JPH09305417A - Cpu use rate optimizing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To operate a CPU with high efficiency by comparing a CPU use rate with a current CPU use rate and changing the priority of jobs composed of work corresponding to the compared result. SOLUTION: The CPU use time per unit time of work acquired by a CPU use time acquiring means 11 of a control part 1 is read out and the current CPU use rate of work is calculated by a current CPU use rate calculating means 12. Besides, the CPU use rate of work acquired from the output of a CPU use rate optimizing means 21 is set by a CPU use rate acquiring means 22 of a statistic part 2. Then, that CPU use rate is compared with the current CPU use rate by a CPU use rate control means 13 and corresponding to the compared result, the priority of jobs consisting of the work is changed. Thus, the CPU can be operated with high efficiency.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a CPU usage rate optimizing method, and more particularly, to a C method for a task executed on an information processing apparatus.
The present invention relates to a PU usage rate optimization method.

[0002]

2. Description of the Related Art As a conventional technique for calculating the CPU usage rate, for example, in Japanese Patent Laid-Open No. 63-68944, a value previously measured as the CPU usage rate after occurrence of an event during business processing is registered. A method has been proposed in which the allocation of various jobs is dynamically determined by giving the value as a CPU usage rate to achieve high-speed and efficient CPU processing. That is, in this conventional technique, the CPU usage rate and the current CPU usage rate are registered and held for each business, and the current CPU usage rate is rewritten when business processing occurs, and the C
The CPU usage rate of the system is generated from the total of the PU usage rates, and the work to be processed next is allocated according to the value.

As a technique related to the CPU usage rate, for example, in Japanese Patent Laid-Open No. 4-245544, data in an asynchronous input / output buffer provided in a main storage device is transferred to an external storage device by an asynchronous input / output execution task. In a data processing system for outputting, a method has been proposed in which the priority of an asynchronous execution input / output task can be periodically changed based on the CPU usage rate. This is because the amount of data written to the asynchronous I / O buffer per unit time generally increases when the CPU usage rate is high, and conversely
When the usage rate is low, it decreases. Therefore, the priority of the asynchronous I / O execution task is changed periodically based on the CPU usage rate, and when the asynchronous I / O buffer stores an almost appropriate amount of data, it is sent to the external storage device. The processing speed is improved by outputting the data.

[0004]

However, the CPU utilization rate calculation method proposed in Japanese Patent Laid-Open No. 63-68944 has the following problems.

The first problem is that it is difficult for a user who executes a job to obtain an optimum value when setting the value of the CPU usage rate allocated to the job. This is because the user sets an empirical value as the CPU usage rate of the work, so it is difficult to predict what value the CPU usage rate should be set to achieve the most efficient operation. It depends. Therefore, there is no guarantee that the CPU usage rate set by the user is an optimum value for system operation.

The second problem is that the CPU usage rate for each job is fixedly provided as one of the system tables, so that the CPU usage rate cannot be dynamically given. Therefore, a detailed CP for each unit time required for optimal use of the CPU of the entire system
U usage rate cannot be set.

Therefore, the present invention has been made in view of the above circumstances, and an object thereof is to use a CPU used for a job from statistical information obtained by previously measuring the CPU usage time of each job and the number of jobs. The rate is calculated and assigned automatically, and C
It is to provide a CPU usage rate optimizing method capable of operating a PU with high efficiency.

[0008]

In order to achieve the above-mentioned object, the CPU usage rate optimizing method according to the present invention generates statistical information on the CPU usage rate of a job executed by an information processing apparatus and stores it in a storage unit. And means for storing the statistical information stored in the storage unit, and optimizing the CPU usage rate based on the analysis result, and for each business time period obtained by the optimizing means. CPU usage rate is set, the set CPU usage rate is compared with the current CPU usage rate calculated by acquiring information on the current CPU usage time of the business, and the business is performed according to the comparison result. And means for controlling the CPU usage rate of the job by changing the priority of the jobs constituting the.

