial coupling design principle, in the serial coupling circulation operation, the execution time of each calculation model is consistent (each calculation model has own time step) when each circulation is performed, so that only one waiting queue ordered according to the actual flow is needed to be established, the professional calculation model in the waiting queue is executed when each circulation is performed, and the synchronization processing is performed once after the execution of each waiting queue is completed once.
Whether serial or parallel coupling is an option that one skilled in the art can make according to his own needs is not limited herein.
S300, carrying out distributed computation scheduling according to the sub-task execution information to obtain scheduling information;
in the embodiment of the invention, the sub-task execution information obtained in the previous step is screened, and then calculation scheduling is performed according to the screened information, so that scheduling information is obtained.
Specifically, as shown in fig. 3, performing distributed computation scheduling according to the subtask execution information to obtain scheduling information, including:
s310, screening the subtask execution information to obtain a list of available simulation computing nodes capable of performing component subtask computation;
and screening a node list capable of carrying out sub-task calculation of the component according to the sub-task execution information, and transmitting the sub-task execution information and the available simulation calculation node list information into a dispatching system.
S320, the subtask execution information and the available simulation calculation node list are sent to a scheduling system, wherein the scheduling system can determine the simulation calculation node corresponding to the subtask execution information according to a scheduling algorithm;
s330, obtaining simulation computing nodes corresponding to the subtask execution information fed back by the scheduling system.
In the embodiment of the present invention, the scheduling system can determine, according to a scheduling algorithm, a simulation computing node corresponding to sub-task execution information, including:
the scheduling system can schedule sub-task execution information according to a scheduling algorithm; and
the state score of each simulation computing node can be determined according to the current state of each simulation computing node in the available simulation computing node list, and the simulation computing node with the highest state score is used as the simulation computing node corresponding to the current subtask execution information;
the current state of the simulation computing node comprises CPU utilization rate, memory use condition, current running task number and priority level information.
It should be understood that the scheduling system can schedule the tasks according to a specified scheduling algorithm, comprehensively score the tasks in the available node list according to the current state of the nodes (such as CPU utilization, memory usage, current running task number, priority and the like), and finally select a node with the highest score as a simulation computing node, and send task running information to the simulation computing node when the opportunity is mature.
S400, the scheduling information, the APP information and the subtask execution information are all sent to a simulation computing node, wherein the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP information and the task information and obtain a subtask execution result;
it should be noted that, in the embodiment of the present invention, the simulation computing node includes a simulation node tool and a computation monitoring node tool, where the computation monitoring node tool is a resident tool and is capable of receiving information of the main control server in real time, and the simulation node tool is used for simulation computing, and is tuned up at each task, and then exits at the end of the task.
In the embodiment of the invention, the scheduling information, the APP information and the subtask execution information are all sent to the simulation computing node, and after the calculation monitoring node tool in the simulation computing node acquires the main control scheduling information, the APP information and the subtask execution information, a local simulation node instance is called up and all the information is transmitted to the simulation node tool.
The simulation node tool opens the locally deployed APP according to the APP information, creates a local task according to the subtask execution information, synchronizes task parameter information in the subtask execution information, and drives the corresponding APP component to execute.
The calculation monitoring node tool reports the subtask executing process and the current subtask state of the simulation node tool to the main control server at any time, and after the subtask is executed, the subtask result information is returned to the main control server.
S500, receiving a subtask execution result fed back by the simulation computing node, and repeating the process of obtaining the subtask execution result to complete all subtask execution information in the APP directed graph information;
in the embodiment of the invention, after receiving the subtask operation result information, searching the next APP component according to the APP directed graph information, generating subtask execution information according to the next APP component, filtering out a node list capable of carrying out subtask calculation again, and transmitting the subtask information and the available node list information into a dispatching system to carry out new subtask calculation again.
Specifically, as shown in fig. 4, receiving the process of repeating the above-mentioned process of obtaining the execution result of the subtask to complete all the subtask execution information in the APP directed graph information includes:
s510, searching the next APP component according to the APP directed graph information, generating sub-task execution information of the next APP component, and repeating the process of obtaining the sub-task execution result to obtain the sub-task execution result of the next APP component;
s520, completing the subtask execution information of all APP components in the APP directed graph information in sequence, and obtaining subtask execution results corresponding to all the subtask execution information.
