JPS5913062B2 - Multiprocessing scheduling processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a multiprocessing scheduling method, and more particularly to a scheduling method for multiprocessing, such as formatting and writing, files stored in a plurality of large storage devices.

Although the subject matter of the present invention is not limited thereto, for example, when formatting and writing a large storage file multiplexed, two methods are conventionally known. That is, one is a method in which an operator performs scheduling (dividing jobs), and the other is a method by a program. However, the scheduling method by dividing jobs requires preparing a large number of job control cards, which requires a large amount of work for the operator, and increases the multiplicity of the system, which reduces efficiency. There are drawbacks. On the other hand, in the case of a scheduling method using a program, in the worst case, accesses may be concentrated on the same device (device). That is, the above scheduling is based on the operating system 05.
Since this is done by the program, the program cannot determine which file corresponds to which device, and cannot prevent the above situation in advance. In addition, in both of the above two methods, the amount of space for the file is determined in advance by the job control card, and there is no possibility that the amount of space for the file will be undesirably too small or undesirably large. occurs.

The present invention aims to solve the above-mentioned points, and includes the following features:
Dynamically determines the amount of space for files in large storage devices and requests to secure them at runtime. By managing two large storage files for each device, scheduling is automatically performed in consideration of device access efficiency. The purpose is to be able to determine the best value, such as by making the designation of the 3-multiplicity commensurate with the number of devices.

Therefore, the multi-processing scheduling processing method of the present invention stores information regarding individual modules to be processed in multiple ways in a scheduling processing method that multi-processes files stored in a plurality of large storage devices. A module table and a device table that stores information that links multiple files belonging to the device corresponding to each device registered for each module in the module table, etc., for each given file. creating a file table for storing unique information necessary for processing the file, and creating each of the tables by at least determining and securing space for the file before executing the processing on the large storage device; When executing processing, the program refers to the above module table, extracts one unprocessed device belonging to a free module, determines one file linked by the above device table, and selects one file linked by the above device table. Start processing based on the content, and start processing execution for the file in sequence until all of the plurality of modules are in the busy state or all of the plurality of devices are in the busy state. , when the above-mentioned busy state occurs, it waits for the completion of any of the processes based on the start of process execution for the file that was activated first, and starts them sequentially at the time of completion;
The feature is that when processing for one file belonging to one device is completed, processing for other files belonging to the device is sequentially executed based on the contents of the device table. This will be explained with reference to the drawings. Fig. 1 is a configuration of an embodiment for carrying out the method of the present invention, Fig. 2 is a flowchart of an embodiment for explaining the method of the present invention, and Fig. 3 is a diagram showing the operation of the present invention. It shows a time chart to explain.

In Fig. 1--, 1 represents a main storage device including a central processing unit, 2A and 2B are large storage devices, 3A1, 3A2, 3A3, . . . 3B1
, 3B2, 3B3... are files Alta.
29a39゜゛''.

゜〜Bl9b29b39゜O゜...are stored, 4 is the operating system, 5 is the program,
Reference numeral 6 represents a module table, 7 a device table, 8 a file table, and 9 an input section (or input information) such as file information. In the case of the present invention, (1) module table 6, (Ii) device table 7, and (i
[e-file table 8 is prepared.

The module table 6 is adapted to store information regarding individual modules to be processed multiplexed.

In the illustrated case, the first module L is shown to use device A, and the second module is shown to use device B. Also, corresponding to each module, control information such as whether the module is in use (that is, processing is being performed on the module) and whether device A is processing the module is also stored. be made to do.
Furthermore, if, for example, file A is currently being processed in response to device A, "a1" is written in the file area. The device table 7 stores information that links multiple files Al, a2, and a3 belonging to the device A, for example, corresponding to each device A or B registered for each module in the module table. be made into

In the illustrated case, files Al, a2, and a3 are shown linked to device A, and files B, b2, and b3 are shown linked to device B. Also, necessary control information is stored. The file table 8 contains specific information (file control program, amount of space, file control block, amount of space,
(number of created blocks, etc.) is stored.