In the present invention, the statistical information includes the total number of job jobs, CPU usage time information for each job,
And, the optimizing means calculates the optimum CPU usage rate by giving a predetermined weight to the CPU usage rate from the statistical content, including statistical information including information on whether or not the work is completed within the operating time. Is characterized by.

[0010]

BEST MODE FOR CARRYING OUT THE INVENTION Preferred embodiments of the present invention will be described below. The present invention, in a preferred embodiment, referring to FIG. 1, CPU utilization statistics means (23)
(25), and outputs it to the file (4). Then, the statistical information is read from this file (4), and the statistical information analysis means (24) analyzes the CPU usage status for each business from the statistical information and calculates the optimum CPU usage rate in terms of resource utilization of the system ( 21).

In the embodiment of the present invention, the statistical information includes the number of job jobs, job execution order, and C of each job.
The PU usage time is included, and the statistical information analysis means (24)
From the statistical information, the average CPU usage rate for one day of work is calculated.

The CPU usage rate optimizing means (21) uses the value of the average daily CPU usage rate of the job to assign the CPU usage rate to each job in the job according to the CPU usage time of the statistical information. Calculated automatically. And business CPU
The CPU usage rate acquisition means (22) acquires the value of the usage rate set for each time period.

The CPU usage rate control means (13) uses the current CP.
The CPU usage rate is controlled by changing the priority change of the execution job according to the U usage rate and the CPU usage rate automatically set for each job by the CPU usage rate obtaining means (22).

An embodiment of the present invention will be described below in order to describe the above-described embodiment of the present invention in more detail.

[0015]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. Referring to FIG. 1, a CPU usage optimization method according to an exemplary embodiment of the present invention includes a control unit 1.
And a statistical unit 2, an output device 3, and a file device 4.

The control unit 1 has a CPU usage time acquisition means 11
And a current CPU usage rate calculation means 12 and a CPU usage rate control means 13.

The statistic section 2 includes a CPU usage rate optimizing means 21.
The CPU usage rate acquisition means 22, the CPU usage rate statistical means 23, the statistical information analysis means 24, the statistical information generation means 25, the display means 26, the reading means 27, and the output means 28 are included.

In this embodiment, the system operating time is 9:
From 00 to 17:00, three operations are in operation,
Statistics shall be taken on an hourly basis. Further, in order to simplify the explanation, it is assumed that there is nothing that operates other than the three jobs.

Here, the CPU usage time represents the total time of the user time and the kernel space usage time, and CP
The U usage rate refers to the ratio of the CPU usage time of work to the CPU usage time used in the entire system. That is, the CPU usage time of the system within the target time is S
(S), if the CPU usage time of task 1 is S (1),
The CPU usage rate of job 1 is (S (1) / S (s)) × 1
00.

FIG. 3 shows an example of the statistical information of work 1 to work 3 in the time zone from 9:00 to 10:00.

The CPU usage time of the entire system in the time zone from 9:00 to 10:00 is a value obtained by adding the total CPU usage time of each of work 1, work 2, and work 3, that is, 25+
5 + 20 = 50 (minutes).

The current CPU usage rate of job 1 during this time period is
It is calculated as (25/50) × 100 = 50 (%). Similarly, the current CPU usage rates of business 2 and business 3 are 10 respectively.
(%) And 40 (%).

The operation of this embodiment will be described below.

First, referring to FIG. 1, the CPU of the control unit 1
By the usage time acquisition means 11, the CPU usage time per unit time of business and C per unit time of the entire system
The PU usage time is obtained. This CPU usage time is read out, and the current CPU usage rate calculation means 12 determines the current CPU of the job.
Calculate the usage rate. This value is the current CPU usage rate (A).

The CPU usage rate acquisition means 22 of the statistics section 2 is C
C of the business obtained from the output of the PU usage rate optimizing means 21.
Let the PU usage rate be the CPU usage rate (B).

