JP2007334782A - Computer apparatus and computer control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a computer apparatus in which effective use of a CPU and suppression of power consumption are made compatible. <P>SOLUTION: The computer apparatus 100 has: a plurality of CPUs 101; a memory 102; an OS 121 and one or more applications 112. Each application 112 has functions to execute unused CPU use processing by receiving unused CPU usable events and to stop execution of the unused CPU use processing by receiving unused CPU use cancellation events. A CPU monitoring means 122 of the OS 111 monitors a CPU load state, transmits the unused CPU usable events to the applications 112 when a CPU load falls to equal to or below a first threshold and transmits the unused CPU use cancellation events to the applications 112 when the CPU load rises to equal to or more than a second threshold which is larger than the first threshold. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a computer apparatus, and more particularly to a technique for effectively using a CPU in a computer apparatus having a plurality of CPUs.

  With the advancement of hardware technology, computer devices having a plurality of CPUs have been supplied at a relatively low cost. A computer device having a plurality of CPUs improves the processing speed, throughput, and reliability by operating a plurality of CPUs in parallel, and has a higher performance computer system than a computer device having only a single CPU. realizable. On the other hand, power consumption increases according to the number of operating CPUs. For this reason, it is often possible to dynamically add (activate) and delete (deactivate) CPUs depending on the situation.

  In a computer apparatus having a plurality of CPUs, an OS (operating system) for a multiprocessor is generally used. The OS manages the operation status of each CPU, and implements parallel processing by assigning each thread of an application to each CPU. Also, a scheduling algorithm called gang scheduling is known in which threads constituting the same application are simultaneously executed by different processors in order to reduce execution synchronization waiting between threads (see, for example, Patent Document 1).

  Further, although not limited to a multiprocessor, there is a method for controlling the allocation of resources (CPU, memory, etc.) to an application so as to match a service level target specified by a parameter (see, for example, Patent Document 2). . Also, when not working, reduce the frequency of the reference clock supplied to the CPU to operate the CPU in the power saving mode or stop the clock to the CPU to further reduce power consumption There is also a technique (see, for example, Patent Document 3).

JP 2004-280293 A Japanese Patent Laid-Open No. 2004-21982 JP 2006-23803 A

  When the user is not performing an operation, the CPU load on the computer device may be very low. In such a case, as a conventional technical trend, as seen in Patent Document 3, reduction of power consumption is given top priority, and the computer apparatus is set to a power saving mode. However, this cannot make effective use of the CPU. This is because, for example, depending on the application, when processing of the next operation of the user can be predicted, if the processing is performed in advance by background processing etc., the response can be greatly improved, and when the user actually performs the operation This is because a high load on the CPU can be avoided. However, in this case, if background processing is permitted unconditionally for the application, the CPU can be used effectively, but there is a problem that the power consumption becomes too large. As seen in Patent Document 2, if a method of controlling the allocation of resources to applications so as to meet the service level target specified by the parameters can be used, unlimited use can be suppressed, but the resources are limited. Therefore, there is a risk of adversely affecting the original processing of the application.

  Accordingly, an object of the present invention is to provide a computer device that achieves both effective use of a CPU and suppression of power consumption.

  In addition, in a computer apparatus having a plurality of CPUs, CPUs are dynamically added and deleted according to the situation as described above, but there is no mechanism for quickly notifying an application of the addition or deletion of CPUs. For this reason, when developing an application that dynamically changes the number of distributed processes (number of threads) in accordance with the dynamically changing number of CPUs, logic for periodically checking the current number of CPUs in each application. As a result, it is difficult to develop an application, and periodic checks are executed for each application, so that the efficiency of the entire system is reduced.

  Therefore, another object of the present invention is to enable an application to quickly recognize a change in the number of CPUs due to addition or deletion of CPUs.

  A first computer apparatus according to the present invention is a computer apparatus comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS. Storing in the memory an identifier of an application that executes idle CPU utilization processing by receiving an idle CPU available event and stops execution of idle CPU utilization processing by receiving an idle CPU utilization cancellation event The CPU monitors the CPU load status to the OS, and when the CPU load drops below the first threshold, an idle CPU available event is sent to any application whose identifier is stored in the storage unit When the CPU load rises above a second threshold value greater than the first threshold value, the idle C Characterized in that it comprises a CPU monitoring means for transmitting an idle CPU usage cancel events to the application that sent the U available events.

