JP2015026132A - Resource control device, method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow a user to perform comfortable resource allocation for the user dynamically in accordance with a use situation of the user.SOLUTION: When a resource allocation request of a task intended by a user is issued, on the basis of the resource allocation request, use situation information of a calculation resource of resource use situation information storage means and resource allocation selection list storage means are referred to, and a resource allocation method is selected. When the vacancy of a calculation resource of the resource allocation request exceeds a threshold depending on priority, the calculation resource is allocated, and when the vacancy of the calculation resource is less than the threshold depending on priority, another calculation resource having vacancy is allocated.

Description

  The present invention relates to a resource control technique, and more particularly to a resource control apparatus, method, and program for controlling allocation of computing resources for processing a task in accordance with a user's intention.

  In addition to the CPUs that have been used as computational resources in the past, tasks that use various computational resources such as GPUs are increasing year by year. However, depending on the combination of calculation resource allocation, the processing speed of tasks with high importance may be reduced. Therefore, in a machine having a plurality of computing resources, a technology for dynamically allocating computing resources to tasks (see, for example, Non-Patent Document 1) and GPU resources are time-divided for tasks that use GPUs. A technique (for example, see Non-Patent Document 2) that enables setting the priority of use is known.

AUGONNET, C., THIBAULT, S., NAMYST, R., AND WACRENIER, P.-A.StarPU: a unified platform for task scheduling on heterogeneous multicore architectures.Proc. Of Concurrency and Computation: Practice and Experience 23, 2 ( feb 2011), 187-198. KATO, S., MCTHROW, M., MALTZAHN, C., AND BRANDT, S. Gdev: First-Class GPU Resource Management in the Operating System.In Proc. Of USENIX ATC (2012).

  However, since the conventional technology aims to allocate computing resources to tasks according to a specific target on a machine having a plurality of different computing resources, the usage status of the user and the user's intention (the importance of the task indicated by the user) (Priority)) resource allocation is not performed. In the prior art disclosed in Non-Patent Document 1, a plurality of allocation algorithms such as allocation of calculation resources aiming at maximizing the calculation amount per unit time can be used, but the target environment is high performance computing. For this reason, the user's usage status and preferences are not taken into consideration when assigning.

  In the prior art disclosed in Non-Patent Document 2, the priority of each task can be set and the GPU resource can be allocated, but the resource cannot be allocated between different types of calculation resources. That is, it is possible to statically specify the task priority according to the usage status and preferences of the user, but it does not have a function to specify dynamically. Therefore, even when calculation resources are allocated using the prior art disclosed in Non-Patent Document 1 and Non-Patent Document 2, there is a possibility that calculation resources appropriate for a task important to the user may not be allocated. The resources used by the tasks important to the user are simultaneously used by other less important tasks, and as a result, the computational resources available for the important tasks may be reduced. For this reason, even when optimal resource allocation can be realized in accordance with a specific target, there is a problem that the user is not always comfortable by those resource allocation methods.

  The present invention has been made in view of the above points, and allows the user to dynamically allocate more comfortable resources according to his / her usage situation according to his / her usage situation, which is important task processing for the user. It is an object of the present invention to provide a resource control apparatus, method, and program capable of preventing the speed reduction.

One aspect of the present invention is a resource control apparatus that allocates a plurality of different types of computing resources for processing tasks,
A resource allocation method selection list storage means storing the resource allocation method;
Resource usage status information storage means for storing usage status information of computing resources;
When a resource allocation request for a task including a user's intention is given, a resource allocation method selection unit that selects a resource allocation method from the resource allocation method selection list storage unit according to a usage mode of a device that executes the task;
If the calculation resource vacancy specified in the resource allocation request with reference to the resource usage status information storage means exceeds a threshold according to the priority, the calculation resource is allocated, and the calculation resource vacancy is a priority. Resource allocation means for allocating other computational resources that are free,
A resource control device is provided.

  According to one aspect, based on a resource allocation method according to the usage status of each task by the user and the user's intention for each task, the computing resource is allocated together with the management of the resource usage status by the resource control device. Thus, the resources used by each task can be distributed in a more comfortable manner for the user.

