JP3374480B2 - Data processing device - Google Patents

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to execution using a CPU.
The present invention relates to a data processing device that executes body dispatch, and more particularly, to a data processing device that realizes reduction of overhead in the dispatch.

[0002] multitasking, a data processing apparatus which operates in multi-process environment, to use the CPU program
Executable body (hereinafter, in this specification, this executable body is composed of
Is sometimes referred to as a resource, and the resources that use the CPU
The source (execution form) may be named as CPU resources) is grayed Le <br/>-loop the sometimes adopt a configuration for executing the distribution control of the CPU resources. When adopting such a CPU resource dispatch processing mode, it is necessary to construct a configuration that reduces the overhead of the operating system in dispatching the CPU resource in order to reduce the load on the operating system.

[0003]

2. Description of the Related Art In a data processing apparatus having a configuration in which CPU resources are grouped and CPU resource allocation control is performed, conventionally, the assigned amount of CPU resource usage designated for each group and the actual group The CPU usage is compared with the used CPU usage, and if the actual CPU usage is less than the allocated capacity, the CPU is preferentially allocated and the group that uses more CPU than the allocated capacity is allocated. On the other hand, a method of dispatching CPU resources was adopted according to the method of reducing the CPU usage.

That is, conventionally, CPU resources are grouped and assigned to a CPU resource group.
CPUs belonging to each CPU resource group according to the allocated amount
When adopting a configuration that dispatches resources, by actually measuring and evaluating the CPU usage of each group,
The method of allocating CPUs was adopted.

[0005]

However, according to such a conventional technique, it is necessary to actually measure and evaluate the CPU usage of each group, so that the number of groups increases or each group increases. If the number of execution units belonging to the CPU increases, the overhead of the operating system for measuring / evaluating the CPU usage increases, and thus the overhead of the operating system for dispatching CPU resources increases. there were.

The present invention has been made in view of the above circumstances, and has a configuration in which CPU resources are grouped and the distribution control of the CPU resources is executed, and the overhead in dispatching the CPU resources is reduced. It is an object of the present invention to provide a new data processing device that realizes the above.

[0007]

FIG. 1 shows the principle configuration of the present invention. In the figure, reference numeral 1 is a data processing apparatus equipped with the present invention, which groups CPU resources and executes prescribed data processing while executing CPU resource distribution control processing.

This data processing device 1 is provided with table means 2
, Generating means 3, issuing means 4, and queuing means 8
With . The table means 2 is composed of a plurality of entries. Each entry of this table means 2 is assigned to each CPU resource group according to the CPU allocation amount of the CPU resource group, and each CPU
It points to the group representative table 7-i (i = 1 to n) located at the head of the queue 6-i (i = 1 to n) that queues the execution bodies 5 belonging to the resource group.

The generating means 3 is composed of, for example, a timer means, a random number generating means and a counter means, and sequentially or regularly generates numerical information associated with the identifier of the entry of the table means 2.

The issuing means 4 is an execution unit 5 which is a CPU resource.
Queue 6 when there is a dispatch request for
-Dispatch the execution body 5 queued at i (i = 1 to n). The queuing means 8 is the above-mentioned group.
If a new executable 5 is created by specifying
Newly created at the end of queue 6-i for the selected group
Queue executor 5 and do not specify the above group
If a new execution body 5 is created, it is necessary to create it.
The maximum of queue 6-i of the group to which the executing body 5 of the request-issuing source belongs
The newly created execution body 5 is queued later.

[0011]

In the present invention, the issuing means 4 executes
If there is de Isupatchi request that a new execution body 5 to stop, read the numerical information generated by the generating means 3, the read numerical table means associated with the information 2
Identify the entry.

The entry specified in this way is
A large number of CPU resource groups are assigned to a CPU resource group with a large CPU allocation amount, and a small number of CPU resource groups are assigned to a CPU resource group with a small CPU allocation amount because of a small number. In this way, the CPU resource group is specified according to the CPU allocation amount.

When the entry of the table means 2 is specified, the issuing means 4 specifies the queue 6-i connected to the group representative table 7-i (i = 1 to n) indicated by the entry, and the queue 6-i. Executive body 5
One execution unit 5 is selected from among these and dispatched.

As described above, according to the present invention, the CPU resources are grouped and assigned to the CPU resource group.
C that belongs to each CPU resource group according to the PU allocation amount
At the time of adopting a configuration of dispatching PU resources, the dispatch processing can be executed without performing the processing of measuring and evaluating the CPU usage amount. The overhead can be greatly reduced.

[0015]

EXAMPLES The present invention will be described in detail below with reference to examples. FIG. 2 illustrates an embodiment of the data processing device 1 including the present invention. In the figure, the same components as those described in FIG. 1 are indicated by the same symbols.