  According to a second computer apparatus of the present invention, in the first computer apparatus, the number of CPUs that execute the idle CPU utilization process is limited, thereby suppressing an increase in CPU load due to the idle CPU utilization process. To do.

  According to a third computer apparatus of the present invention, in the first or second computer apparatus, the CPU monitoring unit monitors increase / decrease in CPU and transmits a CPU number change event to the application when the CPU increases / decreases. Features.

  According to a fourth computer apparatus of the present invention, in the third computer apparatus, the application performs a CPU number change process for changing a distributed process number by receiving a CPU number change event.

  A fifth computer apparatus according to the present invention is a computer apparatus comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS. The OS includes CPU monitoring means for monitoring an increase / decrease in CPU and transmitting a CPU number change event to an application when the CPU increases / decreases.

  According to a sixth computer apparatus of the present invention, in the fifth computer apparatus, the application performs a CPU number change process for changing a distributed process number by receiving a CPU number change event.

  A first computer control method of the present invention is a computer apparatus comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS. In the control method, the OS monitors the CPU load status, and the OS transmits an idle CPU available event to the application when the CPU load drops below a first threshold. 2, a third step in which the application executes idle CPU utilization processing by receiving the idle CPU available event, and the OS exceeds a second threshold greater than the first threshold. A fourth step of transmitting an idle CPU use cancellation event to the application when the CPU load increases; The application is characterized by comprising a fifth step of stopping the execution of the idle CPU usage processing by receiving the idle CPU usage cancel events.

  A second computer control method of the present invention is a computer apparatus comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS. In the control method, the OS monitors the CPU increase / decrease, the OS transmits a CPU number change event to the application when the CPU increases / decreases, and the application includes: And a third step of changing the number of distributed processes by receiving the CPU number change event.

  The first program of the present invention is a plurality of CPUs, a memory, and one or more applications executed on the CPU, and executes an idle CPU utilization process by receiving an idle CPU available event; A computer having an application that stops execution of the idle CPU utilization process by receiving an idle CPU utilization cancellation event is monitored for the CPU load status, and when the CPU load drops below the first threshold, An idle CPU available event is sent to the application that sent the idle CPU available event when the CPU load rises above a second threshold that is greater than the first threshold. It is characterized by functioning as a CPU monitoring means.

  The second program of the present invention monitors a CPU increase / decrease for a computer having a plurality of CPUs, a memory, and one or more applications executed on the CPU. It is made to function as CPU monitoring means which transmits a number change event.

"Action"
In the present invention, when there is no user operation on the computer device and the CPU load drops below the first threshold, this is detected by the CPU monitoring means of the OS, and the idle CPU is used to encourage the use of the idle CPU. A possible event is issued to the application. After that, when the CPU load becomes equal to or higher than the second threshold, for example, when a user operation is resumed, an idle CPU usage cancel event is issued to the application in order to reduce power consumption and reduce the influence on the original processing. Then, the idle CPU utilization process is stopped.

  When a CPU is added or deleted, this is detected by the CPU monitoring means of the OS, and a CPU number change event is issued to each application. As a result, the application can appropriately execute a process of changing the number of distributed processes in accordance with the number of CPUs after the change.

  According to the present invention, when the CPU load of the computer device is lower than the first threshold value, an idle CPU available event is notified to the application to prompt the idle CPU utilization process, and when the second threshold value is exceeded, the application is idled. Since the CPU utilization cancellation event is notified and the idle CPU utilization processing is stopped, it is possible to achieve both effective utilization of the CPU and suppression of power consumption.

  Further, according to the present invention, the CPU monitoring means monitors the CPU increase / decrease, and transmits the CPU number change event to the application when the CPU increases / decreases, so that the application can quickly change the CPU number by adding or deleting CPUs. Can be recognized. As a result, it is possible to easily develop an application that dynamically changes the number of distributed processes (number of threads) according to the number of dynamically changing CPUs.

“First Embodiment”
Referring to FIG. 1, a computer apparatus 100 according to the first embodiment of the present invention includes a plurality of CPUs 101, a memory 102, and an input / output device 103.