It is an example of system configuration in a 1st embodiment of the present invention. It is a structural example of the resource control apparatus in the 1st Embodiment of this invention. It is an example of the resource allocation selection list memory | storage part in the 1st Embodiment of this invention. It is an example of the resource utilization condition information storage part in the 1st Embodiment of this invention. It is a flowchart of the process in the 1st Embodiment of this invention. It is an example of resource allocation in the 1st Embodiment of this invention. It is an example of system configuration in a 2nd embodiment of the present invention. It is an example of resource allocation in the 2nd Embodiment of this invention. It is a flowchart of the process in the 2nd Embodiment of this invention. It is a structural example of the system in which the resource control apparatus exists in a different machine in another embodiment of the present invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
FIG. 1 shows an example of a system configuration in the first embodiment of the present invention.

  A machine 1 shown in FIG. 1 includes middleware 10 and a resource group 20.

  The middleware 10 includes the resource control device 100 illustrated in FIG.

  The resource control device 100 is a device that selects a computing resource 200 to be used for task processing on the machine 1 composed of a plurality of different computing resources 200, and includes a user interface unit 110, a resource allocation method selection unit 120, and a resource allocation intention. It has a confirmation unit 130, a resource allocation unit 140, a resource usage management unit 150, a resource allocation selection list storage unit 160, a resource usage status information storage unit 170, and a resource allocation method list storage unit 180.

  The user interface unit 110 transmits and receives data and messages between the resource control apparatus 100 and a user terminal (not shown).

  The resource allocation method selection unit 120 acquires a resource allocation request {task, resource} input from the user, and selects a resource allocation method from the resource allocation method list storage unit 180.

  The resource allocation intention confirmation unit 130 follows the resource allocation method for the task selected by the resource allocation method selection unit 120, for example, when it is necessary to confirm the user's intention, the user intention via the user interface unit 110. Confirm. Also, for example, when selecting automatically, resource allocation information to a task is extracted according to the priority of the task in the resource allocation selection list storage unit 160 and the resource usage status of the resource usage status information storage unit 170. .

  The resource allocation unit 140 assigns a task to a task based on the allocation resource according to the user's intention, the resource usage status in the resource usage status information storage unit 170, and the resource priority for each task in the resource allocation selection list storage unit 160. Allocate the best resources.

  When a computing resource is assigned to a task, the resource usage management unit 150 executes the task in the resource usage status information storage unit 170 in which the resource is executed (used / unused), and the task using the resource Update the name and its usage rate.

  2 shows an example in which the resource use management unit 150 and the resource use state information storage unit 170 are provided in the resource control apparatus 100, but the resource control apparatus 100 is not limited to this example. It may be provided outside.

  As shown in FIG. 3, the resource allocation selection list storage unit 160 stores calculation resources and their priorities for each task.

  As shown in FIG. 4, the resource usage status information storage unit 170 stores the usage status of calculation resources. In the example of FIG. 4, “1” indicates that the resource is being used, and “0” indicates that the resource is not being used. Also, as shown in FIG. 4, the resource utilization rate for each task is stored as the resource utilization status. In the example of the first line in FIG. 4, the resource (CPU1) indicates that 50% of the resource is used by the task (Job1).

In the resource allocation method list storage unit 180, for example,
a. Determine whether the resource is in use, and if so, present the resource not in use and let the user select it;
b. Determine whether the resource (for example, CPU) is in use. If it is in use, N resources in the order of priority from other unused resources of the same type (for example, CPU) Present resource candidates and let the user choose;
c. Determine whether the resource (for example, CPU) is in use, and if it is in use, N cases from other unused heterogeneous (for example, GPU) resources in descending order of priority Present resource candidates and let the user choose;
d. Determine whether the resource (for example, CPU) is in use, and if so, select the resource with the highest priority from other unused resources of the same type (for example, CPU). Select automatically;
e. Determine whether the resource (eg, CPU) is in use, and if so, select the resource with the highest priority from other unused heterogeneous (eg, GPU) resources. Select automatically;
Are stored.