It should be understood that, in the embodiment of the present invention, after the execution information of all the subtasks of the APP components in the APP directed graph information is sequentially completed, the execution results of the subtasks corresponding to all the subtask execution information are obtained.
S600, feeding back the sub-task execution results of all the sub-task execution information to the client.
According to the distributed task operation method provided by the embodiment of the invention, the main control server can analyze the APP related information sent by the client, generate APP directed graph information and subtask execution information according to the analyzed APP information and task information, send the subtask execution information to the simulation computing node to realize simulation after computing and scheduling the subtask execution information, obtain subtask execution results, and feed back all final subtask execution results to the client.
As another embodiment of the present invention, a master server is provided for implementing the distributed task running method described above, where, as shown in fig. 5, the master server 100 includes:
the first receiving module 110 is configured to receive and parse APP related information sent by the client, and obtain APP application information and task information;
the generating module 120 is configured to generate APP directed graph information and subtask execution information according to the APP application information and the task information;
the scheduling module 130 is configured to perform distributed computing scheduling according to the subtask execution information, so as to obtain scheduling information;
the first sending module 140 is configured to send the scheduling information, the APP application information, and the subtask execution information to a simulation computing node, where the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP application information, and the task information and obtain a subtask execution result;
the second receiving module 150 is configured to receive the subtask execution result fed back by the simulation computing node, and repeat the process of obtaining the subtask execution result to complete all the subtask execution information in the APP directed graph information;
and the feedback module 160 is configured to feed back the execution results of all the subtasks in the subtask execution information to the client.
The main control server provided by the embodiment of the invention can analyze the APP related information sent by the client, generate APP directed graph information and subtask execution information according to the analyzed APP information and task information, send the APP directed graph information and the subtask execution information to the simulation computing node to realize simulation after computing and scheduling the subtask execution information, obtain subtask execution results, and feed back all final subtask execution results to the client.
The specific working principle and process of the master control server provided by the invention can refer to the description of the distributed task operation method, and are not repeated here.
As another embodiment of the present invention, there is provided a distributed task running method, including, as shown in fig. 6:
s610, receiving scheduling information, APP information and task information sent by a main control server, wherein the main control server can receive and analyze APP related information sent by a client, obtain APP information and task information, generate APP directed graph information and subtask execution information according to the APP information and the task information, and perform distributed computation scheduling according to the subtask execution information to obtain scheduling information;
s620, calling a matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining a subtask execution result;
in the embodiment of the present invention, according to the scheduling information, the APP application information and the subtask execution information, a matched simulation node instance is called and a subtask execution result is obtained, as shown in fig. 7, including:
s621, calling a local simulation node instance, and opening a local deployment APP according to the APP information;
s622, carrying out local task creation according to the subtask execution information and synchronizing task parameter information of the subtask execution information;
s623, driving the APP component in the APP information to execute the simulation task according to the subtask execution information synchronized to the local, and obtaining a subtask execution result.
S630, sending the sub-task execution result to a main control server, wherein the main control server can feed back the sub-task execution results of all the sub-task execution information to the client.
In summary, according to the distributed task operation method provided by the invention, after receiving the scheduling information, the APP information and the task information sent by the main control server, the matched simulation node instance is called, the sub-task execution result is obtained, and finally the sub-task execution result is fed back to the main control server. The distributed task operation method can separate the operation logic of the APP from the core algorithm, so that expert simulation experience sharing is realized, and the knowledge pain point of a user is solved.
As another embodiment of the present invention, a simulated computing node is provided for implementing the distributed task running method described above, where, as shown in fig. 8, the simulated computing node 800 includes:
the third receiving module 810 is configured to receive scheduling information, APP application information, and task information sent by a master control server, where the master control server is capable of receiving and analyzing APP related information sent by a client, obtaining APP application information and task information, generating APP directed graph information and subtask execution information according to the APP application information and task information, and performing distributed computation scheduling according to the subtask execution information to obtain scheduling information;
a calling module 820, configured to call the matched simulation node instance according to the scheduling information, the APP application information and the subtask execution information and obtain a subtask execution result;
and a second sending module 830, configured to send the subtask execution result to a main control server, where the main control server can feed back the subtask execution results of all the subtask execution information to the client.
