Simulation task scheduling method, device, equipment and system
By obtaining the status information of simulation tasks and computing nodes, and finely allocating simulation tasks, the problems of low scheduling efficiency and unbalanced resource utilization in the existing technology are solved, and more efficient resource utilization and task execution are achieved.
Patent Information
- Application Number
- CN202410925769.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-10
- Publication Date
- 2024-11-08
AI Technical Summary
When facing large-scale and high-complex simulation tasks, existing simulation task scheduling methods have problems such as inefficient task execution and unbalanced resource utilization, which may cause the simulation computing node to be overloaded or idle.
By obtaining the configuration parameters of the simulation task to be assigned and the current status information of the candidate simulation calculation node, determining the task attribute information and the performance index of the candidate node, using a preset allocation strategy to allocate the most suitable computing nodes for the simulation task to achieve more refined .
It improves the resource utilization rate of the simulation computing cluster, shortens the waiting and execution time of simulation tasks, and improves the overall execution efficiency and user experience.
Smart Images

Figure CN118916128A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of simulation technology, and in particular to a simulation task scheduling method, device, equipment and system. Background Art
[0002] In the field of simulation technology, with the continuous advancement of technology and the complexity of application scenarios, the scale and complexity of simulation systems are also increasing significantly. At present, simulation systems are often composed of simulation computing clusters consisting of multiple simulation computing nodes. Each simulation computing node is responsible for executing the simulation tasks assigned to it. This distributed computing architecture can significantly improve the execution efficiency and processing power of simulation tasks, but it also brings challenges to simulation task scheduling.
[0003] Existing simulation task scheduling methods usually rely on simple polling or random allocation strategies. Although these methods can achieve the scheduling of simulation tasks to a certain extent, when faced with large-scale and highly complex simulation tasks, there are often problems such as low task execution efficiency and unbalanced resource utilization. There may even be situations where the simulation computing nodes are overloaded or idle. Summary of the invention
[0004] In view of this, the embodiments of the present application are committed to providing a simulation task scheduling method, device, equipment and system, which can enable simulation tasks to be assigned to simulation computing nodes that are suitable for their execution requirements, improve the resource utilization of the simulation computing cluster, and improve the overall execution efficiency of simulation tasks.
[0005] According to a first aspect of an embodiment of the present application, a simulation task scheduling method is provided, which is applied to a simulation system, wherein the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node being configured to execute a simulation task assigned to the simulation computing node; the method includes:
[0006] Obtaining configuration parameters of the simulation task to be assigned, and obtaining current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information;
[0007] Determining task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time;
[0008] According to the task attribute information and the current state information of each candidate simulation computing node, the corresponding candidate simulation computing node is assigned to the simulation task to be assigned, so that the candidate simulation computing node assigned to the simulation task to be assigned executes the simulation task to be assigned.
[0009] Optionally, the task attribute information further includes: task urgency and / or task priority, wherein the task urgency indicates the urgency with which the assigned simulation task needs to be executed, and the task priority indicates the degree to which the assigned simulation task needs to be executed with priority.
[0010] Optionally, determining the task attribute information of the simulation task to be assigned according to the configuration parameters includes:
[0011] Determining, according to the configuration parameters, the task characteristics of the simulation task to be assigned, wherein the task characteristics include at least one of the task type, task complexity, and computing resource requirements;
[0012] The predicted execution time of the simulation task to be assigned is determined according to the task characteristics.
[0013] Optionally, the current status information also includes the current task execution status;
[0014] Determining the predicted execution time of the simulation task to be assigned according to the task characteristics includes:
[0015] According to the task characteristics and the current state information of each candidate simulation computing node, predict the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned;
[0016] The shortest execution time is determined among the execution times corresponding to the candidate simulation computing nodes, and the shortest execution time is determined as the predicted execution time of the simulation task to be assigned.
[0017] Optionally, predicting the execution time corresponding to each candidate simulation computing node executing the to-be-allocated simulation task according to the task characteristics and the current state information of each candidate simulation computing node includes:
[0018] The task characteristics and the current state information of each candidate simulation computing node are input into a pre-trained first prediction model, and the first prediction model predicts the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned based on the task characteristics and the current state information of each candidate simulation computing node.
[0019] Optionally, the simulation task to be assigned includes a plurality of tasks, and assigning corresponding candidate simulation computing nodes to the simulation task to be assigned according to the task attribute information and the current state information of each candidate simulation computing node includes:
[0020] Determine the allocation priority of each simulation task to be allocated according to the task attribute information of each simulation task to be allocated, and determine the performance index of each candidate simulation computing node according to the current state information of each candidate simulation computing node;
[0021] The preset allocation strategy is executed, and the corresponding candidate simulation computing node is allocated to each simulation task to be allocated according to the allocation priority of each simulation task to be allocated and the performance index of each candidate simulation computing node.
[0022] Optionally, the method further includes:
[0023] Determining the task type of the simulation task to be assigned according to the configuration parameters;
[0024] According to the task type of the simulation task to be assigned, a data monitoring strategy for the simulation task to be assigned is determined, so that data monitoring of the simulation task to be assigned throughout its life cycle is performed according to the data monitoring strategy for the simulation task to be assigned.
[0025] According to a second aspect of an embodiment of the present application, a simulation task scheduling device is provided, which is applied to a simulation system, wherein the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node is used to execute a simulation task assigned to the simulation computing node; the device includes:
[0026] The first unit is used to obtain configuration parameters of the simulation task to be assigned, and obtain current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information;
[0027] A second unit is used to determine task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time;
[0028] The third unit is used to allocate corresponding candidate simulation computing nodes to the simulation task to be assigned according to the task attribute information and the current state information of each candidate simulation computing node, so that the candidate simulation computing nodes allocated to the simulation task to be assigned execute the simulation task to be assigned.
[0029] According to a third aspect of an embodiment of the present application, a simulation task scheduling device is provided, which is applied to a simulation system, wherein the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node is used to execute a simulation task assigned to the simulation computing node; the device includes a memory and a processor;
[0030] The memory is connected to the processor and is used to store programs;
[0031] The processor is used to implement the simulation task scheduling method as described in any one of the first aspects of the embodiments of the present application by running the program in the memory.
[0032] According to a fourth aspect of an embodiment of the present application, a simulation system is provided, including:
[0033] The simulation task scheduling device, simulation computing cluster, input module, and monitoring module as described in the third aspect of the embodiment of the present application, wherein the input module, the monitoring module, and each simulation computing node in the simulation computing cluster are respectively connected to the simulation task scheduling device for communication;
[0034] The input module is used to receive input information from a user, wherein the input information includes configuration parameters of the simulation task to be assigned;
[0035] The simulation computing cluster includes a plurality of simulation computing nodes, each of which is configured to execute a simulation task assigned to the simulation computing node;
[0036] The monitoring module is used to collect current status information of each simulation computing node, and the current status information includes current load information and / or current health status information.
[0037] The simulation task scheduling method provided in the present application first obtains the configuration parameters of each simulation task to be assigned, and obtains the current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information; then, based on the configuration parameters of each simulation task to be assigned, the task attribute information of each simulation task to be assigned is determined, wherein the task attribute information includes a predicted execution time; finally, based on the task attribute information of each simulation task to be assigned, and the current status information of each candidate simulation computing node, a corresponding candidate simulation computing node is assigned to each simulation task to be assigned, so that the candidate simulation computing node assigned to each simulation task to be assigned executes the simulation task to be assigned.