The processing operation will be described below with reference to FIG. 2 and FIG. 3. In the case of the present invention, the amount of file space in the large storage device is determined and the reservation request is dynamically performed at the time of execution.

In other words, the operator does not have to decide in advance using a job card as in the past, but the operator can simply instruct, for example, how many blocks worth of data area should be secured, and the It should be done dynamically prior to that. The process 10 shown in FIG. 2 represents the above process, and the process first performs the following. That is, (1) it is specified in the module table 6 that device A is used for the first module, and device B is used for the second module.

Note that at this time, information instructing that file A be executed in correspondence with device A is not written at this time. (
2) In the device table 7, there is a file a corresponding to device A.
Link information is written that instructs that steps 1, A2, and A3 be executed in order.

(3) In the file table 8, each file A,
, a2, . . . , B, . . . are written.

When the process 10 shown in FIG. 2 is performed as described above, the program 5 refers to the above-mentioned tables and performs format writing on each file, for example.

That is, (4) the program 5 checks the control information of each module in the module table 6 and processes 11 shown in FIG.
, 12, it is checked whether there is a module that is not in use or whether there is an unprocessed device A.

As shown in the time chart shown in FIG. 3, at START, it can be seen that, for example, the first module is in a vacant state and that the device A is unprocessed. (6) Next, it is checked from the device table 7 that the file a1 is linked to the device A.

Then, the file A is written on the module table 6, and the process 13 shown in FIG.
of module table 6? Sir Til instructs during the information. (6) Then, through processes 14, 15, and 16, access to file A is performed by OS4 based on the contents of file table 8. This timing is represented by T1 in FIG. (7) Next, the program 5 immediately checks the contents of the module table 6 in the same manner as in steps (4) and (5) above, and learns that the second module is free and device B is pending. .

Then, as in the above 6), the timing T2 shown in FIG.
At this point, file B is accessed. (
g) Next, program 5 examines the contents of module table 6 in the same way as in request 4) above, but since both the first module and second module are in use, program 5 With a single event (S
WAIT).

This timing is designated as T3 in FIG. Note that "simple → elephant holding" means, for example, waiting until one of the plurality of accesses issued earlier is completed as described above. (9) Timing T4 shown in Figure 3
For example, if the access to the file a1 is completed, the program 5 executes the process 18 shown in FIG. 2, and then executes the process 19 shown in FIG.

Access to file A is performed, for example, in units of one track, and when format writing for one track is completed, completion of access to file A is notified. If file A spans multiple tracks, this does not mean that the formatting and writing of the entire file A has been completed yet. Therefore, "NO" is returned to the process 19 shown in FIG. 2, and the second access to the file a1 is performed at the timing shown in FIG. (1) Program 5 again has a single event (SWAIT) at timing T6 shown in FIG.
).

Then, at the timing T7 shown in FIG. 3, the file B,
When the access to the file b1 is completed, the program "5 goes through the processes 18 and 19 shown in FIG. 2 again and prepares for the second access to the file b1. Then, at timing T8 shown in FIG. 3, the access to the file B is performed. After that, similar processing is performed, for example, in the third image.
Assume that at timing T9 in the figure, access to file A is completed, and at this time all format writes to file A are completed.

At this time, around program 5, the processes 20 and 21 shown in the second diagram are executed, and the contents of the device table 7 are checked. give instructions to that effect. Then, at the timing TlO shown in FIG. 3, 0S4
accesses file A2. The same applies to αa file B, and the timing Tl shown in FIG.
File B2 is accessed at 1.

(1) After that, format writing is performed on files A, B3, etc., and at timing Tl2 shown in FIG. 3, the program 5 enters the process 22 shown in FIG.

(b) Also, at timing Tl3 shown in FIG. 3, processing 22 and processing 23 shown in FIG. 2 are similarly entered.

In addition, in the above explanation, the processing executed by the program 5 has been outlined, but once again, the steps shown in Figure 3 *1, *2, . . .
. . . If the processing indicated by *7 is listed accurately, it can be considered as follows. That is, (b) processing *1.