The specific working principle and process of the master control server provided by the invention can refer to the description of the distributed task operation method, and are not repeated here.
As another embodiment of the present invention, there is provided a distributed task execution system including: the simulation system comprises a main control server and a plurality of simulation calculation nodes, wherein each simulation calculation node is in communication connection with the main control server, the main control server comprises the main control server, and each simulation calculation node comprises the simulation calculation node.
According to the distributed task operation system provided by the invention, the main control server can analyze the APP related information sent by the client, generate APP directed graph information and subtask execution information according to the analyzed APP information and task information, send the subtask execution information to the simulation computing node to realize simulation after computing and scheduling the subtask execution information, obtain subtask execution results, and feed back all final subtask execution results to the client. And the simulation computing node calls the matched simulation node instance after receiving the scheduling information, the APP information and the task information sent by the main control server, obtains a sub-task execution result, and finally feeds back the sub-task execution result to the main control server. The distributed task operation system can separate the operation logic of the APP from the core algorithm, so that expert simulation experience sharing is realized, and the knowledge pain point of a user is solved.
The specific working principle and process of the distributed task operation system provided by the embodiment of the present invention may refer to the description of the distributed task operation method, which is not repeated here.
It is to be understood that the above embodiments are merely illustrative of the application of the principles of the present invention, but not in limitation thereof. Various modifications and improvements may be made by those skilled in the art without departing from the spirit and substance of the invention, and are also considered to be within the scope of the invention.
Claims (8)
1. A method for running a distributed task, comprising:
receiving and analyzing APP related information sent by a client, and obtaining APP information and task information;
generating APP directed graph information and subtask execution information according to the APP information and the task information;
performing distributed computation scheduling according to the sub-task execution information to obtain scheduling information;
the scheduling information, the APP information and the subtask execution information are all sent to a simulation computing node, wherein the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP information and the task information and obtain a subtask execution result;
receiving a subtask execution result fed back by the simulation computing node, and repeating the process of obtaining the subtask execution result to complete all subtask execution information in the APP directed graph information;
feeding back sub-task execution results of all sub-task execution information to the client;
and carrying out distributed computation scheduling according to the sub-task execution information to obtain scheduling information, wherein the method comprises the following steps:
screening the subtask execution information to obtain an available simulation calculation node list capable of carrying out component subtask calculation;
the subtask execution information and the available simulation calculation node list are sent to a scheduling system, wherein the scheduling system can determine simulation calculation nodes corresponding to the subtask execution information according to a scheduling algorithm;
obtaining a simulation computing node corresponding to the subtask execution information fed back by the scheduling system;
the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP information and the task information and obtain a subtask execution result, and the simulation computing node comprises:
calling a local simulation node instance, and opening a local deployment APP according to the APP information;
local task creation is carried out according to the subtask execution information, and task parameter information of the subtask execution information is synchronized;
and driving the APP component in the APP information to execute the simulation task according to the subtask execution information synchronized to the local, and obtaining a subtask execution result.
2. The distributed task execution method according to claim 1, wherein generating APP digraph information and subtask execution information from the APP application information and task information includes:
generating APP directed graph information according to the APP information, wherein the APP directed graph information comprises APP attribute information and related information of an APP component;
and generating sub-task execution information according to execution logic according to the task information and the related information of the APP component.
3. The distributed task execution method according to claim 1, wherein the scheduling system is capable of determining a simulated computing node corresponding to the subtask execution information according to a scheduling algorithm, comprising:
the scheduling system can schedule sub-task execution information according to a scheduling algorithm; and
the state score of each simulation computing node can be determined according to the current state of each simulation computing node in the available simulation computing node list, and the simulation computing node with the highest state score is used as the simulation computing node corresponding to the current subtask execution information;
the current state of the simulation computing node comprises CPU utilization rate, memory use condition, current running task number and priority level information.