  In the above configuration, the resource allocation unit 140 acquires the user intention confirmed by the resource allocation intention confirmation unit 130 and the resource usage status at that time from the resource usage status information storage unit 170, and performs optimal resource allocation. . Here, the resource allocation according to the user's intention means that the resource control apparatus 100 allocates a resource desired by the user from a task having a high priority according to the user's resource allocation request {task, resource}. means.

  Hereinafter, processing of the resource control device 100 having the above configuration will be described.

  FIG. 5 is a flowchart of processing in the first embodiment of the present invention.

  Step 100) When a resource allocation request {task, resource} specifying a task and a resource to be allocated to the task is input from the user, the user interface unit 110 acquires this and outputs it to the resource allocation method selection unit 120 To do. The resource allocation method selection unit 120 selects a rule set from the resource allocation method selection list storage unit 180 according to the usage mode of the machine 1.

  Step 110) The resource allocation intention confirmation unit 130 determines whether the calculation resource matches the user's intention according to the rule set selected by the resource allocation method selection unit 120. There is a method for prompting the user to make a determination via the user interface unit 110 to determine whether or not the user's intention is met. When resource allocation is automatically performed based on the resource allocation request, the process proceeds to step 120.

  Step 120) When the resource allocation intention confirmation unit 130 inputs that the user's intention is matched, or when the resource allocation is performed automatically, the resource allocation unit 140 performs resource allocation. With reference to the selection list storage unit 160, a calculation resource having a high priority for the task to use the resource is selected. In the resource allocation selection list storage unit 160, for example, the priority for using the resource A by the task is 50, and when the resource A is difficult to use, the priority for using the resource B is set as 20,. The computing resource is selected according to the priority. Further, the resource usage status information storage unit 170 is referred to determine whether the free status of the calculation resource exceeds a threshold value based on the relative priority between tasks. Here, the relative priority is expressed by a value from 1 to 100, and the priority of a task to be newly executed is set based on the priority of resources already used for execution by the denominator. It is expressed as a fraction (or decimal) as a numerator. If the relative priority exceeds 1, the priority of the task to be newly executed becomes higher. In this case, if the optimal computational resources of the task that is already being executed and the task that is to be newly executed are the same, the allocation of the resource is determined by looking at the free capacity (utilization rate) of the computational resource in the resource usage status information storage unit 170. . For example, when the utilization rate is 30% (free space is 70%), a task to be newly executed starts processing with the calculation resource, and a task already being executed is processed with another resource. The degree of utilization to be used as a threshold is determined by the user in advance or by considering how much utilization of each task is optimal for each resource.

  If the resource availability exceeds the threshold value based on the relative priority, the process proceeds to step 130.

  Step 130) The resource assignment unit 140 assigns a task to the resource and waits for the next request.

  Step 140) When the resource availability is equal to or less than the threshold based on the relative priority, the resource allocation unit 140 selects the resource that is the most available other than the resource from the resource usage status information storage unit 170.

  Step 150) The resource assignment unit 140 assigns a task to the resource selected in Step 140 and waits for the next request.

  Next, the above process will be specifically described along the above flowchart.

  In the present embodiment, resource allocation in the case where a user performs only two tasks, ie, a task of playing a moving image and a task of compressing a file, on a computer having one CPU and one GPU as resources will be described. . There are two methods of decoding for video playback: processing using the CPU and using the decoding support function implemented in the GPU, such as noise reduction and edge enhancement to improve video quality. As for effects, there are a method of processing by the CPU and a method of processing by the GPU. In addition, in a computer on which a dedicated chip is mounted for taking charge of the moving image reproduction process, there is a method of performing the reproduction process by the chip. Since it is well known to those skilled in the art that a CPU or GPU or the like can be used for moving image reproduction, detailed description thereof is omitted.

  Below, the case where GPU is used for animation reproduction is described.

  When the resource allocation method selection unit 120 acquires a resource allocation request from the user via the user interface 110, the resource allocation method selection unit 120 refers to the resource allocation method selection list storage unit 180 and selects a resource allocation method based on the resource name ( Step 100).