The current CPU usage rate (A) and the CPU usage rate (B) are compared, and C of the control unit 1 is set so that the CPU usage rate (A) of the work is brought close to the target CPU usage rate (B).
The PU usage rate control means 13 controls. At that time, CPU
The usage rate control means 13 controls using a priority change method.

That is, the current CPU usage rate (A) <CPU
If it is the usage rate (B), the execution priority of the job in the job is increased so that the CPU can be used more. On the contrary, when the current CPU usage rate (A)> CPU usage rate (B), the execution priority of the job in the job is lowered to suppress the CPU usage.

When the current CPU usage rate (A) = CPU usage rate (B), priority control is not performed.

Then, control of the CPU usage rate by the above-mentioned priority change system is performed at regular time intervals.

Next, the CPU usage rate statistics means 23 acquires the CPU usage time per unit time of work, the current CPU usage rate, and the number of jobs.

These information and the CPU usage rate acquisition means 2
CPU usage rate from 2 and current CPU usage rate calculation means 12
Current value obtained by (current value, current CPU usage rate)
In addition, the statistical information generating means 25 outputs to the file 4 as statistical information for each work. The statistical information is preferably obtained by averaging the information up to the previous day and updating it every day to obtain more accurate information.

The file 4 is used as the information of the work subject to the CPU usage rate control, the number of jobs per time zone and its unit time, the job execution order, the CPU usage time, the current CPU.
It is a file that stores the usage rate and the CPU usage rate, and an example of the stored contents is as shown in FIG.

Next, the statistical information analysis means 24 will be described.

In this embodiment, the system operation time is
Since it is from 9:00 to 17:00, the maximum is 8 for the entire work.
× 60 = 480 minutes real time can be used.

The statistical information analysis means 24 reads the information of the file 4 via the reading means 27.

FIG. 2 schematically shows an example of the statistical information for each job of the business stored in the file 4. That is, the statistical information read from the file 4 includes the number of jobs per day for each job, the job execution order, and information on whether the job is completed within the operating time.

Referring to FIG. 2, the total number of jobs to be executed in one day of job 1 is 5 (job 1 to job 5), and the total CPU usage time of all jobs is 174 minutes. Further, the total number of jobs executed in the day of work 2 is 3, and the total CPU usage time of all jobs is 160 minutes. The total number of jobs executed in one day of work 3 is 3, and the total CPU usage time of all jobs is 1
10 minutes.

Therefore, in the whole work, 174 + 160 +
It can be seen that 110 = 444 minutes of CPU time is consumed. Generally, the real time is longer than the CPU usage time, but in the example shown in FIG. 2, 480> 444, which satisfies this condition.

The CPU usage rate for the first day of work 1 is (174
/444)×100=39.1(%) is calculated. This value is set as C (1). Calculating assuming that real time is also used at this ratio, 480 × 0.391 = 187.6
8 (minutes).

Next, the CPU usage rate optimizing means 21 will be described. FIG. 4 is a flowchart showing the processing flow of the CPU usage rate optimizing means 21 in this embodiment. CPU
The usage rate optimizing means 21 obtains the optimum CPU usage rate for each job as follows.

Referring to FIG. 4, the statistical information on the work is read from the statistical information file 4 (step 101),
It is determined whether or not the work is completed within the operating time zone (step 102), and if the work is completed, the process proceeds to step 103.

On the other hand, if the job is not completed within the operating time, a weight p is given to the value C (1) calculated as the CPU usage rate of the job, and the new CPU usage rate is C (1) × p = p.
Let it be C (1) (step 104).

Next, since the number of jobs is proportional to the CPU usage time, the number of jobs for each job is given as a weight to the CPU usage rate by the weighting coefficient k (step 103), and a new C
The PU usage rate is pC (1) × k = kpC (1).

In the present embodiment, k for the work 1
It is assumed that p = 1 and the CPU usage rate is C (1).

For the weighting factors p and k, the most effective day is searched by referring to the statistical information, and the value used on that day is set. Here, what can be said to be "effective" is
This is the case where the difference between the current CPU usage rate and the CPU usage rate is the smallest.