4. A distributed task execution method as claimed in claim 1, wherein receiving the process of repeating the above-mentioned process of obtaining the execution result of the subtasks to complete all the execution information of the subtasks in the APP directed graph information includes:
searching the next APP component according to the APP directed graph information, generating sub-task execution information of the next APP component, and repeating the process of obtaining the sub-task execution result to obtain the sub-task execution result of the next APP component;
and sequentially completing the subtask execution information of all APP components in the APP directed graph information, and obtaining subtask execution results corresponding to all the subtask execution information.
5. A method for running a distributed task, comprising:
receiving scheduling information, APP information and task information sent by a main control server, wherein the main control server can receive and analyze APP related information sent by a client, obtain APP information and task information, generate APP directed graph information and subtask execution information according to the APP information and the task information, and perform distributed computation scheduling according to the subtask execution information to obtain scheduling information;
calling a matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining a subtask execution result;
the sub-task execution result is sent to a main control server, wherein the main control server can feed back the sub-task execution result of all sub-task execution information to the client;
and carrying out distributed computation scheduling according to the sub-task execution information to obtain scheduling information, wherein the method comprises the following steps:
screening the subtask execution information to obtain an available simulation calculation node list capable of carrying out component subtask calculation;
the subtask execution information and the available simulation calculation node list are sent to a scheduling system, wherein the scheduling system can determine simulation calculation nodes corresponding to the subtask execution information according to a scheduling algorithm;
obtaining a simulation computing node corresponding to the subtask execution information fed back by the scheduling system;
the method for calling the matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining the subtask execution result comprises the following steps:
calling a local simulation node instance, and opening a local deployment APP according to the APP information;
local task creation is carried out according to the subtask execution information, and task parameter information of the subtask execution information is synchronized;
and driving the APP component in the APP information to execute the simulation task according to the subtask execution information synchronized to the local, and obtaining a subtask execution result.
6. A master server for implementing the distributed task running method as claimed in any one of claims 1 to 4, comprising:
the first receiving module is used for receiving and analyzing the APP related information sent by the client and obtaining APP information and task information;
the generation module is used for generating APP directed graph information and subtask execution information according to the APP information and the task information;
the scheduling module is used for performing distributed computation scheduling according to the sub-task execution information to obtain scheduling information;
the first sending module is used for sending the scheduling information, the APP information and the subtask execution information to a simulation computing node, wherein the simulation computing node can call a matched simulation node instance according to the scheduling information, the APP information and the task information and obtain a subtask execution result;
the second receiving module is used for receiving the subtask execution result fed back by the simulation computing node and repeating the process of obtaining the subtask execution result so as to complete all the subtask execution information in the APP directed graph information;
and the feedback module is used for feeding back the subtask execution results of all the subtask execution information to the client.
7. A simulated computing node for implementing the distributed task execution method of claim 5, comprising:
the third receiving module is used for receiving scheduling information, APP information and task information sent by the main control server, wherein the main control server can receive and analyze APP related information sent by the client, obtain APP information and task information, generate APP directed graph information and subtask execution information according to the APP information and the task information, and perform distributed computation scheduling according to the subtask execution information to obtain scheduling information;
the calling module is used for calling the matched simulation node instance according to the scheduling information, the APP information and the subtask execution information and obtaining a subtask execution result;
the second sending module is used for sending the sub-task execution result to the main control server, wherein the main control server can feed back the sub-task execution results of all the sub-task execution information to the client.