  For example, in the machine 1 in which one CPU and one GPU are mounted, the resource allocation method selection unit 120 determines whether the resource (for example, CPU) is in use as the resource allocation method described above. If selected, image editing is performed when N resource candidates are presented in descending order of priority from other unused heterogeneous (for example, GPU) resources and selected by the user. Assuming that the task related to and the video playback task are faster in GPU processing. When a task related to image editing has already been processed on the GPU, and the user wants to start a video playback task with a higher priority than a task related to image editing, the resource control device 100 is related to image editing. Hooks the task API (Application Programming Interface), switches to processing on the CPU, and processes high-priority video playback tasks on the GPU.

  The selection of the resource allocation method is not limited to the above example.

  Next, the process waits until a change in the user's usage status or a new resource allocation request occurs. If the generated event is a resource allocation request, the resource allocation intention confirmation unit 130 stores the resource allocation method selection list. It is checked in section 160 whether a user's preferred assignment method for the requested task is defined. For example, according to a resource (“GPU” in this example) for a task of playing a moving image according to the user's intention defined in advance in the resource allocation method selection list, a GPU is allocated to the task. For example, by hooking the API of the video playback task, the task is set to be processed on the GPU. When the status of the resource used changes, the resource usage management unit 150 updates the resource status, task, and usage rate in the resource usage status information storage unit 170.

  Next, when the user tries to start the execution of the file compression task while playing the moving image, the file compression task requests a calculation resource from the resource allocation unit 140 (step 110). The resource allocation unit 140 allocates a resource candidate as a resource when the resource allocation method according to the user's intention exceeds the threshold according to the priority (Yes in step 120) when the resource candidate to be allocated exceeds the threshold according to the priority. (Step 130). On the other hand, if the available resource candidate allocated by the resource allocation method according to the user's intention is below the threshold according to the priority, select another available resource candidate different from the resource candidate (step 140), The most free computing resource is allocated (step 150).

  As an example, if it is requested to assign the same calculation resource, specifically “GPU”, to the “Video Playback Task” for the “File Compression Task” as the user ’s intention, In the resource allocation used by the task “task”, if the threshold derived by the relative priority of the “file compression task” with respect to the “video playback task” exceeds the free space of the GPU calculation resource, the GPU is allocated. Conversely, if the GPU space is not enough and the threshold is not reached, a CPU is allocated.

  Note that this threshold setting is statically specified. There are various methods such as dynamically changing according to the usage status of the computing resources of the entire machine, but the computing resources are compared by comparing the necessity of the computing resources based on the relative importance of the task and the availability of the computing resources. Any designation method may be used as long as it is possible to determine whether or not the allocation is possible.

  On the other hand, if the user's intention is to select the CPU allocation method for the “File compression task”, if the CPU usage status is not used by other tasks, there is enough free computing resources. Therefore, a CPU that is a calculation resource according to the user's intention is allocated.

  In this way, with respect to the allocation of calculation resources of the “file compression task”, it is possible to allocate a CPU that is a resource according to the user's intention.

  Regarding the calculation resource allocation of the “file compression task”, the task (in this case, “file compression task”) is not defined in the resource allocation method selection list storage unit 180, or the user has assigned the task to the task. When there is no particular resource allocation method according to the user's intention, such as when resource allocation is not considered important (other tasks may be allocated with the highest priority), the usage status of the resource usage status information storage unit 170 Depending on, CPUs that are more computational resources are allocated. In a situation where the user does not place importance on task allocation, it is not always necessary to select a large free resource, and the resource allocation intention confirmation unit 130 may prompt the user to input an instruction on which resource to allocate. Is possible.

  In addition, in response to a resource allocation request, for a task that does not have a resource allocation method according to the user's intention, a method of allocating a calculation resource with a larger vacancy according to the usage status of the usage status information storage unit 170 There are various methods, such as a method for assigning calculation resources and a method for assigning available calculation resources in order, but it is important for the user, and a task for which a resource assignment method according to the intention of assignment of calculation resources is defined. Any method may be used as long as it is possible to prevent a decrease in the processing speed of those important tasks.

  In addition, when a resource allocation method is defined in accordance with the user's intention on the assumption that video playback is extremely important for other tasks for the user (for example, the priority of the resource allocation selection list storage unit 160) Since the GPU is allocated to video playback and the CPU is allocated to file compression, the GPU's computing power can be maximized for each task of the operation playback task. In the example of FIG. 6, real-time (frame rate 60fps) playback of moving images is possible. Therefore, it is possible to prevent the occurrence of an event that the user feels uncomfortable due to a decrease in the processing capability related to moving image reproduction.

