JPH05241860A - Time slice optimization system - Google Patents

Abstract

PURPOSE:To provide a time slice optimization system capable of estimating what characteristic of program is often executed at present, adjusting the time slice value of each program automatically and easily performing an efficient and optinum system operation. CONSTITUTION:This time slice optimization system is composed of a time slice information storage part 1 storing the time slice information on this program every program being executed now, a dispatch information storage part 2 storing the dispatch information for each program during a fixed period, a time slice optimizing part 4 estimating the characteristic of each program being executed now from the dispatch information stored in the dispatch information storage part 2 every fixed period and optimizing the time slice value of each program stored in the time slice information storage part 1 and a dispatch part 3 dispatching a CPU to a program by the time slice value optimized by the time slice optimizing part 4 and stored in the time slice information storage part 1.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a time slice optimizing method, and more particularly to a time slice optimizing method applied to a multi-programming system that dispatches CPUs using different time slice values for each program.

[0002]

2. Description of the Related Art A conventional time slice optimization method, for example, when applied to a multi-programming system, dispatches a CPU using a time slice in order to prevent a specific program from occupying the CPU for a long time. Was becoming. At this time, the time slice is set for each program or for each program group including a plurality of programs.

Generally, in the case of a program mainly composed of calculation processing, the dispatched CPU is often abandoned by using time slices. Conversely, in the case of a program mainly composed of interactive processing and input / output processing, it is dispatched. In many cases, the CPU is abandoned without using the time slice for the reason such as waiting for completion of input / output and waiting for other events.

Therefore, if a long time slice value is set as the time slice of each program, the number of dispatching processes in the entire system decreases when many programs whose main processing is calculation are executed, and therefore dispatching processes are performed. Although the overhead is reduced and the processing capability of the system is enhanced, when a program mainly for interactive processing is simultaneously executed, the responsiveness of the program becomes slow.
On the contrary, if a short time slice value is set, the responsiveness of the program mainly for interactive processing can be secured, but the overhead of dispatching processing increases because the number of dispatching processing in the entire system increases and the processing as a system The ability decreases and the strength and weakness are switched depending on the use.

[0005]

In the conventional time slice optimization method described above, the CPU pre-estimates the processing contents and sets the time slice value. However, the time slice set in each program is fixed, and even if there are many processes different from the presumed processing contents after the system is operated, it is not possible to automatically change the time slice value according to the current state of the system. There is a problem that you cannot do it.

An object of the present invention is to estimate what characteristics of programs are currently being executed, and to adjust the time slice value of each program automatically.
It is to provide a time slice optimization method that can easily and efficiently perform optimal system operation.

[0007]

According to the time slice optimization method of the present invention, in a multi-programming system for dispatching a CPU by using a different time slice value for each program, this program is executed for each program currently being executed. Time slice information storage unit that stores time slice information about the program, dispatch information storage unit that stores dispatch information for each program during a certain period, and current execution from the dispatch information stored in the dispatch information storage unit every certain period A time slice optimizing unit that estimates the characteristics of each program and optimizes the time slice value of each program stored in the time slice information storing unit; and the time slice information is optimized by the time slice optimizing unit. Store in storage And a dispatch unit for dispatching the CPU to program a time slice value,
The configuration is such that the time slice value is automatically adjusted according to the state of the multi-programming system.

[0008]

As described above, according to the present invention, according to the state of the system such as whether a large number of programs mainly for calculation processing are currently being executed or conversely, a large number of programs mainly for interactive processing are being executed. By automatically optimizing the time slice value of each program, optimum system operation can be easily performed.

[0009]

Embodiments of the present invention will now be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention.

The time slice optimization method of the present invention stores a time slice information storage unit 1 which stores time slice information relating to each program currently being executed, and dispatch information for each program during a fixed period. The time slice value of each program stored in the time slice information storage unit 1 is estimated by estimating the characteristics of each program currently being executed from the dispatch information storage unit 2 and the dispatch information stored in the dispatch information storage unit 2 at regular intervals. The time slice optimization unit 4 for optimization and the dispatch unit 3 for dispatching the CPU to the program with the time slice value optimized by the time slice optimization unit 4 and stored in the time slice information storage unit 1 are configured.

FIG. 2 is a diagram showing an example of the time slice information stored in the time slice information storage unit.

In order from the top, the time slice value 11 initially set in the program and the current time slice value 1
2, the maximum possible time slice value 1
3 and the minimum value 14 are shown. These pieces of information are newly prepared and store values when the corresponding program becomes executable. In this case, the current time slice value 12 stores the same value as the time slice value 11 initially set in the program.

FIG. 3 is a diagram showing an example of the dispatch information stored in the dispatch information storage unit.

From the top, the number of dispatches 21 and the number of times 22 when the time slice is used up are shown in order from the top.