Checking the input information (9 shown in Figure 1), determining the amount of space, determining the degree of multiplicity, and initial processing for each table 6, 1, and 8 (day processing*2).

Determining free modules, determining unprocessed devices, displaying modules in use, displaying devices in processing, securing and opening files, editing buffers, and requesting file access.

(O treatment*3.

Holds a blank module and a condolence event.

O treatment*4.

Judgment and check of access completion module, completion judgment of each file, buffer editing, and file access request.

(g) Processing*5.

Judgment and check of access completion module, completion judgment for each file, closing and returning of files, completion judgment for each device, securing and opening of new files, editing buffers, and file access requests.

[F] Processing*6.

Judgment and check of access completed module, completion judgment for each file, file closing and return, completion judgment for each device, empty display for module, processed display for device, judgment of empty module, unprocessed device judgment, and judgment of the processing device.

6 processing*7.

In addition to processing *6, it has a single event.

In the case of the present invention, the program 5 processes as described above using the module table 6, device table 7, and file table 8.

On the other hand, one problem with conventional processing systems is that a portion of the module table 6 and the file
Table 8 was prepared, but device table 7 did not exist (...For this reason, in the case of conventional processing, program 5 uses module table 6.
First, based on device A, file A,
While requesting access to , the following happened.

That is, when the program 5 attempts to perform processing corresponding to the second module again, the contents of the file table 8 are examined in order and an access request is made to, for example, file A2. That is, the link relationships among modules, devices, and files are not clear. On the other hand, in the case of the present invention, since the device and the file are linked by the device table 7, files A,, a, . . . Access requests are not made one after another. In addition, in the case of the present invention, as mentioned above, the single event (SWAIT)
Therefore, the program 5 can run immediately when any one event ends. When the number of events is large, that is, when the degree of multiplicity is relatively large, the occurrence of the event-holding state is extremely rare, making it possible to effectively request access. Furthermore, since each of the above-mentioned tables is generated based on the information 9 that is actually given when the process is actually executed, there is no need to take too much or too little space in advance as in the past. can be largely prevented.

As explained above, according to the present invention, it is possible to perform, for example, format write processing on a large storage file very effectively, and the burden on the operator can be significantly reduced.

[Brief explanation of drawings]

FIG. 1 shows the configuration of an embodiment for carrying out the method of the present invention, FIG. 2 is a flow chart of an embodiment for explaining the method of the present invention, and FIG. 3 is a time chart for explaining the operation of the present invention. In the figure, 1 represents a main storage device including a central processing unit, 2A and 2B are large storage devices, Al,
a2ta3, ゛゜゜゜゜゜, Bl, b2, b3,...
. . . respectively represent files, 4 is an operating system, 5 is a program, 6 is a module table, 7 is a device table, 8 is a file table, and 9 is an input section (or input information) such as file information. ing.

Claims (1)

[Claims] 1. In a scheduling processing method for multiplexing files stored in a plurality of large storage devices, there is provided a module table that stores information regarding individual modules to be multiplexed, and the module - A device table that stores information that links multiple files belonging to the device corresponding to each device registered in each module of the table, and a device table that stores information that links multiple files belonging to the device, and a unique information necessary for processing the file for each given file. A file table for storing information is created, and each of the above tables is created by at least determining and securing the file space amount prior to execution of processing on the large storage device, and when executing the processing, the program After referring to the module table and extracting one unprocessed device belonging to an empty module, extract the one that is linked by the device table above.
Determine one file, start processing execution based on the contents of the file table, and start processing execution for the file if all of the plurality of modules mentioned above are busy or all of the plurality of devices mentioned above are busy. When the above-mentioned busy state occurs, wait for one of the processes based on the process execution start for the file that was activated first to finish, and then start them sequentially at the end of the process. and further characterized in that when the processing for one file belonging to the one device is completed, the processing for other files belonging to the device is sequentially executed based on the contents of the device table. Multi-processing scheduling processing method. 2. The multiprocessing schedule according to claim 1, wherein the number of modules stored in the module table is determined at the time of creating each table to correspond to the number of available devices. Ring processing method.