  On the other hand, in the following cases, the CPU is allocated to the task of the file compression processing in exchange for the frame rate decrease (35 fps) of the video playback by allocating the GPU to the file compression.

When a resource allocation method according to the user's intention is defined as video playback is not particularly important for the user;
・ When the resource allocation method according to the user's intention is not defined;
-When the resource allocation method according to the user's intention is defined as file compression is important;
This reduces the time required for file compression processing from 12 minutes to 9 minutes and 30 seconds compared to assigning a CPU to the file compression processing task in exchange for lowering the frame rate (35fps) for video playback. Therefore, it is possible to reflect the user's intention to finish the file compression process quickly.

  In this embodiment, it is assumed that the user shifts the GPU allocation to the video playback. However, the present invention can be similarly implemented even when the user shifts the CPU allocation to the video playback. It is natural. Further, in this embodiment, in a machine on which one CPU and one GPU are mounted, calculation tasks are allocated by executing two tasks. However, when three or more tasks are executed or three calculation resources are allocated. Even if there are two or more, based on the definition of the calculation resource allocation method according to the user's intention for the task and the availability information of each calculation resource, the same as the case of two calculation resources and two tasks Of course it is possible.

[Second Embodiment]
FIG. 7 is a system configuration example according to the second embodiment of the present invention, and shows a configuration having a CPU 300 and a GPU 400 as calculation resources managed by the resource control device 100.

  In the present embodiment, an example will be described in which calculation resources of some tasks are fixed and a task that has been executed first is switched.

  Since the configuration of the resource control device 100 is the same as that of the first embodiment, the configuration diagram and the description thereof are omitted.

  In this embodiment, resource allocation when a 3D game is newly executed in a state where a user already uses a GPU for video encoding processing on a machine with one CPU and one GPU is described. To do.

  In video encoding processing, which has traditionally used CPUs and dedicated chips, in recent years the method of using a dedicated circuit on the GPU to speed up video encoding processing uses the GPU's general-purpose processing capabilities. A method for speeding up the encoding process of a moving image is used. Since it is well known to those skilled in the art that these GPU computing capabilities can be used for the encoding process of moving images, detailed description will be omitted. On the other hand, in 3D games, it is generally necessary to use a GPU for graphics processing. The GPU is originally a processor prepared for performing graphics processing, and since it is well known to those skilled in the art that it is necessary to use the GPU for 3D game execution, detailed description thereof is omitted.

  Since the processing flow is substantially the same as in the first embodiment, only the differences will be described.

  As shown in FIG. 8, in this embodiment, when a GPU is used as a calculation resource for video encoding processing, the frame rate of the game is reduced to 27 fps, and it is difficult for the user to play the game comfortably. By assigning a CPU to the encoding process, the frame rate of the game can be increased to 55 fps, and the user can play the game more comfortably. However, in the example of FIG. 8, when a CPU is assigned as a calculation resource for moving image encoding processing, the time required for moving image encoding processing is extended from 25 minutes to 70 minutes.

  FIG. 9 is a flowchart of processing in the second embodiment of the present invention.

  Step 200) The resource allocation method selection unit 120 selects a resource allocation method from the resource allocation method selection list storage unit 180 according to the usage mode of the machine 1.

  Step 210) Wait until a change in user usage, a new resource allocation request, or a fixed resource allocation notification with high importance occurs, and if the event that occurred is a resource allocation request or resource allocation notification, The resource allocation intention confirmation unit 130 confirms the user's preferred resource allocation for the task that has issued the resource allocation request or the resource allocation notification.

  Step 220) The resource allocation unit 140 refers to the resource usage status information storage unit 170, and determines whether there is sufficient free space in the favorite resource of the target task or the allocation notification destination resource. If there is a vacancy, the process proceeds to step 230. If not, the process proceeds to step 240.

  Step 230) The resource allocation unit 140 allocates a preferred calculation resource to the task and waits for the next resource allocation request or resource allocation notification.