[0038] The simulation task scheduling method provided in the present application is applied to a simulation system including a simulation computing cluster. By comprehensively considering the task attribute information of the simulation task to be assigned, including the predicted execution time, and the current status information of all candidate simulation computing nodes, more refined resource allocation can be achieved. Not only can the simulation task be assigned to a node that meets its execution requirements, but it can also avoid excessive concentration or waste of resources, achieve balanced utilization of resources in the simulation computing cluster, avoid overload or idleness of simulation computing nodes, thereby improving the resource utilization of the entire simulation computing cluster. In addition, by optimizing resource allocation, the present application can shorten the waiting time and execution time of simulation tasks, thereby improving the overall execution efficiency of simulation tasks and enhancing user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0040] Figure 1 A schematic diagram of the structure of a simulation system provided in an embodiment of the present application.
[0041] Figure 2 A flowchart of a simulation task scheduling method provided in an embodiment of the present application.
[0042] Figure 3 A flowchart of another simulation task scheduling method provided in an embodiment of the present application.
[0043] Figure 4 A schematic diagram of the structure of a simulation task scheduling device provided in an embodiment of the present application.
[0044] Figure 5 A structural diagram of a simulation task scheduling device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0045] The technical solution of the embodiment of the present application is applicable to various scenarios where simulation tasks need to be scheduled, and is applicable to various fields such as manufacturing, scientific research, autonomous driving and intelligent transportation, aerospace, etc. In the application scenario, the user has submitted the simulation task to the simulation system, and the simulation system needs to assign the corresponding candidate simulation computing node to the simulation task submitted by the user to execute the simulation task. The technical solution of the embodiment of the present application enables the simulation task to be assigned to the simulation computing node that is suitable for its execution requirements, improves the resource utilization of the simulation computing cluster, and improves the overall execution efficiency of the simulation task.
[0046] The technical solution provided in the embodiments of the present application can be exemplarily applied to hardware devices such as processors, electronic devices, servers (including cloud servers), or packaged into software programs to be run. When the hardware device executes the processing of the technical solution of the embodiments of the present application, or the above-mentioned software program is run, the automatic splitting of the target task and the automatic calling of the application program interface required for the task can be achieved to complete the purpose of the target task. The embodiments of the present application only exemplarily introduce the specific processing of the technical solution of the present application, and do not limit the specific implementation form of the technical solution of the present application. Any technical implementation form that can execute the processing of the technical solution of the present application can be adopted by the embodiments of the present application.
[0047] In the field of simulation technology, with the continuous advancement of technology and the complexity of application scenarios, the scale and complexity of simulation systems are also increasing significantly. At present, simulation systems are often composed of simulation computing clusters consisting of multiple simulation computing nodes. Each simulation computing node is responsible for executing the simulation tasks assigned to it. This distributed computing architecture can significantly improve the execution efficiency and processing power of simulation tasks, but it also brings challenges to simulation task scheduling.
[0048] Existing simulation task scheduling methods usually rely on simple polling or random allocation strategies. Although these methods can achieve the scheduling of simulation tasks to a certain extent, when faced with large-scale and highly complex simulation tasks, there are often problems such as low task execution efficiency and unbalanced resource utilization. There may even be situations where the simulation computing nodes are overloaded or idle.
[0049] In view of this, the embodiments of the present application are committed to providing a simulation task scheduling method, device, equipment and system, which can enable simulation tasks to be assigned to simulation computing nodes that are suitable for their execution requirements, improve the resource utilization of the simulation computing cluster, and improve the overall execution efficiency of simulation tasks. The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0050] Exemplary Systems
[0051] For ease of understanding, firstly, an exemplary introduction to the implementation environment of the simulation task scheduling method provided in the embodiment of the present application is given. Figure 1 , Figure 1 This is a structural diagram of a simulation system provided in an embodiment of the present application. The simulation task scheduling method provided in the present application can be exemplarily applied to the simulation system.
[0052] like Figure 1 As shown, the simulation system includes an input module 110, a monitoring module 120, a simulation computing cluster 130 and a simulation task scheduling device 140, wherein the simulation computing cluster 130 includes a plurality of simulation computing nodes, each of which is configured to execute a simulation task assigned to the simulation computing node. The input module 110, the monitoring module 120 and each of the simulation computing nodes in the simulation computing cluster 130 are respectively connected to the simulation task scheduling device 140 for communication.
[0053] The input module 110 is used to receive user input information, and the input information includes configuration parameters of the simulation task to be assigned. The input module 110 is also used to send the configuration parameters of the simulation task to be assigned to the simulation task scheduling device 140 through the communication network. Specifically, the simulation system provides a user interaction interface, and the user can log in to the relevant system through the user interaction interface, and propose the simulation task to be completed in the input module 110, and input the configuration parameters of the simulation task. The input module 110 generates a data packet including the configuration parameters of the simulation task, and sends the data packet to the simulation task scheduling device 140 through the communication network. The simulation task scheduling device 140 can obtain the configuration parameters of the simulation task and determine the simulation task as the simulation task to be assigned.
[0054] The monitoring module 120 is used to collect and manage the current status information in the simulation computing cluster 130 in real time, and send the collected current status information to the simulation task scheduling device 140 through the communication network. The simulation task scheduling device 140 can obtain the current status information of each simulation computing node in the simulation computing cluster 130.
[0055] The current status information includes the current load information and / or current health status information of each simulation computing node. The current load information includes but is not limited to any one or more combinations of data such as CPU usage, memory usage, network bandwidth occupancy, etc. By analyzing the current load information, the current load status of each simulation computing node can be determined. The current health status information includes but is not limited to any one or more combinations of data such as network delay and number of error logs, and the current health status of each simulation computing node can be determined by analyzing the current health status information.
[0056] Optionally, the monitoring module 120 is also used to collect simulation task scheduling logs and / or simulation task execution logs in real time, perform fine-grained monitoring of the simulation task throughout its life cycle, and collect key task allocation and / or operation data, such as task start time, execution time, completion time, task status change records, and track task failure or error logs. Optionally, after collecting relevant data, the monitoring module 120 first performs preliminary cleaning and formatting of the data to ensure data quality and consistency.
[0057] The simulation task scheduling device 140 is used to execute the simulation task scheduling method provided by each embodiment of the present application after receiving the configuration parameters of the simulation task, including: obtaining the configuration parameters of the simulation task to be assigned, and obtaining the current state information of each candidate simulation computing node, the current state information including the current load information and / or the current health state information; according to the configuration parameters, determining the task attribute information of the simulation task to be assigned, the task attribute information including the predicted execution time; according to the task attribute information, and the current state information of each candidate simulation computing node, assigning the corresponding candidate simulation computing node to the simulation task to be assigned, after determining the candidate simulation computing node assigned to the simulation task to be assigned, sending the configuration parameters of the simulation task to be assigned to the candidate simulation computing node through the communication network, so that the candidate simulation computing node executes the simulation task to be assigned. As for the specific content of the simulation task scheduling method, it will be described in the subsequent embodiments, and the content of the subsequent embodiments can be referred to, and no detailed description will be given here.
[0058] Exemplary Methods
[0059] Figure 2 A flowchart of a simulation task scheduling method provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the simulation task scheduling method can be exemplarily applied to Figure 1 The simulation system in the simulation system is specifically executed by the simulation task scheduling device in the simulation system. The method includes steps S201-S203:
[0060] S201, obtaining configuration parameters of a simulation task to be assigned, and obtaining current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information.
[0061] The simulation task to be assigned can be understood as a simulation task that has been submitted to the simulation system by a user but has not yet been assigned to a specific simulation computing node for execution.
[0062] The configuration parameters of the simulation task to be assigned can be understood as a series of parameters or set values related to the simulation task provided by the user when submitting the simulation task, which can include any one or more combinations of simulation task related parameters such as task description, computing resource requirements, simulation model parameters, simulation step size, simulation algorithm parameters, input and output specifications, time limits, task urgency, boundary conditions and initial conditions. In the actual application process, the content of the configuration parameters is specifically determined according to the simulation task requirements and the selection of simulation software or system, and this application does not limit this.