FIG. 4 is a flow chart showing the algorithm of the dispatch unit.

First, when the operation is started, step (hereinafter S
1), each value in the dispatch information storage unit 2 is initialized to "0" at system startup. Next, the dispatch unit 3 adds "1" to the dispatch count 21 of the dispatch information storage unit 2 in S2. Next, in S3, the program to be executed is determined, and in S4, the time slice information storage unit 1 that stores the time slice information of the previously determined program is referred to. Dispatch the CPU. Then S6
And then wait for this program to give up the CPU,
When this program abandons the use of the CPU, the dispatch unit 3 checks the reason for the abandonment in S7, and if the use of the CPU is abandoned because the time slice is used up, S
In step 8, "1" is added to the number 22 of times the time slice has been used up in the dispatch information storage unit 2. Next, in S9, the number of dispatches 21 in the dispatch information storage unit 2 is checked, and if it is a certain value or more, the time slice optimization unit 4 is executed in S10, and then in S11, each value in the dispatch information storage unit 2 is set to Clear 0 ”. After that, the dispatch unit 3
Returns to the selection of the program to be executed next and repeats the operations from S2.

FIG. 5 is a flow chart showing the algorithm of the time slice optimizing unit 4.

The time slice optimizing unit 4 is executed each time the dispatch unit 3 dispatches a fixed number of times, and in step S21, the dispatch number 21 in the dispatch information storage unit 2 is compared with the time slice exhaustion number 22 to dispatch. If the number of times 22 that the time slice is used up is smaller than the number of times 21, it can be considered that there are many programs mainly for interactive processing or input / output processing for peripheral devices at present, so that the time slice value from the next time is reduced in S22. Set a negative value as an adjustment value. In the case of this embodiment, when the number of times 22 when the time slice has been used up is 20% or less of the number of times of dispatch 21, less than Δ is set as the adjustment value. On the other hand, in S23, the number of dispatches 21 in the dispatch information storage unit 2 is compared with the number of times 22 when the time slice is used up.
If there are many, it can be considered that there are many programs mainly composed of calculation processing at present. Therefore, in S24, a positive value is set as the adjustment value in order to increase the time slice value from the next time.
In the case of this example, when the number 22 of times the time slice is used up is 80% or more of the number 21 of dispatches, the adjustment value is +
Set Δ. If the number 22 of times the time slice has been used up is not too large or extremely small compared to the number of dispatches 21, the current time slice value can be considered to be valid, and there is no need to adjust the time slice value. . In the case of this example, when the number of times 22 when the time slice is used up is more than 20% and less than 80% of the dispatch number 21, it is considered that the optimization is not necessary, and the process of the time slice optimizing unit 4 is ended. Next, in S25, the time slice information storage unit 1 corresponding to each program being executed.
Is checked in order, and if all are not checked, the above adjustment value is added to the value of the current time slice value 12 in the time slice information storage unit 1 in S26. However, if it is determined that the value after addition in S27 is larger than the maximum value 13 in the time slice information storage unit 1, the current time slice value 12 is set to the maximum value 13 in S28. This is to prevent the time slice value from becoming infinitely large. When it is determined that the value after addition in S29 is smaller than the minimum value 14 in the time slice information storage unit 1, the current time slice value 12 is set to the minimum value 14 in S30. This is to prevent the time slice value from becoming 0 or less.

[0020]

As described above, the present invention compares the number of times a time slice has been used up during this period with the total number of times dispatched during this period to determine the current characteristics. It is possible to guess whether many programs are being executed and automatically adjust the time slice value of each program, and it is possible to easily and efficiently perform optimum system operation.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is a diagram showing an example of time slice information stored in a time slice information storage unit.

FIG. 3 is a diagram showing an example of dispatch information stored in a dispatch information storage unit.

FIG. 4 is a flowchart showing an algorithm of a dispatch unit.

5 is a flowchart showing an algorithm of the time slice optimizing unit 4. FIG.

[Explanation of symbols]

 1 time slice information storage unit 2 dispatch information storage unit 3 dispatch unit 4 time slice optimization unit

Claims (1)

[Claims] 1. In a multi-programming system for dispatching a CPU by using a different time slice value for each program, a time slice information storage unit for storing time slice information relating to this program for each program currently being executed, The dispatch information storage unit that stores dispatch information for each program during a fixed period, and the time slice information by estimating the characteristics of each program currently being executed from the dispatch information stored in the dispatch information storage unit at fixed intervals A time-slice optimizing unit that optimizes the time-slice value of each program stored in the storage unit, and the CPU in the program with the time-slice value optimized by the time-slice optimizing unit and stored in the time-slice information storage unit. Dispa Has a dispatch unit for
A time slice optimizing method characterized in that a time slice value is automatically adjusted according to the state of the multi-programming system.