  The memory 102 is a storage device such as a RAM that can be accessed from a plurality of CPUs 101, and includes an OS 111, one or more application programs (hereinafter simply referred to as applications) 112, an idle CPU usage control table 113, and first and first 2 set load values 114 and 115 are stored.

  The OS 111 is a multiprocessor operating system, and includes an OS kernel 121 that performs basic functions such as memory management, file management, and device management, and CPU monitoring means 122 that is newly added in the present invention. .

  The CPU monitoring unit 122 has a function of monitoring the load status of the CPU 101 and transmitting an idle CPU available event to the application 112 when the CPU load drops below the first set load value 114. Also, a function of transmitting an idle CPU use cancellation event to the application 112 that has transmitted an idle CPU available event when the CPU load increases to a second set load value 115 greater than the first set load value 114 or more. Have

  The first set load value 114 and the second set load value 115 are fixedly stored as one of the initial setting parameters of the computer apparatus 100, or parameters that can be changed by a user operation from the input / output device 103 Is stored in the memory 102. Of the two set load values, the first set load value 114 indicates a CPU load that is a threshold value for starting the idle CPU use process, and the second set load value 115 is a threshold value for stopping the idle CPU use process. CPU load. The value of the second set load value 115 is determined by the balance between performance and power consumption. For example, in the case of a computer device driven by a battery, the value is set to a low value in order to reduce power consumption. The first set load value 114 is set to a value smaller than the second set load value 115. When the difference between the first set load value 114 and the second set load value 115 is small, there is a high possibility that the activated idle CPU utilization process is immediately stopped. On the contrary, if the difference is large, the number of scenes where the idle CPU utilization process is activated decreases, so that the determination is made in consideration of the balance between the two.

  The application 112 is a program such as a document viewer, and executes predetermined processing by being executed by any of the CPUs 101. The application 112 includes an I / O event processing unit 131, an idle CPU utilization processing unit 132, and an event dispatch unit 133.

  The I / O event processing means 131 is a part that executes processing corresponding to an I / O event issued to the own application 112, and exists for each type of I / O event.

  The idle CPU usage processing means 132 is a means for executing a process to be executed when an idle CPU available event is issued to the own application 112 (hereinafter referred to as an idle CPU usage process).

  The event dispatch unit 133 is a unit that receives an event from the OS 111 and controls the processing units 131 and 132 according to the type of the received event.

  The idle CPU usage control table 113 is a storage unit that records the transmission status of an idle CPU available event for each application 112 having the idle CPU usage processing unit 132.

  Referring to FIG. 2, the idle CPU usage control table 113 is provided with a plurality of entries for recording sets of application identifiers and idle CPU available event transmission statuses. In the application identifier, a character string that uniquely identifies the application 112 is set. In the idle CPU available event transmission status, either “transmitted” or “not transmitted” is set. The entry sequence indicates the priority order for transmitting the idle CPU available event.

  The input / output device 103 includes a keyboard that receives data and instructions from the user, a display that presents various data to the user, and the like.

  Next, the operation of the computer apparatus 100 according to the present embodiment will be described.

  When the computer apparatus 100 is turned on and the system is started up, the OS kernel 121 and the CPU monitoring unit 122 of the OS 111 start operation.

  Whenever the application 112 is activated on the computer apparatus 100, the CPU monitoring unit 122 sets an entry in which the application identifier is set if the activated application 112 is an application having the idle CPU utilization processing unit 132. It is added to the end of the usage control table 113. At this time, the idle CPU available event transmission status of the entry is set to “untransmitted”.

  As a method of determining whether the CPU monitoring unit 122 is a type of application having the idle CPU utilization processing unit 132, the type of application 112 is notified to the CPU monitoring unit 122 of the OS 111 when starting up. There is a method in which the CPU monitoring unit 122 has a function and determines the type based on the presence or absence of notification. As another method, there is a method in which whether or not the type has the idle CPU utilization processing unit 132 is set as one of the application attributes, and the CPU monitoring unit 122 determines with reference to the application attributes.