[0063] The configuration parameters of the simulation task to be assigned can characterize the characteristics, requirements, constraints and other related information of the simulation task to be assigned, and are the basis for scheduling and executing the simulation task to be assigned. By analyzing the configuration parameters, the simulation system can understand the relevant attributes of the simulation task to be assigned and reasonably allocate resources accordingly.
[0064] Optional, such as Figure 1 As shown, after receiving the configuration parameters of the simulation task input by the user, the input module sends the configuration parameters of the simulation task to the simulation task scheduling device through the communication network. The simulation task scheduling device can obtain the configuration parameters of the simulation task and determine the simulation task as the simulation task to be assigned.
[0065] The candidate simulation computing node can be understood as a set of computing nodes in a simulation computing cluster of a simulation system that can be used to execute the simulation task to be assigned, and the set includes at least one simulation computing node.
[0066] Optionally, the candidate simulation computing nodes may be all simulation computing nodes in the simulation computing cluster. Optionally, the candidate simulation computing nodes may be a subset of the simulation computing cluster, which is selected from the simulation computing cluster according to specific screening conditions (such as computing power, memory size, network latency, health status, specific software or hardware requirements, etc.). Optionally, the candidate simulation computing nodes may be a node set dynamically determined based on a certain preset strategy.
[0067] The current status information of the candidate simulation computing node can be understood as relevant information used to describe the operating status of the candidate simulation computing node at the current moment, including but not limited to current load information (such as CPU usage, memory usage, disk I / O and network bandwidth, etc., which can reflect the current load status of the candidate simulation computing node) and / or current health status information (such as network latency, number of error logs, hardware fault detection results, operating system stability assessment, and software service status monitoring data, etc., which can be used to evaluate the current health status of the candidate simulation computing node).
[0068] Optionally, the current status information of the candidate simulation computing node may also include current task execution status information, security and compliance information, other specific information and other related information, which is not limited in this application.
[0069] The current state information of the candidate simulation computing node can represent the current properties of the candidate simulation computing node. By analyzing the current state information, the simulation system can be helped to comprehensively evaluate the candidate simulation computing nodes so as to make the best task allocation decision.
[0070] Optional, such as Figure 1 As shown, after the monitoring module collects the current status information of each candidate simulation computing node in real time, it sends the current status information of the candidate simulation computing node to the simulation task scheduling device through the communication network, and the simulation task scheduling device can obtain the current status information of each candidate simulation computing node.
[0071] S202. Determine task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time.
[0072] The task attribute information of the simulation task to be assigned can be understood as relevant information representing key attributes of the simulation task to be assigned. These attribute information are directly related to the actual execution requirements of the simulation task to be assigned, and have vital guiding significance for subsequent task scheduling.
[0073] According to the foregoing content, the configuration parameters of the simulation task to be assigned include data that characterizes the characteristics, requirements, constraints and other related information of the simulation task to be assigned. By conducting in-depth analysis of the configuration parameters of the simulation task to be assigned, the task attribute information of the simulation task to be assigned can be determined.
[0074] The task attribute information includes a predicted execution time, which refers to the estimated time required to execute the simulation task to be assigned. The predicted execution time can directly affect the execution efficiency and completion time of the simulation task to be assigned on the candidate simulation computing node. A longer predicted execution time may mean that the task requires more computing resources or a longer execution cycle.
[0075] As an optional implementation, the predicted execution time can be understood as the expected execution time required to complete the assigned simulation task in a specific execution environment (such as a pre-set, standardized simulation computing node). This implementation provides a standardized execution time evaluation.
[0076] In this implementation, the predicted execution time can be determined based on a variety of implementations, such as determining the predicted execution time through the processing concept of an algorithm or model. As an optional implementation, determining the predicted execution time of the simulation task to be assigned according to the configuration parameters includes: determining the task characteristics of the simulation task to be assigned according to the configuration parameters, and determining the predicted execution time of the simulation task to be assigned according to the task characteristics, wherein the task characteristics include at least one of the task type, task complexity, and computing resource requirements. This will be described in detail in subsequent embodiments and will not be repeated here.
[0077] As an optional implementation, the predicted execution time can be understood as the shortest execution time of the estimated execution time of the simulation task to be assigned on each candidate simulation computing node. This implementation provides a more specific and practical execution time evaluation, which helps to optimize task scheduling and resource allocation decisions.
[0078] In this implementation, optionally, determining the predicted execution time of the simulation task to be assigned includes: determining the task characteristics of the simulation task to be assigned according to the configuration parameters; and predicting the shortest execution time of the simulation task to be assigned on each candidate simulation computing node among the expected execution times according to the task characteristics and the current state information of each candidate simulation computing node. This will be described in detail in subsequent embodiments and will not be elaborated here.
[0079] As an optional implementation, the task attribute information may also include task urgency and / or task priority, wherein the task urgency indicates the urgency with which the assigned simulation task needs to be executed, and the task priority indicates the degree to which the assigned simulation task needs to be executed with priority.
[0080] The task urgency can be understood as the urgency with which the simulation task to be assigned needs to be executed immediately. The simulation tasks to be assigned with high urgency usually need to be started as soon as possible to reduce delays or avoid potential adverse consequences. Understanding the urgency of the simulation tasks to be assigned helps the simulation task scheduling device to consider not only the task attribute information such as the predicted execution time, but also the urgency of the simulation tasks when assigning time-sensitive simulation tasks, so as to meet their time urgency requirements.
[0081] Optionally, the task urgency of the simulation task to be assigned may be determined based on the time constraints in the configuration parameters of the simulation task to be assigned. For example, the configuration parameters of the simulation task to be assigned include the deadline for completion of the simulation task, and the task urgency may be determined based on the deadline for completion and the current time, such as a simulation task to be assigned that is close to or exceeds the deadline may have a higher task urgency.
[0082] Optionally, when inputting the configuration parameters of the simulation task to be assigned, the user directly inputs the corresponding task urgency, that is, the configuration parameters of the simulation task to be assigned include the task urgency.
[0083] The task priority can be understood as the importance of the simulation task to be assigned relative to other simulation tasks to be assigned. The task priority helps the simulation task scheduling device to assign simulation tasks when system resources are limited. It not only considers task attribute information such as predicted execution time, but also comprehensively considers the importance of the simulation task relative to other simulation tasks to be assigned, so as to meet the priority requirements of the simulation task.
[0084] Task priority can be determined based on a variety of factors, such as the business value of the simulation task, the impact on the overall system performance, dependencies, etc. For example, certain critical simulation tasks or simulation tasks that have a significant impact on the overall project schedule may be considered more important and therefore be given a higher task priority.
[0085] Optionally, determining the task priority of the simulation task to be assigned according to the configuration parameters of the simulation task to be assigned includes: determining the task type of the simulation task to be assigned according to the configuration parameters of the simulation task to be assigned, and determining the task priority of the simulation task to be assigned according to the task type of the simulation task to be assigned and the preset task priority corresponding to the task type. For example, by identifying key information such as a specific fluid dynamics model and fluid mechanics related parameters in the configuration parameters of the simulation task to be assigned, it is determined that its task type is a fluid simulation task, and in the pre-set mapping relationship between different simulation types and task priorities, the task priority of the fluid simulation task is medium, then the task priority of the simulation task to be assigned is medium.
[0086] In order to facilitate the processing of the scheduling algorithm, the urgency and priority of the task can be quantified into specific values or levels. For example, a level of 1-10 can be used to represent the urgency or priority of the task, where level 1 represents the lowest and level 10 represents the highest. In addition, when determining the urgency and priority of the task, it is necessary to comprehensively consider the relevant business rules to ensure that the urgency and priority of the task are accurately reflected.
[0087] Task urgency and / or task priority provide an additional dimension of decision-making basis for simulation task scheduling, which helps the simulation task scheduling equipment to more reasonably allocate candidate simulation computing nodes to the simulation tasks to be assigned, making the decision-making process more scientific and accurate, ensuring that high-priority and urgent tasks receive sufficient resource support, while avoiding non-critical tasks from occupying too many resources.
