JP2890569B2 - Process control equipment - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a process control system, and more particularly to a process control system capable of efficiently allocating resources of processes executed in parallel.

(Prior Art) In a normal process, round-robin scheduling is performed.

FIG. 5 is a flowchart showing an example of the round robin scheduling.

When executing a plurality of processes, focusing on one of the processes, first, a maximum CPU occupation time (= Tmax) is set (step S1). Next, the continuous CPU occupation time (= Tseq) is reset (step S2), and measurement of the CPU occupation time is started (step S3). Next, Tmax ≦
It is determined whether the condition of Tseq is satisfied (step
S4) While the determination is negative, the process proceeds to step S5 to execute the target program (step S5). Next, it is determined whether the process has been completed (step S6),
When a negative determination is made, the process returns to step S3 and the processes in steps S3 to S6 are continued.

If the determination in step S4 is affirmative, the process proceeds to step S7, where the process is switched. That is, the processing of the CPU is transferred to another process.

Thereafter, the process waits until the process of the CPU is transferred from another process to the own process (step S8). When the process is transferred, the process returns to step S2 and the continuation of the above-described process is executed again.

For example, as shown in FIG. 6, a plurality of workstations 2,
It is assumed that there is a LAN to which the print servers 3 and 4 are connected. Then, there is a request to the print servo 5 to first finish printing 100 pages from the workstation 2 by 5 hours, and then to finish printing 20 pages from the workstation 3 by 1 hour. Then, it is assumed that there is a request from the workstation 4 to finish printing one page by two hours. It is also assumed that it takes two minutes for the print server 5 to print one page.

In such a case, according to the round-robin scheduling, the unit time of process execution is, for example,
If it is determined to be 20 minutes, first perform 10 pages of printing requested by the workstation 2, then perform 10 pages of printing requested by the workstation 3, and then perform the printing of the page requested by the workstation 4. One page is printed.

Next, 10 pages of printing requested by the workstation 2 are performed again, and then 10 pages of printing requested by the workstation 3 are performed. After that, the printing requested by the workstation 4 has been completed, so the printing requested by the workstation 2 is performed again for 10 pages.

 By repeating the above, printing is continued for every 10 pages.

The A process has the first priority, the B process has the second priority, and the C process has the third priority. For the first priority, the process execution unit time is, for example, 20 units. For example, for 10 minutes,
For example, if the priority is determined to be 5 minutes for the third priority, process A (20 minutes) → process B (10 minutes)
Process control was performed in the order of process C (5 minutes) → process A (20 minutes) → process B (10 minutes) → process C (5 minutes) →.

(Problems to be Solved by the Invention) As described above, conventionally, the resource allocation of processes executed in parallel has been fixed, so that there is a problem that a process processing request cannot be sufficiently answered. Was.

For example, in the above example, the request from workstation 2 to finish printing 100 pages by 5 hours will be completed after about 4 hours and 2 minutes, so the workstation 2 will not need to be able to fulfill it with sufficient time. The finish of the request from the workstation 3 is 1 hour and 22 minutes later, and there is a problem that the request from the workstation 3 to finish printing 20 pages by 1 hour later cannot be realized.

An object of the present invention is to eliminate the problems of the conventional method,
It is an object of the present invention to provide a process control method capable of maximally satisfying the requirements of each process by allowing the resource allocation of processes executed in parallel to be dynamically switched.

(Means and Actions for Solving the Problems) In order to achieve the above-mentioned object, the present invention provides a method in which a plurality of processes executed in parallel determine in parallel whether or not to end their processing by an end target time. Judgment is made in consideration of the number of processes that are being executed, and when it is judged that the processing is completed by the end target time, the processing time of the own process is reduced or the priority is lowered, and conversely, the processing is The feature is that when it is determined that the process is not completed by the end target time, the processing time of the own process is extended or the priority is raised.

According to the present invention, each process dynamically determines whether or not its own process is completed by the target end time. When it is determined that the processing is to be terminated, the processing of the CPU is passed to another process, and control is performed so that the processing proceeds quickly. On the other hand, when it is determined that the own processing is not completed by the end target time, the control is performed so that the occupation time of the CPU is lengthened and the processing of the own process is promoted.

As a result, the target end time of each process is satisfied by the greatest common denominator.

(Example) Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 shows a flowchart of one embodiment of the present invention.
First, an initial value for controlling the process is set (step S1). Specifically, the target end time (= Ten
d), the maximum CPU occupation time (= Tmax), and the estimated total CPU occupation time (= Tall) are set.

For example, in the case of the process 1 in FIG. 7, as shown in FIG. 2, the end target time (= Tend) is 5 hours,
The maximum CPU occupation time (= Tmax) is 20 minutes, and the estimated total CPU occupation time (= Tall) is 200 minutes.

Next, sampling information is set (step
S2). Specifically, the current time (= Tnow), the consumed CPU occupation time (= Tpass), the estimated remaining CPU occupation time (= Tstill),
Current number of processes (= Pnow), continuous CPU occupation time (= T
seq) is set. Note that Tstill = Tall−Tpass.

In the example of FIG. 7, the number of processes (Pnow) at the present time becomes 3 because there are three processes, Process 1 to Process 3, at the beginning.
The current number of processes (Pnow) is 2.