  The activated application 112 is executed on any one or a plurality of CPUs 101. When any I / O event occurs from the input / output device 103 or the like to the application 112, the OS kernel 121 transmits the I / O event to the corresponding application 112, and the event dispatch unit 133 outputs the corresponding I / O event. Processing means 131 is executed. These operations are the same as in the prior art.

  Further, the CPU monitoring unit 122 always executes the process shown in FIG. 3 during the operation of the computer apparatus 100. First, the CPU load status is periodically checked (step S101). The CPU load status to be investigated here is the CPU usage rate of the entire system considering the loads of all the plurality of CPUs 101. For example, if there are four CPUs 101 in total, three of which have a CPU usage rate of 0% and the remaining one has a CPU usage rate of 100%, the CPU usage rate of the entire system is 25%.

  Next, it is determined whether or not the CPU load is lower than the first set load value (step S102). If the CPU load is lower than the first set load value, the process proceeds to step S103, and if not, the process proceeds to step S110.

  In step S103, the idle CPU usage control table 113 is referred to, and it is investigated whether or not there is an application 112 whose idle CPU available event transmission status is “sent” (step S104). If there is no application 112 that has transmitted an idle CPU available event (NO in step S104), the process proceeds to step S107. On the other hand, if there is already an application 112 that has transmitted an idle CPU available event (YES in step S104), an idle CPU use cancel event is sent to this application 112 (step S105), and an idle CPU use control table is sent. The idle CPU available event transmission status of the application 112 in 113 is changed to “unsent” and the entry is moved to the end in order to lower the order (step S106). Then, the process proceeds to step S107. The reason for sending the idle CPU use cancellation event in step S105 is that the CPU load is still lower than the first set load value even though the idle CPU available event has already been sent to a certain application. This is because the idle CPU usage process by the application assumes that the CPU is hardly used, so that the usage is canceled and used by another application.

  In step S107, the application 112 having the application identifier set in the top entry of the idle CPU usage control table 113 is selected (step S107). Next, an idle CPU available event is transmitted to the selected application 112 (step S108), and the idle CPU available event transmission status of the application 112 in the idle CPU usage control table 113 is changed to “sent”. (Step S109). Then, it waits for a predetermined time (step S115), and then returns to step S101 to check the CPU load status again.

  As shown in FIG. 4, the application 112 that is the destination of the idle CPU available event in step S108 receives the idle CPU available event transmitted (step S201), and the event dispatch means 133 uses the idle CPU. The processing means 132 is executed (step S202). As a result, the CPU load increases in accordance with the amount of idle CPU utilization processing.

  On the other hand, if it is determined in step S102 that the CPU load is not lower than the first set load value, is the CPU load higher than the second set load value set higher than the first set load value? It is investigated whether or not (step S110). If the CPU load is not higher than the second set load value, the timer waits for a fixed time (step S115), and then the process returns to step S101 to check the CPU load status again.

  When the CPU load is higher than the second set load value, the CPU monitoring unit 122 refers to the idle CPU usage control table 113 (step S111), and the idle CPU available event transmission status is “sent”. It is checked whether or not the application 112 exists (step S112). If it exists, an idle CPU utilization cancel event is transmitted to the application 112 (step S113), and the idle CPU available event transmission status of the application 112 in the idle CPU utilization control table 113 is “not transmitted”. And the entry is moved to the end in order to lower the order (step S114). Then, it waits for a predetermined time (step S115), and then returns to step S101 to check the CPU load status again.

  As shown in FIG. 5, the application 112 that is the destination of the idle CPU usage cancellation event in step S113 receives the idle CPU usage cancellation event transmitted (step S211), and the event dispatch unit 133 uses the idle CPU usage. The execution of the processing means 132 is stopped (step S212). As a result, the CPU load decreases.

  FIG. 6 shows an example of temporal changes in CPU load. FIG. 6A shows an example of a temporal change in CPU load when there is no application 112 of the type having the idle CPU usage processing means 132, that is, the same situation as that of a conventional computer device. Even when the user operation is not performed and the CPU load becomes smaller than the first set load value, the idle CPU utilization process is not executed. For this reason, the CPU load changes with a low level. Thereafter, the CPU load is increased because the user operation is resumed, for example, the processing corresponding to the I / O event is executed by the application.