[0088] S203. According to the task attribute information and the current state information of each candidate simulation computing node, the corresponding candidate simulation computing node is assigned to the simulation task to be assigned, so that the candidate simulation computing node assigned to the simulation task to be assigned executes the simulation task to be assigned.
[0089] According to the foregoing content, the current status information of the candidate simulation computing node can be understood as relevant information used to describe the operating status of the candidate simulation computing node at the current moment. Therefore, by analyzing the current status information of each candidate simulation computing node, the operating status of each candidate simulation computing node at the current moment can be determined separately (including but not limited to the current load situation and / or current health status).
[0090] Specifically, by analyzing the current load information in the current state information, the current load condition of each candidate simulation computing node can be determined, and the load condition can reflect the workload, resource usage status and potential processing capacity of each candidate simulation computing node at the current moment.
[0091] For example, the current load situation of the candidate simulation computing node can be determined by analyzing the occupancy rate of various resources (such as CPU, memory, disk I / O, etc.) on the candidate simulation computing node, and / or analyzing the number and type of tasks currently being executed on the candidate simulation computing node, and / or analyzing the length of the queue of tasks waiting to be executed on the candidate simulation computing node.
[0092] Optionally, in order to facilitate the processing of the scheduling algorithm, the current load condition of the candidate simulation computing node may be quantified into a specific value or level, for example, the load condition may be expressed in the form of a load index.
[0093] Specifically, by analyzing the current health status information in the current status information, the current health status of each candidate simulation computing node can be determined. The health status can reflect the operating health status of the hardware and / or software of each candidate simulation computing node at the current moment, and can indicate whether the candidate simulation computing node currently has health problems, and / or the severity of the existing health problems, and / or the impact of the existing health problems on subsequent task allocation.
[0094] For example, by analyzing the health status information of hardware components (such as processors, memory modules, network interfaces, etc.), checking whether the hardware components of the candidate simulation computing nodes have hardware failures or performance degradation problems, and / or, analyzing the health status information related to software operation, evaluating the stability and compatibility of the operating system, simulation software and other related software running on the candidate simulation computing nodes, and / or, analyzing historical fault information (fault type, occurrence frequency and recovery time), etc., evaluating the instability factors of the candidate simulation computing nodes, and then determining the current health status of the candidate simulation computing nodes.
[0095] Optionally, in order to facilitate the processing of the scheduling algorithm, the current health status of the candidate simulation computing node can be quantified into a specific value or level. For example, the health status can be divided into five levels: "excellent, good, average, poor, and bad", and each level can correspond to a specific health value range.
[0096] Specifically, after determining the task attribute information of the simulation task to be assigned and the current status of each candidate simulation computing node, the preset allocation strategy is executed to determine the target simulation computing node among the candidate simulation computing nodes. The target simulation computing node is the candidate simulation computing node assigned to the simulation task to be assigned and is used to execute the simulation task to be assigned.
[0097] The allocation strategy is a pre-set one that determines the processing of the target simulation computing node through a series of preset rules or algorithms based on the task attribute information of the simulation task to be assigned and the current status of each candidate simulation computing node. The allocation strategy is pre-stored on the cloud or device side. After the simulation task scheduling device determines the task attribute information of the simulation task to be assigned and the current status of each candidate simulation computing node, it calls and executes the pre-set allocation strategy to determine the target simulation computing node among the candidate simulation computing nodes.
[0098] As an optional implementation, there is only one simulation task to be assigned at the current moment. In this implementation, exemplarily, the allocation strategy may include: comprehensively analyzing the task attribute information (especially the predicted execution time) of the simulation task to be assigned, determining the actual demand of the simulation task to be assigned, and comprehensively evaluating the current status of each candidate simulation computing node, determining the candidate simulation computing node that best matches the actual demand of the simulation task to be assigned based on the above analysis content, and assigning the simulation task to be assigned to the candidate simulation computing node.
[0099] Optionally, when system resources are sufficient, candidate simulation computing nodes with relatively light load, sufficient resources to execute simulation tasks, and good health status are preferentially selected as target computing nodes, so as to ensure that the simulation tasks to be assigned can be completed within the scheduled time, while reducing the risk of interruption due to node failure. Optionally, when system resources are relatively tight, candidate simulation computing nodes that best match the actual needs of the simulation tasks to be assigned are selected as target computing nodes, so as to reduce system resource usage as much as possible while meeting the actual needs of the simulation tasks to be assigned.
[0100] As an optional implementation manner, there are multiple simulation tasks to be assigned at the current moment.
[0101] In this implementation, the simulation task scheduling device determines the task attribute information of each simulation task to be assigned according to steps S201 and S202, and when executing step S203 "allocating the corresponding candidate simulation computing node to the simulation task to be assigned according to the task attribute information and the current state information of each candidate simulation computing node", it includes steps A1-A2:
[0102] A1. Determine the allocation priority of each simulation task to be allocated according to the task attribute information of each simulation task to be allocated, and determine the performance index of each candidate simulation computing node according to the current state information of each candidate simulation computing node.
[0103] The allocation priority of a simulation task to be allocated can be understood as the degree to which the simulation task to be allocated is allocated first among multiple simulation tasks to be allocated. A higher allocation priority means that the corresponding simulation task to be allocated has a higher priority allocation degree and is more likely to be allocated to a suitable simulation computing node for execution.
[0104] Specifically, after obtaining the task attribute information of each simulation task to be assigned, a first comprehensive analysis is performed on the task attribute information of each simulation task to be assigned, so as to determine the assignment priority of each simulation task to be assigned.
[0105] Optionally, the first comprehensive analysis includes: assigning corresponding weights to each item of information in the task attribute information of the simulation task to be assigned, and comprehensively calculating the assignment priority score of the simulation task based on each item of information and the corresponding weight, and then determining the assignment priority of each simulation task to be assigned based on the assignment priority scores of each simulation task to be assigned. Exemplarily, the task attribute information includes predicted execution time, task urgency, and task priority, and different weights are assigned to these three items of attribute information, and then the assignment priority score of the simulation task to be assigned is obtained by weighted summation.
[0106] The performance index can be understood as the execution capability of the candidate simulation computing node in the current state. A higher performance index means that the corresponding candidate simulation computing node has better comprehensive execution capability, such as more available resources or better health status.
[0107] Specifically, after obtaining the current state information of each candidate simulation computing node, for each candidate simulation computing node, determine its current state according to its current state information, and perform a second comprehensive analysis on various states in its current state to determine the performance index of the candidate simulation computing node.
[0108] Optionally, the second comprehensive analysis includes: assigning corresponding weights to various states in the current state of the candidate simulation computing node, and comprehensively calculating the comprehensive performance index of the candidate simulation computing node based on the various states and the corresponding weights, and the relative execution capability of the candidate simulation computing node can be determined based on the calculated performance index. Exemplarily, the current state includes the current load condition and the current health state, and the current load condition and the current health state have been quantified into specific values or levels. Different weights are assigned to the two current states, and then the performance index of the candidate simulation computing node is obtained by weighted summation.
[0109] A2. Execute a preset allocation strategy, and allocate a corresponding candidate simulation computing node to each simulation task to be allocated according to the allocation priority of each simulation task to be allocated and the performance index of each candidate simulation computing node.
[0110] As an optional implementation method, the allocation strategy can be an optimal matching strategy, including: sorting each simulation task to be allocated in order of allocation priority from high to low, and sorting each candidate simulation computing node in order of performance index from high to low; then selecting the simulation task to be allocated with the highest allocation priority from the sorted list of simulation tasks to be allocated, selecting the candidate simulation computing node with the highest performance index from the sorted list of candidate simulation computing nodes, and assigning the candidate simulation computing node to the candidate simulation computing node; repeating the previous step until all simulation tasks to be allocated are assigned to suitable candidate simulation computing nodes, and in each iteration, selecting the next simulation task to be allocated with the highest matching priority from the list of simulation tasks to be allocated, and selecting the next candidate simulation computing node with the highest performance index from the list of candidate nodes for matching.