  Step 240) The resource allocation unit 140 determines whether there is any other task using the resource that has a low relative priority with the task. If there is, the process proceeds to Step 250. Goes to step 260.

  Step 250) The resource allocation unit 140 changes the allocated resource of the task having the lowest relative priority to one having the most free resources other than the resource, and proceeds to Step 220.

  Step 260) The resource allocation unit 140 determines whether the resource allocation of the target task can be changed. If the resource allocation can be changed, the process proceeds to Step 270. If the resource allocation cannot be changed, the process proceeds to Step 230.

  Step 270) The resource allocation unit 140 selects the most available calculation resource other than the corresponding calculation resource.

  Step 280) The resource allocation unit 140 allocates the most available calculation resource and waits for the next resource allocation request or resource allocation notification.

  Hereinafter, specific processing will be described along the above flowchart.

  When a user starts a game task and acquires a resource allocation request to try to start executing a game task in the middle of executing a video encoding process, the resource allocation intention confirmation unit 130 Confirm that it is defined that the processing speed of the game is prioritized as a method of allocating the calculation resource according to the intention of (Step 210), and the resource allocation unit 140 performs the task of the encoding process of the video by the CPU, For the task of encoding the moving image, the allocated computing resource is switched from the GPU to the CPU (step 230). It is well known to those skilled in the art that regarding calculation resource switching between different calculation resources, although there are restrictions such as temporary interruption of processing, such as process migration, calculation resources can be switched by various methods. Therefore, the description is omitted. On the other hand, if the processing speed of the game task is not so important for the user in the usage situation (step 240, No), the GPU is continuously used without changing the calculation resource allocation for the task of the video encoding process ( (Step 260, No, Step 230), the processing time of the moving image encoding process can be made relatively shorter than when the assignment time is changed to the CPU. In addition, when a GPU is assigned to the game and it is determined that the importance is low (there is a low relative priority) (step 240, Yes), the computing resource is changed from the GPU to the CPU (step 250).

  In the present embodiment, it has been described that the present invention is applied even when two different tasks are executed on a machine having two different calculation resources and the calculation resource cannot be switched for one of the tasks. Even when a machine having more than one computing resource or three or more tasks are executed, as long as the computing resources of some tasks can be switched, the same as in the second embodiment Of course, it can be implemented.

  In the first and second embodiments described above, an example in which a CPU or GPU mounted on the machine 1 is used as a plurality of different calculation resources is not limited to this example. As a computing resource that can be used in the present invention, another processor or DSP having a computing ability capable of executing task processing may be used, and it is managed by a computing resource or a resource control device abstracted by a technology such as a virtual machine. It may be a computing resource mounted on a non-identical machine.

  Further, the available calculation resources may be divided and allocated as smaller calculation resources.

  The resource control apparatus 100 may not be executed on the same machine as the calculation resource 200, but may perform resource control by different machines and assign calculation resources as shown in FIG. Further, the resource control device 100 is not limited to a form that is executed on a single machine, and the resource devices may be operated cooperatively on a plurality of machines.

  The method of allocating resources to tasks according to the user's intention is not limited to a mode in which a user directly assigns a specific calculation resource to a specific task. Priority is given to the task related to image editing, and the processing speed is given priority. For example, the user's request with a high level of abstraction with respect to the importance (priority) of the task is input and the input is converted into a specific resource allocation method. It may be used above. For example, in a device equipped with one CPU and one GPU, it is assumed that the processing speed of an image editing task and a video playback task is higher when processing with the GPU. When a task related to image editing has already been processed on the GPU, and the user newly activates an operation playback task with a higher priority than a task related to image editing, the resource control device 100 performs a task related to image editing. Hook the API, switch to processing on the CPU, and process high-priority video playback tasks on the GPU.

  Furthermore, the number of simultaneous users of the machine in the present invention is not limited to one. When two or more users are using at the same time, the situation used by a plurality of specific users is used as a form of use situation, and resources are allocated to each task according to the situation and each user's intention. May be.

  As described above, in the present invention, the resource control device manages the use status of the calculation resource by using the user's intention to assign the calculation resource to each task according to the use status of the user's calculation resource. After confirming the resource allocation method according to the intention, the computing resources used by each task can be distributed in a more comfortable manner to the user.