  FIG. 6B shows a temporal change example of the CPU load when the type of application 112 having the idle CPU utilization processing unit 132 exists. When no user operation is performed and the CPU load becomes smaller than the first set load value, execution of the idle CPU utilization process by the application 112 of the type having the idle CPU utilization processing means 132 is started. For this reason, the CPU load is increased. In the example of FIG. 6B, the transition is stably performed with a CPU load smaller than the second set load value. Thereafter, as in the case of FIG. 6A, for example, when processing corresponding to an I / O event is executed by the application, the CPU load exceeds the second set load value by superimposing the corresponding load. Therefore, an idle CPU utilization cancel event is issued, and the idle CPU utilization process is stopped. In the case of FIG. 6B, the idle CPU utilization process is executed from time t1 to time t2.

  By the way, the CPU load during the idle CPU utilization process may be lower than the second set load value as shown in FIG. 6B, but if the second set load value is exceeded, FIG. As shown in (c), the idle CPU utilization process is stopped by an idle CPU utilization cancellation event when the second set load value is exceeded. As a result, when the CPU load decreases and becomes lower than the first set load value, an idle CPU available event is issued again, and the idle CPU use process is executed. The situation where the idle CPU utilization process is stopped to exceed the set load value is repeated. Even in such a situation, the idle CPU utilization process proceeds to some extent, but the efficiency is poor. Therefore, it is desirable to introduce a mechanism for suppressing an increase in CPU load due to idle CPU utilization processing.

  An example of a method for suppressing an increase in CPU load due to idle CPU usage processing is to notify the application of the number of CPUs available for idle CPU usage processing by a CPU available event, and the application is idle with the number of threads matching that number. CPU usage processing is executed. For example, when there are a total of n CPUs 101, assuming that the number of CPUs that can be used in the idle CPU utilization process is m (m <n), the fact that m CPUs can be used is indicated by the CPU monitoring means 122 by the idle CPU availability event. The application 112 is notified. The application 112 executes the idle CPU usage processing of the idle CPU usage processing unit 132 by dividing it into the same number of threads as the number of available CPUs m.

  Next, the effect of this embodiment will be described.

  According to the present embodiment, it is possible to achieve both effective use of the CPU and suppression of power consumption. The reason is that when the CPU load of the computer device is lower than the set load value, the idle CPU available event is notified to the application to prompt the idle CPU use processing, and when the set load value is exceeded, the idle CPU use cancel event is sent to the application. This is because the idle CPU utilization process is stopped by notifying the above.

  Next, a specific example of idle CPU usage processing will be described.

1) Specific example 1
The application 112 is a document viewer, and the idle CPU utilization processing unit 132 is a unit that performs a preceding process of rendering. Specifically, in order to request the display of the previous page or the next page with a high probability, the user performs this rendering in advance using the idle CPU utilization processing unit 132.

  FIG. 7A shows an execution state of the application 112 in a computer device having four CPUs. The second set load value is set to about 25%, and when the CPU of the computer device is not particularly used, an idle CPU available event that one corresponding to 25% of the four CPUs can be used. Is sent from the CPU monitoring means 122 to the application 112. The application 112 executes and outputs the predicted next operation result within a range where one CPU is used from the sent event, and uses it when the actual operation of the user matches the prediction. . For this reason, when the prediction is successful, the actual processing time required for the response can be shortened. In addition, the CPU usage rate at that time can be about 25% at the maximum. Therefore, it is possible to achieve both effective use of the CPU and suppression of power consumption.

  On the other hand, as shown in FIG. 7B, when all processes are performed after the user starts the operation, a certain amount of processing time is required even when the four CPUs are fully used. And the response is slow.

2) Specific example 2
As another example, in a computer device participating in a type of mechanism (for example, grid computing) that executes a process in response to a request from another, the CPU load of the computer device is changed according to the request from the other device. An example in which the application 112 executes the idle CPU utilization process when is lower than the set load value can be considered. This makes it possible to achieve both effective use of the CPU and suppression of power consumption.

“Second Embodiment”
Referring to FIG. 8, a computer apparatus 200 according to the second embodiment of the present invention is different from the computer apparatus 100 according to the first embodiment shown in FIG. In that the CPU monitoring means 222 is provided instead of the CPU monitoring means 122 and the CPU device state 116 is further provided.