[0111] Optionally, after selecting the to-be-assigned simulation task with the highest allocation priority from the sorted list of to-be-assigned simulation tasks and selecting the candidate simulation computing node with the highest performance index from the sorted list of candidate simulation computing nodes, you can first check whether the candidate simulation computing node meets the actual requirements of the to-be-assigned simulation task. If so, assign the candidate simulation computing node to the candidate simulation computing node. If not, you can wait for a while until the current state of a computing node changes (such as completing a simulation task and releasing resources, etc.) and the performance index is higher than the performance index of the candidate simulation computing node before matching.
[0112] This optional implementation method can ensure that the simulation tasks to be assigned with higher assignment priorities can preferentially obtain more suitable computing resources, thereby improving the overall task execution efficiency and resource utilization.
[0113] Optionally, the allocation strategy can also be a load balancing strategy, a demand matching strategy, a combination strategy, or other feasible allocation strategies, which are not limited in this application. For example, the allocation strategy is a demand matching strategy, which includes: sorting each simulation task to be allocated in order of allocation priority from high to low, and then matching each simulation task to be allocated with a candidate simulation computing node in turn according to the sorting order, and the matching includes: for each simulation task to be allocated, searching for a first candidate simulation computing node whose performance index meets the actual demand corresponding to the simulation task attribute of the simulation task to be allocated from the remaining candidate simulation computing nodes, and selecting a candidate simulation computing node with the highest performance index or the most suitable one from the first candidate simulation computing node for allocation.
[0114] The simulation task scheduling method provided in the present application first obtains the configuration parameters of each simulation task to be assigned, and obtains the current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information; then, based on the configuration parameters of each simulation task to be assigned, the task attribute information of each simulation task to be assigned is determined, wherein the task attribute information includes a predicted execution time; finally, based on the task attribute information of each simulation task to be assigned, and the current status information of each candidate simulation computing node, a corresponding candidate simulation computing node is assigned to each simulation task to be assigned, so that the candidate simulation computing node assigned to each simulation task to be assigned executes the simulation task to be assigned.
[0115] The simulation task scheduling method provided in the present application is applied to a simulation system including a simulation computing cluster. By comprehensively considering the task attribute information of the simulation task to be assigned, including the predicted execution time, and the current status information of all candidate simulation computing nodes, more refined resource allocation can be achieved. Not only can the simulation task be assigned to a node that meets its execution requirements, but it can also avoid excessive concentration or waste of resources, achieve balanced utilization of resources in the simulation computing cluster, avoid overload or idleness of simulation computing nodes, thereby improving the resource utilization of the entire simulation computing cluster. In addition, by optimizing resource allocation, the present application can shorten the waiting time and execution time of simulation tasks, thereby improving the overall execution efficiency of simulation tasks and enhancing user experience.
[0116] As an optional implementation, in step S202, "determining the predicted execution time of the simulation task to be assigned according to the configuration parameters" includes steps B1-B2:
[0117] B1. Determine the task characteristics of the simulation task to be assigned according to the configuration parameters.
[0118] The task characteristics may be understood as key information elements that can describe the inherent properties of the simulation task itself, including but not limited to at least one of the task type, task complexity, and computing resource requirements.
[0119] The task type may be understood as a specific category or field to which the simulation task belongs, such as structural simulation, fluid simulation, thermal simulation, etc. Different types of simulation tasks may require different calculation methods, resources, and estimated execution time.
[0120] Optionally, the configuration parameters of the simulation task to be assigned can be matched with a predefined simulation task feature library to determine the task type of the simulation task to be assigned. Taking a fluid simulation task as an example, by comparing the similarity between the configuration parameters and the fluid dynamics model, key parameters and other information in the feature library, the simulation task can be identified as a fluid simulation task.
[0121] In order to manage and execute simulation tasks more efficiently, attribute templates can be defined for each type of simulation task. The attribute templates cover key information such as the type of computing resources required for task execution, the input and output formats of data, and the applicable computing model. When entering configuration parameters, users can directly enter them according to the module, which facilitates subsequent simulation type matching.
[0122] The task complexity reflects the complexity of the simulation task, including the number of calculation steps, the complexity of the required algorithm, etc. A task with high complexity usually requires more computing resources and time. It directly affects the time it takes to execute the task and the amount of resources required.
[0123] Quantifying the complexity of a simulation task may involve factors such as the size of the model, the time span of the simulation, the accuracy of the required calculations, etc. Tasks with high complexity may require more computing resources and time.
[0124] Optionally, the task complexity of the simulation task to be assigned is determined according to the configuration parameters of the simulation task to be assigned, including: first, extracting features related to the task complexity from the configuration parameters of the simulation task to be assigned, the features may include but are not limited to the complexity of the model (such as the number and type of equations, the number of variables, etc.), the number of required calculation steps, the scale and complexity of the data (such as the size, dimension, resolution, etc. of the input data); then, based on the extracted features related to the task complexity, a predefined complexity evaluation standard or algorithm is used to quantify the complexity of the task, for example, weighting and summing the features or using a more complex mathematical model to calculate the task complexity score.
[0125] Optionally, after extracting features related to task complexity, the extracted features can also be matched with a predefined task complexity template. The task complexity template represents typical task features of different complexity levels. By comparing the similarity between the features related to task complexity of the simulation task to be assigned and the task complexity template, the complexity level of the simulation task to be assigned can be determined.
[0126] The computing resource requirements can be understood as the computing resources required during the execution of the simulation task, such as computing density (CPU), memory, network bandwidth, etc., which intuitively represents the requirements of the simulation task to be assigned for the hardware environment.
[0127] Optionally, by analyzing the computational steps and complexity required for the simulation task to be assigned, estimate its CPU or GPU usage, and dynamically adjust the resource allocation strategy by using a reinforcement learning algorithm combined with historical data of the simulation task. Optionally, by analyzing the scale and data volume of the simulation model, determine the memory requirements. Optionally, by analyzing the size and transmission frequency of the input and output data, estimate the bandwidth requirements.
[0128] In this implementation, not only are simulation tasks classified, but resource demand analysis is also performed in multiple dimensions, such as computational density analysis, memory usage analysis, and bandwidth analysis. This deeply integrated feature analysis can more accurately predict and allocate necessary resources.
[0129] B2. Determine the predicted execution time of the simulation task to be assigned according to the task characteristics.
[0130] The task type, task complexity, and computing resource requirements all have a certain impact on the simulation task execution time. Optionally, the expected basic time corresponding to each task type is pre-set so that the task type can characterize the basic time required to execute the task of that type. For example, a fluid simulation task may require longer computing time than a structural simulation task. Optionally, simulation tasks with high task complexity usually require more computing steps and longer execution time. Optionally, computing resource requirements will also affect the execution time. If the task requires a large amount of computing resources, and the resources of the candidate computing nodes are limited, the execution time may be extended.
[0131] Specifically, after determining the task characteristics of the simulation task to be assigned according to its configuration parameters, the simulation task scheduling device performs a first processing, conducts further comprehensive analysis on the task characteristics of the simulation task to be assigned, and determines the predicted execution time of the simulation task to be assigned through the comprehensive impact of various task characteristics on the execution time of the simulation task.
[0132] The first processing includes multiple implementation methods, such as using empirical data, historical execution records or prediction models to process task characteristics and determine the predicted execution time of the assigned simulation task, which is not limited in this application.