Next, the continuous CPU occupation time is updated (step S3), and it is determined whether or not Tmax ≦ Tseq is satisfied (step S4). If not, the target program of this process is executed (step S4). S5). Subsequently, it is determined whether the process has been completed (step S6),
If not, the process returns to step S3 and returns to steps S4 and S5.
Processing is continued.

If step S4 is positive, the process proceeds to step S7,
The sampling information set in step S2 is updated. Thereafter, it is determined whether (Tend-Tnow) / Pnow is greater than Tstill (step S8).

Here, (Tend-Tnow) represents the time from the current time (Tnow) to the end target time (Tend), and Pnow represents the number of processes executed in parallel at the current time. Therefore, (Tend-Tnow) / Pnow indicates the resource allocation time of the process of interest. In step S8, this resource allocation time is calculated as the estimated remaining CPU occupation time (Tstil
l), when the former is larger than the latter (Step S8 is affirmative), since the processing time has a margin, the maximum CPU occupation time (Tmax) of the own process is reduced (Step S10), An operation to pass the processing of the CPU to another process quickly is performed. On the other hand, when the former is smaller than the latter (step S8
Is negative), there is no room for the processing time, so the maximum CPU occupation time (Tmax) of the own process is increased (step S
9) An operation of giving priority to the process of the own process without passing the process of the CPU to another process quickly. Note that the right side of the inequality in step S8 is represented by (Tstill + α) (where α
May be a fixed time).

When the processing in step S9 or S10 ends, step S
Proceeding to 11, the process is switched, that is, the process of passing the CPU to another process is performed (step S11). afterwards,
Waits for the CPU to be passed to its own process (step S12),
When the CPU is passed again (Yes at step S12), the step
Returning to S1, the maximum CPU occupation time (Tmax) is determined in step S9.
Alternatively, the process of the process is executed by updating to the value obtained in S10.

As described above, according to the present embodiment, the resource allocation time of the own process is calculated from the time from the current time to the target end time and the number of processes currently being executed in parallel. When it is larger than the estimated remaining CPU occupation time,
If the result is smaller than the expected remaining CPU occupation time, the maximum CPU occupation time is increased and the processing of the own process is prioritized. Therefore, there is an effect that efficient resource allocation can be performed.

FIG. 3 shows a modification of the above embodiment. This variant is
The point is that steps S101 and S102 are inserted between step S9 or S10 in FIG. 1 and step S11.

In step S101, the maximum CPU occupation time (Tmax) and the estimated remaining C
Compare the PU occupation time (Tstill). If the former is larger than the latter, the process proceeds to step S102 and the continuous CPU occupation time (Tse
Reset q) and return to step S3. As a result, the maximum C
If the PU occupation time (Tmax) is longer than the expected remaining CPU occupation time (Tstill), in other words, the CPU is set to Tmax once more.
If the operation is performed only for a certain time and the processing of the own process is completely terminated, the processing of the own process is continued without passing the CPU to another process.

In the above embodiment, the maximum CPU occupation time (Tmax)
Is increased or decreased, but the present invention is not limited to this, and the amount of work to be processed, for example, the number of prints may be adjusted.

Next, a second embodiment of the present invention will be described with reference to FIG. This embodiment is an example of a case where priorities are assigned to a plurality of processes that operate in parallel. The difference from the embodiment of FIG. 1 is that after step S8, the priority is increased (step S8). (S91) or the process of lowering (step S92).

According to the present embodiment, if there is room in the process of the own process, the priority is lowered so that other processes are preferentially processed. Conversely, if there is no room, the priority is raised and the process is performed. Demand for promotion. As a result, the priorities of the plurality of processes are not fixed, and the processing of each process can be operated flexibly. Therefore, an efficient resource ratio becomes possible.

(Effect of the Invention) As is apparent from the above description, according to the present invention, a plurality of processes executing in parallel adjust resource allocation by time slice in consideration of their own processing conditions. Therefore, there is an effect that efficient resource allocation can be performed.

In addition, since the priority is not taken as an absolute number but as a relative distance to a target, there is an effect that the priority of parallel execution can be made flexible.

[Brief description of the drawings]

1 is a flowchart of one embodiment of the present invention, FIG. 2 is an explanatory diagram of the flowchart, FIG. 3 is a diagram showing a modification of this embodiment, and FIG. 4 is a flowchart of a second embodiment of the present invention. FIG. 5 is a flowchart of a conventional system, FIG. 6 is a system diagram showing a preferred use example of the present invention, and FIG. 7 is a diagram showing an example of processing conditions of processes executed in parallel. 1 ... coaxial cable, 2, 3, 4 ... workstation, 5 ... print server

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06F 9 / 46,3 / 12,15 / 16 G06F 15 / 00,1 / 00

Claims (1)

(57) [Claims] 1. A process control device for a system for executing a plurality of processes each having at least a target processing end time and an expected total processing time in a time-sharing manner. Means for obtaining an expected remaining processing time from the processing time already consumed; means for obtaining a remaining time until the processing end target time from the target processing end time and the current time; execution in a time-sharing manner with the remaining time Means for comparing the resource allocation time of the own process obtained based on the number of processes being performed with the expected remaining processing time; and, when the resource allocation time of the process is larger than the expected remaining processing time, the own process. Means for reducing the processing time of the process or lowering the priority to promote the processing of other processes; and conversely, the resource allocation time of the process When the is less than between, to increase the elongation to or priority processing time of its own process, the process control device being characterized in that and means so as to facilitate the process of its own process.