  The CPU monitoring unit 222 has the same function as the CPU monitoring unit 122, and further has a function of monitoring an increase / decrease in CPU and transmitting a CPU number change event to the application 212 when the CPU increases / decreases. In the CPU number change event, the latest number of CPUs in the activated state is notified.

  The application 212 includes a CPU number change processing unit 134 in addition to an I / O event processing unit 131, an idle CPU utilization processing unit 132, and an event dispatch unit 133 similar to those of the application 112.

  The used CPU number change processing means 134 is a means for changing the number of distributed processes of processing of the application 212 (for example, processing executed by the I / O event processing means 131) in accordance with the number of CPUs notified by the CPU number change event. is there.

  The CPU device state 116 is a storage unit that records the state (activation, deactivation, etc.) of all the CPUs 101 existing in the computer device 200.

  Next, the operation of the computer apparatus 200 according to the present embodiment will be described. In the computer apparatus 200, the same operation as the computer apparatus 100 described in the first embodiment and an operation peculiar to the present embodiment are performed. The same operations as those of the computer apparatus 100 are not described because they are redundantly described. Hereinafter, operations unique to the present embodiment will be described.

  The CPU monitoring unit 222 always executes the process shown in FIG. 9 during the operation of the computer apparatus 200. First, the CPU device state is periodically checked (step S301). The CPU device state investigated here is a state in which each CPU 101 existing in the computer device 200 is activated or deactivated. Next, a past CPU device state is obtained from the CPU device state 116 (step S302), and a change in the number of active CPUs is detected by comparing with the investigated CPU device state (step S303). If there is no change, the process proceeds to step S306. When there is a change, a CPU number change event for notifying the number of active CPUs is transmitted to each application 212 (step S304). Next, the CPU device state 116 is updated with the investigated CPU device state (step S305), and the process proceeds to step S306.

  In step S306, a timer is waited for a certain time and an event notifying that the hardware has been changed. If a certain time has elapsed or a hardware change event has occurred before that, the process returns to step S301. The CPU device status is checked again.

  As shown in FIG. 10, each application 212 receives the CPU number change event transmitted from the CPU monitoring means 222 in step S304 (step S401), and the event dispatch means 133 executes the used CPU number change processing means 134. (Step S402). Thereby, the number of distributed processes is changed according to the number of CPUs after the change, and the CPU can be used efficiently.

  FIG. 11A shows the CPU usage status when two CPUs are newly added when the two CPUs are in the active state. In the present embodiment, when two CPUs are added, this is detected by the CPU monitoring unit 222 of the OS 111, and a system event (CPU number change event) for transmitting the information to each application 212 is sent. Upon receiving this event, the application 212 changes the number of distributed processing divisions according to the number of added CPUs based on the information, thereby quickly and effectively using the added CPUs.

  On the other hand, when no processing is performed for the addition of CPUs, as shown in FIG. 11B, the processing remains distributed with two CPUs, and the two added CPUs are replaced. It cannot be used. In addition, when each application periodically checks the number of CPUs and changes the number of distributed processes in accordance with that, the process of periodically checking the number of CPUs runs for each application as shown in FIG. Therefore, loss increases. Furthermore, in a method in which a large number of processes are distributed in advance and the number of CPUs is increased or decreased by the CPU scheduling function of the OS, the distribution number exceeds the number of CPUs that can actually be used, as shown in FIG. In such a case, a CPU context switch by time division for assigning a plurality of distributed processes to one CPU occurs, and a loss occurs at this time.

  Next, the effect of this embodiment will be described.

  According to this embodiment, the same effects as those of the first embodiment can be obtained, and an application can quickly recognize a change in the number of CPUs due to addition or deletion of CPUs. The reason is that the CPU monitoring means monitors the increase / decrease in the CPU and transmits a CPU number change event to the application when the CPU increases / decreases. As a result, it is possible to easily develop an application that dynamically changes the number of distributed processes (number of threads) according to the number of dynamically changing CPUs.