[0133] As an optional implementation method, the first processing can be implemented based on the processing idea of the model. Specifically, the task characteristics of the simulation task to be assigned are input into the pre-trained second prediction model, and the second prediction model predicts the execution time of the simulation task to be assigned based on the input task characteristics, and outputs the predicted execution time of the simulation task to be assigned.
[0134] In the embodiment of the present application, a second prediction model is pre-trained. The second prediction model can be obtained by performing execution time prediction training on the second training sample. The second training sample includes sample task features and the annotated execution time corresponding to the sample task features, and the annotated execution time includes the correct execution time of the simulation task corresponding to the sample task features. The second training sample can use an existing public data set, or it can create and annotate a data set by itself. The execution time prediction training is aimed at enabling the second prediction model to determine the predicted execution time of the simulation task corresponding to the task features based on the input task features.
[0135] The second prediction model can adopt any model for execution time prediction and can be implemented based on a variety of deep learning neural network structures, such as fully connected neural networks (Fully Connected Neural Networks), convolutional neural networks (CNN), recurrent neural networks (RNN) and their variants such as long short-term memory networks (LSTM), etc. This application does not limit this.
[0136] Exemplarily, during the training process of the second prediction model, linear regression is selected to preliminarily predict the execution time, and the LSTM deep learning model is used to capture the time dependency and nonlinear relationship in the execution of the simulation task. The data set is divided into a training set and a validation set, and the training set is used for model training, and the validation set is used to evaluate the model performance. After the training is completed, the second prediction model can capture the time dependency and nonlinear relationship in the execution of the simulation task, thereby predicting the execution time of the simulation task.
[0137] In this implementation, the predicted execution time of the simulation task output by the second prediction model obtained by training the second training sample can be understood as the expected execution time required to complete the simulation task to be assigned in a specific execution environment (such as a pre-set, standardized simulation computing node). This implementation provides a standardized execution time evaluation.
[0138] As an optional implementation, the predicted execution time may also be the shortest execution time of the estimated execution time of the simulation task to be assigned on each candidate simulation computing node. In this implementation, when predicting the execution time, in addition to the task characteristics of the simulation task to be assigned itself, the current state information of each candidate computing node also needs to be considered.
[0139] The execution time of the same simulation task executed by simulation computing nodes with different current status information may not be the same. The load of the simulation computing node may affect the execution efficiency of the simulation task, and the health status of the simulation computing node may determine whether it can stably execute the task. Optionally, the current status information also includes the current task execution status, which can be understood as the execution status of the task being executed on the simulation computing node. Specifically, the current task execution status may include key information such as whether the task is being executed, the progress of the execution, whether errors or blockages are encountered, etc.
[0140] In this implementation, step B2 "determining the predicted execution time of the simulation task to be assigned according to the task characteristics" includes steps C1-C2:
[0141] C1. According to the task characteristics and the current state information of each candidate simulation computing node, predict the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned.
[0142] Specifically, after determining the task characteristics of the simulation task to be assigned according to the configuration parameters of the simulation task to be assigned, and obtaining the current status information of each candidate simulation computing node, the simulation task scheduling device performs a second processing, comprehensively analyzing the task characteristics of the simulation task to be assigned and the current status information of each candidate simulation computing node to determine the predicted execution time of the simulation task to be assigned.
[0143] The second processing includes multiple implementation methods, such as using experience data, historical execution records or prediction models, etc., which is not limited in this application.
[0144] As an optional implementation method, the second processing can be implemented based on the processing idea of the model. Specifically, the task characteristics of the simulation task to be assigned and the current state information of each candidate simulation computing node are input into a pre-trained first prediction model, and the first prediction model predicts the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned based on the task characteristics and the current state information of each candidate simulation computing node.
[0145] In the embodiment of the present application, a first prediction model is pre-trained. The first prediction model can be obtained by performing execution time prediction training on the first training sample. The first training sample includes sample task characteristics, sample current state information of each simulation computing node and annotated execution time, and the annotated execution time includes the correct execution time corresponding to each simulation computing node in the first training sample when executing the simulation task corresponding to the sample task characteristics. The first training sample can use an existing public data set, or create and annotate a data set by itself. The execution time prediction training aims to enable the first prediction model to predict the execution time corresponding to each simulation computing node when executing the simulation task corresponding to the task characteristics based on the input task characteristics and the current state information of each simulation computing node.
[0146] As for other contents of the first prediction model, reference may be made to the specific contents of the second prediction model in the previous embodiment, which will not be elaborated here.
[0147] C2. Determine the shortest execution time among the execution times corresponding to the candidate simulation computing nodes, and determine the shortest execution time as the predicted execution time of the simulation task to be assigned.
[0148] Specifically, after determining the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned, the execution time corresponding to each candidate simulation computing node is compared, and the shortest execution time among the execution times corresponding to each candidate simulation computing node is selected as the predicted execution time of the simulation task to be assigned.
[0149] As an optional implementation, Figure 3 As shown, the simulation task scheduling method provided in the embodiment of the present application also includes steps S301-S302:
[0150] S301. Determine the task type of the simulation task to be assigned according to the configuration parameters.
[0151] The specific content of determining the task type of the simulation task to be assigned according to the configuration parameters can refer to the relevant content of step B1 in the above embodiment, which will not be described in detail here.
[0152] S302: Determine a data monitoring strategy for the simulation task to be assigned according to the task type of the simulation task to be assigned, so that data monitoring of the simulation task to be assigned throughout its life cycle is performed according to the data monitoring strategy for the simulation task to be assigned.
[0153] The data monitoring strategy can be understood as a strategy for monitoring and managing the data involved in the simulation task, covering the entire life cycle of the simulation task, and may include but is not limited to monitoring frequency, monitoring granularity, monitoring data points, data storage and access methods, etc.
[0154] Different simulation task types (such as structural simulation, fluid simulation, thermal simulation, etc.) differ in data processing, calculation models, and expected results. The types, scales, and frequencies of data generated during the simulation process may be different. These differences will directly affect the focus and methods of data monitoring. For example, fluid simulation may need to monitor dynamic data such as flow field velocity and pressure, while structural simulation may focus more on static data such as stress and strain. For another example, for computationally intensive fluid simulation tasks, the collection agent will increase the frequency of data collection to obtain more detailed CPU and memory usage data. For low-load simulation tasks, the collection frequency will be appropriately reduced to save system resources. Therefore, using different data monitoring strategies for simulation tasks of different task types can ensure effective data monitoring of the entire life cycle of the simulation task.
[0155] Optionally, a mapping relationship between each task type and each data monitoring strategy is preset, and the data monitoring strategy corresponding to the task type of the simulation task to be assigned can be determined according to the task type of the simulation task to be assigned and the mapping relationship.
[0156] In this implementation, the data monitoring strategy is dynamically adjusted according to the simulation task type, which can adapt to the specific requirements of different types of tasks. Compared with the traditional fixed-frequency data acquisition method, it can more effectively handle different types of simulation task requirements.
[0157] As an optional implementation, the simulation task scheduling method provided in the embodiment of the present application also includes simulation task anomaly detection and automatic repair, which specifically includes the following contents:
[0158] Anomaly detection: Implement the receipt of logs collected from the monitoring and acquisition module, including simulation data synchronization anomalies, simulation task reception anomalies, task scheduling anomalies, and related heterogeneous integration system interface call anomalies, etc. Use time series-based anomaly detection and clustering-based anomaly detection to improve the accuracy of anomaly detection.
[0159] Automated repair: After an anomaly is detected, the repair mechanism is automatically triggered to reallocate simulation computing resources, re-execute failed tasks, or perform fault recovery processes to ensure the smooth completion of simulation tasks.
[0160] Self-optimization: After the repair is completed, feedback information on the repair effect is collected and compared with the pre-repair data to evaluate the effectiveness of the repair strategy. By analyzing the repair history and anomaly occurrence patterns, the anomaly detection algorithm and repair strategy are continuously optimized.