8. A distributed task execution system, comprising: the master server of claim 6 and a plurality of the emulated compute nodes of claim 7, each emulated compute node communicatively coupled to the master server.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310161094.XA CN115827258B (en) | 2023-02-24 | 2023-02-24 | Distributed task operation method, main control server, simulation computing node and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310161094.XA CN115827258B (en) | 2023-02-24 | 2023-02-24 | Distributed task operation method, main control server, simulation computing node and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115827258A CN115827258A (en) | 2023-03-21 |
CN115827258B true CN115827258B (en) | 2023-05-30 |
Family
ID=85522255
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310161094.XA Active CN115827258B (en) | 2023-02-24 | 2023-02-24 | Distributed task operation method, main control server, simulation computing node and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115827258B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117172633B (en) * | 2023-10-30 | 2024-01-30 | 浙江大学高端装备研究院 | Manufacturing service subgraph simulation method and system for industrial Internet platform |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101436959A (en) * | 2008-12-18 | 2009-05-20 | 中国人民解放军国防科学技术大学 | Method for distributing and scheduling parallel artificial tasks based on background management and control architecture |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114428722A (en) * | 2020-10-29 | 2022-05-03 | 上海阵量智能科技有限公司 | Hardware simulation method, device, equipment and storage medium |
CN112506627A (en) * | 2020-11-19 | 2021-03-16 | 西安电子科技大学 | Method, system, medium, device and terminal for scheduling directed acyclic graph tasks |
CN112506691B (en) * | 2020-12-14 | 2024-04-19 | 贵州电网有限责任公司 | Digital twin application fault recovery method and system for multi-energy system |
CN113848752B (en) * | 2021-09-24 | 2023-11-07 | 北京机电工程研究所 | Distributed real-time simulation method |
-
2023
- 2023-02-24 CN CN202310161094.XA patent/CN115827258B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101436959A (en) * | 2008-12-18 | 2009-05-20 | 中国人民解放军国防科学技术大学 | Method for distributing and scheduling parallel artificial tasks based on background management and control architecture |
Also Published As
Publication number | Publication date |
---|---|
CN115827258A (en) | 2023-03-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106775632B (en) | High-performance geographic information processing method and system with flexibly-expandable business process | |
Ruggiero et al. | Communication-aware allocation and scheduling framework for stream-oriented multi-processor systems-on-chip | |
Wang et al. | Workflow as a service in the cloud: architecture and scheduling algorithms | |
CN115827258B (en) | Distributed task operation method, main control server, simulation computing node and system | |
CN112114950A (en) | Task scheduling method and device and cluster management system | |
CN111181773B (en) | Delay prediction method for multi-component application of heterogeneous border cloud collaborative intelligent system | |
CN107220107B (en) | Multi-clock multi-task parallel real-time simulation system and method | |
US20190213040A1 (en) | Workflow scheduling system, workflow scheduling method, and electronic apparatus | |
CN112764902B (en) | Task scheduling method and system | |
CN114327829A (en) | Multi-core real-time task scheduling analysis and simulation system and method | |
Yildiz et al. | Heterogeneous hierarchical workflow composition | |
CN114625517A (en) | DAG graph computation distributed big data workflow task scheduling platform | |
CN115756822A (en) | Method and system for optimizing performance of high-performance computing application | |
Cuomo et al. | Performance prediction of cloud applications through benchmarking and simulation | |
Feljan et al. | Task allocation optimization for multicore embedded systems | |
Obaida et al. | Simulation of HPC job scheduling and large-scale parallel workloads | |
CN117234697A (en) | Conservative time synchronization parallel event scheduling computing architecture and method | |
CN109426564A (en) | A kind of IT resource data acquisition unifies manufacture timeout control method and system | |
Kim et al. | Cooperative scheduling schemes for explainable DNN acceleration in satellite image analysis and retraining | |
CN114896049A (en) | Method, system, equipment and medium for scheduling operation tasks of electric power artificial intelligence platform | |
Lebedev et al. | DDS: The dynamic deployment system | |
Fang et al. | Meta-process: a noval approach for decentralized execution of process | |
CN114298313A (en) | Artificial intelligence computer vision reasoning method | |
CN109583071B (en) | Parallel optimization method and system based on cloud simulation | |
CN114035919A (en) | Task scheduling system and method based on power distribution network layered distribution characteristics |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
CB02 | Change of applicant information |
Address after: Building T1, Xuelang Town, No. 99, Qingshu Road, Wuxi Economic Development Zone, Jiangsu Province, 214000 Applicant after: China Shipbuilding Orlando Wuxi Software Technology Co.,Ltd. Address before: Building T1, Xuelang Town, No. 99, Qingshu Road, Wuxi Economic Development Zone, Jiangsu Province, 214000 Applicant before: CSIC Orlando Wuxi Software Technology Co.,Ltd. |
|
CB02 | Change of applicant information | ||
GR01 | Patent grant | ||
GR01 | Patent grant |