Further, since the total of the CPU usage rates of all the jobs may exceed 100% due to the addition of the weight, it is determined whether the total of the CPU usage rates of all the jobs exceeds 100%. It is judged (step 105), and the total is 100.
The correction is made so that it becomes not more than% (step 106).

That is, the CP of each job obtained by adding weights
The ratio of the U usage rates is calculated, and the ratio is kept and recalculated so that the total becomes 100.

Here, it is assumed that the coefficients k and p are 1, and the CPU usage rate assigned to the task 1 is C (1) = 3.
It becomes 9.1 (%).

This value is notified to the CPU usage rate acquisition means as CPU usage rate information (step 107).

Next, CPU for each unit of time for business
Similarly, the usage rate allocation method will be described by using the task 1 as an example.

The daily CPU usage rate of task 1 is 39.1.
(%) Was calculated, but if it is calculated assuming that real time is used at this ratio, then 480 × 0.391 = about 188
(Minutes). When this time (188 minutes) is assigned to the jobs 1 to 5 in the job 1, it becomes as shown in FIG. For example, jobs 1 and 2 (CPU usage time is 30 minutes) are allocated 32 minutes as real time, and jobs 4 and 5 are allocated.
49 minutes is allocated as the actual time to (CPU usage time is 45 minutes).

The shaded area in FIG. 6, ie, the area of the shaded area = real time of the entire system × CPU usage rate is represented.

The area of the shaded area is the real time consumed by the job 1. The area of the shaded area = Σ can be represented by the real time of the job in the job.

Then, by dividing the CPU usage time allocated to the job 1 into each job as shown in FIG. 6, the CPU usage rate for each time zone can be determined.

From the statistical information shown in FIG. 3, the execution order condition of the job of job 1 is obtained.

As schematically shown in FIG. 2, job 1 and job 2 are executed in parallel, and job 3 is conditioned on the completion of the jobs.
Is executed. Then, when job 3 ends, job 4
And job 5 are executed in parallel. Based on this condition, the CPU usage rate for each time zone is obtained.

As shown in FIG. 6, the parallel execution real time of jobs 1 and 2 is A time, the execution real time of job 3 is B time, and the parallel execution real time of jobs 4 and 5 is C time. Also, the CPU usage rate assigned to job 1 and job 2 is x (%), and the CPU usage rate assigned to job 3 is y.
(%), And the CPU usage rate assigned to jobs 4 and 5 is z (%). Then the following equation holds.

A (min) × x / 100 = 32 (min) B (min) × y / 100 = 26 (min) C (min) × z / 100 = 49 (min) A + B + C = 480 ... Equation (1)

Two jobs are executed in the time zone A, one job is executed in B, and two jobs are executed in C. Therefore, weighting is performed in proportion to the number of jobs, for example, A: B: C = 2. :
When calculated as 1: 2, A = 192, B = 96, C = 192 ... Formula (2)

Solving the above equations (1) and (2), x = 1
6.6 (%), y = 27.1 (%), z = 25.5
(%).

As a result of the above calculation, the CPU usage rate assigned to the job 1 is 1 for 9:00 to 12:12.
6.6 × 2 = 33.2 (%), 12:13 to 13:49
Is 27.1 (%), 13:50 to 17:00 is 2
It is calculated as 5.5 × 2 = 51 (%).

The CPU usage rate obtained as described above is
The CPU usage rate acquisition means 22 acquires the information as information for each work, and the CPU usage rate control means 13 acquires the information and the current CP.
Control is performed by comparing with the current value obtained by the U usage rate calculating means 12.

Further, the information of the file 4 is read out by the means 2
7, and the output means 3 outputs it as statistical information of the CPU usage rate by using the display means 26.

FIG. 7 is a diagram showing an example of output information.
The CPU usage rate and the current CPU usage rate for each business time zone, and the CPU usage time of each job in the business are displayed.