[0161] Exemplary Devices
[0162] Corresponding to the above-mentioned simulation task scheduling method, an embodiment of the present application also provides a simulation task scheduling device. Figure 4 is a structural diagram of a simulation task scheduling device provided in an embodiment of the present application, such as Figure 4 As shown, the simulation task scheduling device provided in the embodiment of the present application is applied to a simulation system, wherein the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node being used to execute a simulation task assigned to the simulation computing node; the device includes:
[0163] The first unit 401 is used to obtain configuration parameters of the simulation task to be assigned, and obtain current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information;
[0164] The second unit 402 is used to determine the task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time;
[0165] The third unit 403 is used to allocate corresponding candidate simulation computing nodes to the simulation task to be allocated according to the task attribute information and the current state information of each candidate simulation computing node, so that the candidate simulation computing nodes allocated to the simulation task to be allocated perform the simulation task to be allocated.
[0166] The simulation task scheduling device provided by the present application is applied to a simulation system including a simulation computing cluster. By comprehensively considering the task attribute information of the simulation task to be assigned, including the predicted execution time, and the current status information of all candidate simulation computing nodes, more refined resource allocation can be achieved. Not only can the simulation task be assigned to a node that meets its execution requirements, but it can also avoid excessive concentration or waste of resources, achieve balanced utilization of resources in the simulation computing cluster, avoid overload or idleness of simulation computing nodes, thereby improving the resource utilization of the entire simulation computing cluster. In addition, by optimizing resource allocation, the present application can shorten the waiting time and execution time of simulation tasks, thereby improving the overall execution efficiency of simulation tasks and enhancing user experience.
[0167] Optionally, the task attribute information further includes: task urgency and / or task priority, wherein the task urgency indicates the urgency with which the assigned simulation task needs to be executed, and the task priority indicates the degree to which the assigned simulation task needs to be executed with priority.
[0168] Optionally, the second unit 402 includes:
[0169] A task feature determination subunit, used to determine the task feature of the simulation task to be assigned according to the configuration parameters, wherein the task feature includes at least one of the task type, task complexity and computing resource requirement;
[0170] The execution time prediction subunit is used to determine the predicted execution time of the simulation task to be assigned according to the task characteristics.
[0171] Optionally, the current state information also includes current task execution state information, and the execution time prediction subunit can be specifically used for:
[0172] According to the task characteristics and the current state information of each candidate simulation computing node, predict the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned;
[0173] The shortest execution time is determined among the execution times corresponding to the candidate simulation computing nodes, and the shortest execution time is determined as the predicted execution time of the simulation task to be assigned.
[0174] Optionally, the execution time prediction subunit may be specifically used for:
[0175] The task characteristics and the current state information of each candidate simulation computing node are input into a pre-trained first prediction model, and the first prediction model predicts the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned based on the task characteristics and the current state information of each candidate simulation computing node.
[0176] Optionally, the simulation task to be assigned includes multiple ones, and the third unit 403 can be specifically used for:
[0177] Determine the allocation priority of each simulation task to be allocated according to the task attribute information of each simulation task to be allocated, and determine the performance index of each candidate simulation computing node according to the current state information of each candidate simulation computing node;
[0178] The preset allocation strategy is executed, and the corresponding candidate simulation computing node is allocated to each simulation task to be allocated according to the allocation priority of each simulation task to be allocated and the performance index of each candidate simulation computing node.
[0179] Optionally, the device further comprises:
[0180] The fourth unit is used to determine the task type of the simulation task to be assigned according to the configuration parameters, and to determine the data monitoring strategy for the simulation task to be assigned according to the task type of the simulation task to be assigned, so that the data of the simulation task to be assigned is monitored throughout its life cycle according to the data monitoring strategy of the simulation task to be assigned.
[0181] The simulation task scheduling device provided in this embodiment belongs to the same application concept as the simulation task scheduling method provided in the above embodiments of this application, and can execute the simulation task scheduling method provided in any of the above embodiments of this application, and has the corresponding functional modules and beneficial effects of executing the simulation task scheduling method. For the technical details not described in detail in this embodiment, please refer to the specific processing content of the simulation task scheduling method provided in the above embodiments of this application, which will not be repeated here.
[0182] The functions implemented by the above first unit 401, second unit 402 and third unit 403 may be implemented by the same or different processors, respectively, and are not limited in the embodiment of the present application.
[0183] It should be understood that the units in the above devices can be implemented in the form of a processor calling software. For example, the device includes a processor, the processor is connected to a memory, and instructions are stored in the memory. The processor calls the instructions stored in the memory to implement any of the above methods or realize the functions of each unit of the device, wherein the processor can be a general-purpose processor, such as a CPU or a microprocessor, etc., and the memory can be a memory in the device or a memory outside the device. Alternatively, the units in the device can be implemented in the form of hardware circuits, and the functions of some or all units can be realized by designing the hardware circuits. The hardware circuit can be understood as one or more processors; for example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are realized by designing the logical relationship of the components in the circuit; for another example, in another implementation, the hardware circuit can be implemented by PLD, taking FPGA as an example, which can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by the configuration file, so as to realize the functions of some or all of the above units. All units of the above devices can be implemented in the form of a processor calling software, or in the form of hardware circuits, or in part by a processor calling software, and the remaining part is implemented in the form of hardware circuits.
[0184] In an embodiment of the present application, a processor is a circuit with the ability to process signals. In one implementation, the processor may be a circuit with the ability to read and run instructions, such as a CPU, a microprocessor, a GPU, or a DSP; in another implementation, the processor may implement certain functions through the logical relationship of a hardware circuit, and the logical relationship of the hardware circuit is fixed or reconfigurable, such as a hardware circuit implemented by an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document to implement the hardware circuit configuration can be understood as the process of the processor loading instructions to implement the functions of some or all of the above units. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, DPU, etc.
[0185] It can be seen that each unit in the above device can be one or more processors (or processing circuits) configured to implement the above method, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.
[0186] In addition, all or part of the units in the above device can be integrated together, or can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a SOC. The SOC may include at least one processor for implementing any of the above methods or implementing the functions of each unit of the device. The type of the at least one processor may be different, for example, including a CPU and an FPGA, a CPU and an artificial intelligence processor, a CPU and a GPU, etc.
[0187] Exemplary Electronic Devices
[0188] Another embodiment of the present application also provides a simulation task scheduling device, see Figure 5 As shown, the device includes: a memory 200 and a processor 210; wherein the memory 200 is connected to the processor 210 and is used to store programs; the processor 210 is used to implement the simulation task scheduling method disclosed in any of the above embodiments by running the program stored in the memory 200.
[0189] Specifically, the above simulation task scheduling device may further include: a bus, a communication interface 220, an input device 230 and an output device 240. The processor 210, the memory 200, the communication interface 220, the input device 230 and the output device 240 are interconnected via the bus.
[0190] A bus may include a pathway that transfers information between components of a computer system.
[0191] The processor 210 may be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits for controlling the execution of the program of the scheme of the present invention. It may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0192] The processor 210 may include a main processor, and may also include a baseband chip, a modem, and the like.
[0193] The memory 200 stores a program for executing the technical solution of the present invention, and may also store an operating system and other key services. Specifically, the program may include a program code, and the program code includes a computer operation instruction. More specifically, the memory 200 may include a read-only memory (ROM), other types of static storage devices that can store static information and instructions, a random access memory (RAM), other types of dynamic storage devices that can store information and instructions, a disk storage, a flash, and the like.