  Also, the operations of the components of the resource control device 100 shown in FIG. 2 can be constructed as a program, installed in a computer used as the resource control device and executed, or distributed via a network.

  Further, the constructed program may be recorded on a computer-readable recording medium such as a CD-ROM, DVD-ROM, or flash memory.

  The present invention is not limited to the above-described embodiments, and various modifications and applications can be made within the scope of the claims.

1 machine 10 middleware 20 resource group 100 resource control device 110 user interface unit 120 resource allocation method selection unit 130 resource allocation intention confirmation unit 140 resource allocation unit 150 resource usage management unit 160 resource allocation selection list storage unit 170 resource usage status information storage unit 180 Resource allocation method list storage unit 200 Computing resource

Claims (7)

A resource control device that allocates a plurality of different types of computing resources for processing tasks,
A resource allocation method selection list storage means storing the resource allocation method;
Resource usage status information storage means for storing usage status information of computing resources;
When a resource allocation request for a task including a user's intention is given, a resource allocation method selection unit that selects a resource allocation method from the resource allocation method selection list storage unit according to a usage mode of a device that executes the task;
If the calculation resource vacancy specified in the resource allocation request with reference to the resource usage status information storage means exceeds a threshold according to the priority, the calculation resource is allocated, and the calculation resource vacancy is a priority. Resource allocation means for allocating other computational resources that are free,
A resource control apparatus comprising: A resource allocation selection list storage means storing resource allocation information for tasks;
Based on the resource allocation method selected by the resource allocation method selection unit, the user's intention is confirmed, or in response to the resource allocation request, the task priority of the resource allocation selection list storage unit, the resource It further comprises resource allocation intention confirmation means for extracting resource allocation information to the task according to the resource usage status of the usage status information storage means,
The resource allocation means includes:
The resource control according to claim 1, further comprising means for allocating a calculation resource so as to correspond to the user's intention when the selected resource allocation method does not correspond to the user's intention. apparatus. The resource allocation request is
Contains the importance of the task set by the user,
The resource allocation means includes:
The resource control device according to claim 1, further comprising means for allocating a resource during task execution to another resource or determining whether to continue using the resource during task execution according to the importance of the task. . A resource control method for allocating a plurality of different types of computing resources for processing tasks,
A resource allocation method selection list storage means storing the resource allocation method;
Resource usage status information storage means for storing usage status information of computing resources;
A resource allocation selection list storage means storing a resource allocation selection list;
In an apparatus having resource allocation method selection means and resource allocation means,
When the resource allocation method selection unit receives a resource allocation request for a task including the user's intention, the resource allocation method selection unit selects a resource allocation method from the resource allocation method selection list storage unit according to the usage mode of the device that executes the task. A resource allocation method selection step,
The resource allocation means refers to the resource usage status information storage means, and allocates the calculation resource when the calculation resource designated by the resource allocation request exceeds a threshold according to priority, A resource allocation step of allocating a calculation resource having another free space when the calculation resource free space falls below a threshold according to the priority;
The resource control method characterized by performing. A resource allocation intention confirmation step for confirming whether the resource allocation method selected in the resource allocation method selection step is defined in the resource allocation selection list storage means;
In the resource allocation step,
Based on the resource allocation method selected in the resource allocation intention confirmation step, the user's intention is confirmed, or in response to the resource allocation request, the task priority of the resource allocation selection list storage means, the resource Extract resource allocation information to the task according to the resource usage status of the usage status information storage means,
In the resource allocation step,
The resource control method according to claim 4, wherein when the resource allocation method selected in the resource allocation intention confirmation step does not correspond to the user's intention, the calculation resource is allocated to correspond to the user's intention. The resource allocation request is
Contains the importance of the task set by the user,
In the resource allocation step,
6. The resource control method according to claim 4 or 5, wherein, depending on the importance of the task, it is determined whether to allocate a resource during task execution to another resource or to continue using the resource during task execution. Computer
The resource control program for functioning as each means of the resource control apparatus of any one of Claims 1 thru | or 3.

Images