[0065]

As described above, the CP according to the present invention
The U usage rate optimizing unit first obtains the CPU usage rate by calculating the average CPU usage rate for one day and then calculating the CPU usage rate for each business time period. That is, the CPU usage rate, which was conventionally given manually, can be automatically obtained as the number of jobs based on the statistical information in proportion to the actual execution time. It is possible to obtain a CPU usage rate more suitable for efficient use of the CPU than the experience value given by the user.

Further, according to the present invention, since the statistical information is updated by taking the average of daily information, it is possible to improve the performance of the CPU usage rate optimization by using the statistical information.
In addition, by storing the weights p and q of the most efficiently used day and using the values from the next time, it is possible to obtain a more optimal CPU usage rate.

According to the present invention, the CP output is optimally determined by referring to the result output by the display means.
This has the effect of verifying whether the U usage rate is set.

[Brief description of drawings]

FIG. 1 is a functional configuration diagram showing an embodiment of a CPU usage rate optimizing system according to the present invention.

FIG. 2 is a diagram for explaining an embodiment of the present invention and is a diagram showing an example of statistical information regarding a CPU usage state stored in a file 4;

FIG. 3 is a diagram for explaining the embodiment of the present invention and is a diagram showing an example of statistical information of each work in a certain time zone.

FIG. 4 is a flowchart showing a processing flow of a CPU usage rate optimizing unit in the embodiment of the present invention.

FIG. 5 is a diagram for explaining the embodiment of the present invention, and is a diagram showing allocation of CPU usage time of each job in the job 1 obtained from statistical information and real time calculated from the statistical value thereof.

FIG. 6 is a diagram for explaining the embodiment of the present invention and is a diagram showing an example of allocation of the CPU usage rate for each time slot of the job 1;

FIG. 7 is a diagram for explaining the embodiment of the present invention and is a diagram showing an example of outputting statistical information to an output device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 control part 2 statistics part 3 output device 4 file 11 CPU usage time acquisition means 12 current CPU usage rate calculation means 13 CPU usage rate control means 21 CPU usage rate optimization means 22 CPU usage rate acquisition means 23 CPU usage rate statistical means 24 Statistical information analysis means 25 Statistical information generation means 26 Display means 27 Reading means 28 Output means

Claims (5)

[Claims] 1. A means for generating and storing in a storage unit statistical information relating to a CPU usage rate of a job executed by an information processing apparatus, and analyzing the statistical information stored in the storage unit, and based on the analysis result, the CPU An optimizing means for optimizing the usage rate and a CPU usage rate for each business time period obtained by the optimizing means are set, and the set CPU usage rate and the current CPU usage time of the business are set. Current CP calculated by acquiring information
The CPU of the job is compared by comparing the U usage rate and the priority of the job constituting the job according to the comparison result.
A CPU usage rate optimization method comprising: a means for controlling a usage rate. 2. The statistical information is the total number of jobs for business,
Information about the CPU usage time of each job and information about whether or not the work was completed within the operating time are included as statistical contents, and the optimizing means gives a predetermined weight to the CPU usage rate from the statistical contents. The CPU usage rate optimizing method according to claim 1, wherein an optimum CPU usage rate is calculated. 3. The CPU according to claim 1, further comprising means for displaying statistical information stored in the storage unit.
Utilization optimization method. 4. Each job in a job is included in the statistical information based on the statistical information including the total number of jobs of the job executed by the information processing device, the job execution order, and the CPU usage time of each job. Means for automatically calculating the optimum CPU usage rate according to the CPU usage time, and a value for each time zone for the calculated CPU usage rate of the task, and the value is calculated from the current CPU usage time information of the task. Current CP
By comparing the U usage rate with the CPU usage rate set for each job and changing the priority of the execution job according to the comparison result, the CPU usage rate is adjusted so that the job can be operated most efficiently. CPU utilization rate optimizing method characterized by performing the above control. 5. The CPU utilization rate optimizing method according to claim 4, wherein the statistical information includes a value obtained by averaging information for each business, and the value is updated for each operation unit. .