It is a block diagram of a 1st embodiment of the present invention. It is a figure which shows the example of the content of the idle CPU utilization control table in the 1st Embodiment of this invention. It is a flowchart which shows the process example of the CPU monitoring means in the 1st Embodiment of this invention. It is a flowchart which shows the process example concerning the idle CPU utilization event of the application in the 1st Embodiment of this invention. It is a flowchart which shows the process example concerning the idle CPU utilization cancellation event of the application in the 1st Embodiment of this invention. It is a figure which shows the example of a time change of CPU load in the 1st Embodiment of this invention. It is a figure which shows the specific example in the 1st Embodiment of this invention. It is a block diagram of the 2nd Embodiment of this invention. It is a flowchart which shows the process example of the CPU monitoring means in the 2nd Embodiment of this invention. It is a flowchart which shows the process example concerning the CPU number change event of the application in the 2nd Embodiment of this invention. It is explanatory drawing of the effect in the 2nd Embodiment of this invention.

Explanation of symbols

100, 200 ... computer apparatus 101 ... CPU
102 ... Memory 103 ... I / O device 111 ... OS
112, 212 ... Application 113 ... Idle CPU usage control table 114 ... First set load value 115 ... Second set load value 116 ... CPU device state 121 ... OS kernel 122, 222 ... CPU monitoring means 131 ... I / O event Processing means 132 ... Idle CPU utilization processing means 133 ... Event dispatch means 134 ... Used CPU number change processing means

Claims (10)

In a computer apparatus comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS,
A storage unit that stores, in the memory, an identifier of an application that executes an idle CPU use process by receiving an idle CPU available event and stops executing the idle CPU use process by receiving an idle CPU use cancel event With
Monitor the CPU load status to the OS, and send an idle CPU available event to any application whose identifier is stored in the storage unit when the CPU load drops below a first threshold, CPU monitoring means for transmitting an idle CPU use cancellation event to an application that has transmitted the idle CPU available event when the CPU load increases to a second threshold value greater than a first threshold value or more. Computer device.   The computer apparatus according to claim 1, wherein an increase in CPU load due to the idle CPU utilization process is suppressed by limiting the number of CPUs that execute the idle CPU utilization process.   The computer apparatus according to claim 1, wherein the CPU monitoring unit monitors increase / decrease in CPU and transmits a CPU number change event to an application when the CPU increases / decreases.   4. The computer apparatus according to claim 3, wherein the application performs a CPU number change process for changing a distributed process number by receiving a CPU number change event. In a computer apparatus comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS,
A computer apparatus comprising: a CPU monitoring unit that monitors an increase or decrease of a CPU and transmits a CPU number change event to an application when the CPU increases or decreases.   6. The computer apparatus according to claim 5, wherein the application performs a CPU usage change process for changing the number of distributed processes by receiving a CPU number change event. In a control method of a computer device comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS,
A first step in which the OS monitors a CPU load state;
A second step in which the OS sends an idle CPU available event to the application when the CPU load drops below a first threshold;
A third step in which the application performs idle CPU utilization processing by receiving the idle CPU available event;
A fourth step in which the OS transmits an idle CPU use cancellation event to the application when the CPU load increases to a second threshold value greater than the first threshold value;
A computer control method comprising: a fifth step in which the application stops execution of the idle CPU utilization process by receiving the idle CPU utilization cancellation event. In a control method of a computer device comprising a plurality of CPUs, a memory, an OS executed on the CPU, and one or more applications executed on the CPU under the control of the OS,
A first step in which the OS monitors an increase or decrease in CPU;
A second step in which the OS transmits a CPU number change event to the application when the CPU increases or decreases;
And a third step of changing the number of distributed processes by receiving the CPU number change event.   A plurality of CPUs, a memory, and one or more applications executed on the CPU, executing idle CPU usage processing by receiving an idle CPU available event and receiving an idle CPU usage cancel event The computer having an application that stops the execution of the idle CPU utilization process by monitoring the CPU load status, and when the CPU load drops below the first threshold, transmits an idle CPU available event to the application. , For functioning as a CPU monitoring means for transmitting an idle CPU use cancellation event to an application that has transmitted the idle CPU available event when the CPU load increases to a second threshold value greater than the first threshold value. program.   CPU monitoring means for monitoring a CPU increase / decrease in a computer having a plurality of CPUs, a memory, and one or more applications executed on the CPU, and transmitting a CPU number change event to the application when the CPU increases / decreases Program to function as.

Images