[0194] The input device 230 may include a device for receiving data and information input by a user, such as a keyboard, a mouse, a camera, a scanner, a light pen, a voice input device, a touch screen, a pedometer or a gravity sensor, etc. The output device 240 may include a device that allows information to be output to the user, such as a display screen, a printer, a speaker, etc. The communication interface 220 may include a device using any transceiver to communicate with other devices or communication networks, such as Ethernet, a radio access network (RAN), a wireless local area network (WLAN), etc. The processor 210 executes the program stored in the memory 200, and calls other devices, which can be used to implement the various steps of any simulation task scheduling method provided in the above embodiments of the present application.
[0195] An embodiment of the present application also proposes a chip, which includes a processor and a data interface. The processor reads and runs the program stored in the memory through the data interface to execute the simulation task scheduling method introduced in any of the above embodiments. The specific processing process and its beneficial effects can be found in the introduction to the embodiment of the above simulation task scheduling method.
[0196] Exemplary computer program products and storage media
[0197] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the simulation task scheduling method according to various embodiments of the present application described in any of the above embodiments of this specification.
[0198] The computer program product may be written in any combination of one or more programming languages to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0199] In addition, the embodiment of the present application may also be a storage medium on which a computer program is stored. The computer program is executed by a processor to execute the steps of the simulation task scheduling method according to various embodiments of the present application described in any of the above embodiments of this specification, and specifically the following steps may be implemented:
[0200] S201, obtaining configuration parameters of a simulation task to be assigned, and obtaining current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information.
[0201] S202. Determine task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time.
[0202] S203. According to the task attribute information and the current state information of each candidate simulation computing node, the corresponding candidate simulation computing node is assigned to the simulation task to be assigned, so that the candidate simulation computing node assigned to the simulation task to be assigned executes the simulation task to be assigned.
[0203] For the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0204] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The steps in the methods of each embodiment of the present application can be adjusted, merged and deleted in sequence according to actual needs, and the technical features recorded in each embodiment can be replaced or combined. The modules and sub-modules in the devices and terminals of each embodiment of the present application can be merged, divided and deleted according to actual needs. In several embodiments provided in this application, it should be understood that the disclosed terminals, devices and methods can be implemented in other ways. For example, the terminal embodiment described above is only schematic, for example, the division of modules or sub-modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple sub-modules or modules can be combined or integrated into another module, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or module, which may be electrical, mechanical or other forms.
[0205] The modules or submodules described as separate components may or may not be physically separated, and the components of the modules or submodules may or may not be physical modules or submodules, that is, they may be located in one place, or they may be distributed on multiple network modules or submodules. Some or all of the modules or submodules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, each functional module or submodule in each embodiment of the present application may be integrated in a processing module, or each module or submodule may exist physically alone, or two or more modules or submodules may be integrated in one module. The above-mentioned integrated modules or submodules may be implemented in the form of hardware or in the form of software functional modules or submodules.
[0206] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application. The steps of the method or algorithm described in conjunction with the embodiments disclosed herein can be implemented directly with a software unit executed by a hardware or processor, or a combination of the two. The software unit can be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the technical field.
[0207] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements. The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A simulation task scheduling method, It is characterized in that Applied to a simulation system, the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node is configured to execute a simulation task assigned to the simulation computing node; the method includes: Obtaining configuration parameters of the simulation task to be assigned, and obtaining current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information; Determining task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time; According to the task attribute information and the current state information of each candidate simulation computing node, the corresponding candidate simulation computing node is assigned to the simulation task to be assigned, so that the candidate simulation computing node assigned to the simulation task to be assigned executes the simulation task to be assigned.
2. The method according to claim 1, It is characterized in that The task attribute information also includes: task urgency and / or task priority, wherein the task urgency indicates the urgency with which the simulation task to be assigned needs to be executed, and the task priority indicates the degree to which the simulation task to be assigned needs to be executed with priority.
3. The method according to claim 1, It is characterized in that Determining task attribute information of the simulation task to be assigned according to the configuration parameters includes: Determining, according to the configuration parameters, the task characteristics of the simulation task to be assigned, wherein the task characteristics include at least one of the task type, task complexity, and computing resource requirements; According to the task characteristics, the predicted execution time of the simulation task to be assigned is determined.
4. The method according to claim 3, It is characterized in that The current status information also includes the current task execution status; Determining the predicted execution time of the simulation task to be assigned according to the task characteristics includes: According to the task characteristics and the current state information of each candidate simulation computing node, predicting the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned; The shortest execution time is determined among the execution times corresponding to the candidate simulation computing nodes, and the shortest execution time is determined as the predicted execution time of the simulation task to be assigned.
5. The method according to claim 4, It is characterized in that According to the task characteristics and the current state information of each candidate simulation computing node, predicting the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned, including: The task characteristics and the current state information of each candidate simulation computing node are input into a pre-trained first prediction model, and the first prediction model predicts the execution time corresponding to each candidate simulation computing node executing the simulation task to be assigned based on the task characteristics and the current state information of each candidate simulation computing node.
6. The method according to claim 1, It is characterized in that The simulation tasks to be assigned include a plurality of tasks, and according to the task attribute information and the current state information of each candidate simulation computing node, the corresponding candidate simulation computing nodes are assigned to the simulation tasks to be assigned, including: Determine the allocation priority of each simulation task to be allocated according to the task attribute information of each simulation task to be allocated, and determine the performance index of each candidate simulation computing node according to the current state information of each candidate simulation computing node; The preset allocation strategy is executed, and the corresponding candidate simulation computing node is allocated to each simulation task to be allocated according to the allocation priority of each simulation task to be allocated and the performance index of each candidate simulation computing node.
7. The method according to any one of claims 1 to 6, It is characterized in that The method further comprises: Determining the task type of the simulation task to be assigned according to the configuration parameters; According to the task type of the simulation task to be assigned, a data monitoring strategy for the simulation task to be assigned is determined, so that data monitoring of the simulation task to be assigned throughout its life cycle is performed according to the data monitoring strategy for the simulation task to be assigned.
8. A simulation task scheduling device, It is characterized in that Applied to a simulation system, the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node is used to execute a simulation task assigned to the simulation computing node; the device includes: The first unit is used to obtain configuration parameters of the simulation task to be assigned, and obtain current status information of each candidate simulation computing node, wherein the current status information includes current load information and / or current health status information; A second unit is used to determine task attribute information of the simulation task to be assigned according to the configuration parameters, wherein the task attribute information includes a predicted execution time; The third unit is used to allocate corresponding candidate simulation computing nodes to the simulation task to be assigned according to the task attribute information and the current state information of each candidate simulation computing node, so that the candidate simulation computing nodes allocated to the simulation task to be assigned execute the simulation task to be assigned.
9. A simulation task scheduling device, It is characterized in that Applied to a simulation system, the simulation system includes a simulation computing cluster composed of a plurality of simulation computing nodes, each simulation computing node is used to execute a simulation task assigned to the simulation computing node; the device includes a memory and a processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the simulation task scheduling method as described in any one of claims 1 to 7 by running the program in the memory.
10. A simulation system, It is characterized in that include: The simulation task scheduling device, simulation computing cluster, input module, and monitoring module according to claim 9, wherein the input module, the monitoring module, and each simulation computing node in the simulation computing cluster are respectively connected to the simulation task scheduling device for communication; The input module is used to receive input information from a user, wherein the input information includes configuration parameters of the simulation task to be assigned; The simulation computing cluster includes a plurality of simulation computing nodes, each of which is configured to execute a simulation task assigned to the simulation computing node; The monitoring module is used to collect current status information of each simulation computing node, and the current status information includes current load information and / or current health status information.
Citation Information
Cited By
Simulation method and device based on atomization simulation case and computer equipment
CN119203626A
Task scheduling method and system for scheduling platform
CN119597428A
Heterogeneous simulation environment protocol interconnection and intercommunication method and heterogeneous simulation system
CN119603374A
A method for protocol interoperability in a heterogeneous simulation environment and a heterogeneous simulation system
CN119603374B
Embedded multi-core operating system scheduling method and device and storage medium
CN120909738A