Reference numeral 10 is a search table corresponding to the table means 2 in FIG. The search table 10 has, for example, 100 entries, each of which is assigned a group representative table 7-i (i = 1 to n).
Point to.

11 is a table setting program,
Called at the time of IPL or in response to a processing request from the user, a pointer is set between the entry of the search table 10 and the group representative table 7-i (i = 1 to n).

Reference numeral 12 denotes an execution unit creation program, which is called in response to a processing request from a user to create a new execution unit 5 and a CP to which the created execution unit 5 belongs.
Queuing at the end of queue 6-i of the U resource group. Specifically, when a CPU resource group is designated and called, as shown in (a) of FIG. 3, the created execution body 5 is added to the end of the queue 6-i of the designated CPU resource group. On the other hand, when the CPU resource group is called without specifying the CPU resource group, the created execution body 5 is queued to the CPU resource group to which the calling execution body 5 belongs, as shown in (b) of FIG. 6-i
Queue at the end of.

Reference numeral 13 denotes a time slice control program, which divides the CPU monopoly right of the execution unit 5 into fixed time intervals, and the execution unit 5 that has passed that time belongs to the C to which it belongs.
By queuing at the end of the queue 6-i of the PU resource group, it is controlled so that all the execution bodies 5 can use the CPUs evenly.

Reference numeral 14 is a timer, which measures time.
Reference numeral 15 denotes a dispatch control program, which is called at the time of a dispatch request of the execution body 5 which is a CPU resource, and dispatches the execution body 5 queued in any of the queues 6-i (i = 1 to n).

FIG. 4 shows an example of a processing flow executed by the table setting program 11, and FIG. 5 shows an example of a processing flow executed by the dispatch control program 15. Next, the dispatch control processing of the execution unit 5 configured according to the present invention will be described in detail according to these processing flows.

First, the processing flow executed by the table setting program 11 will be described. At the time of IPL or when there is a processing request from the user, the table setting program 11 first, as shown in the processing flow of FIG. 4, first in step 1, the CPU allocation amount x _i of each CPU resource group.
To set. This setting process is performed according to a setting instruction from the user, and for example, the CPU allocation x ₃ of the third CPU resource group is 20%, for example, each CPU allocation x according to the percentage display.
_i will be set.

Next, in step 2, the search table 10
N _i of the number of entries and the CPU allocation x _i set in step 1 are calculated. For example, the number N of entries in the search table 10 is “100”, and the CPU allocation amount x ₃ of the third CPU resource group is “20%”.
If it is, the multiplication value N ₃ indicating the value “20” is calculated according to the multiplication process.

Then, in step 3, the search table 1
The entry of 0 is divided according to the multiplication value N _i calculated in step 2, and a pointer is placed between each entry and the group representative table 7-i located at the head of the queue 6-i in units of the division. Stretch. For example, the third C
When the multiplication value N ₃ of the PU resource group is “20”, 20 entries are acquired from the entries of the search table 10, and those entries are obtained as the group representative table of the third CPU resource group. Draw a pointer to link with 7-3.

In this way, according to the processing of the table setting program 11, each entry of the search table 10 is assigned to each CPU resource group according to the CPU allocation amount of the CPU resource group and belongs to each CPU resource group. Queue 6 for queuing execution body 5
The point is to point to the group representative table 7-i located at the beginning of -i.

Next, the processing flow executed by the dispatch control program 15 will be described. When the dispatch control program 15 receives a new dispatch request for the executing body 5 due to the execution of the executing body 5 being stopped due to termination or waiting for I / O, as shown in the processing flow of FIG. In step 1, the time generated by the timer 14 is read.

Next, in step 2, the numerical value of the last two digits of the time read in step 1 is specified. For example, the numerical value of the last two digits of the time expressed in microsecond units is specified.

Then, in step 3, the entry of the search table 10 having the same number as the numerical value specified in step 2 is accessed, and the execution unit 5 connected to the group representative table 7-i pointed to by the accessed entry. Among them, the executable body 5 which is executable and closest to the head is specified and dispatched.

In this way, the dispatch control program 15 selects an entry in the search table 10 from the time when the timer 14 is generated and dispatches the execution body 5 belonging to the CPU resource group pointed to by the selected entry. It is processed as follows.

In this way, each entry of the search table 10 is allocated to each CPU resource group according to the CPU allocation amount according to the processing of the table setting program 11, and the dispatch control program 1
According to the processing of No. 5, the selection is made in a distributed manner without being concentrated on a specific one. As a result, the execution bodies 5 belonging to each CPU resource group can be dispatched according to the CPU allocation amount allocated to the CPU resource group.

Although the illustrated embodiment has been described, the present invention is not limited to this. For example, in the example,
Although the configuration in which the entry of the search table 10 that is the access destination is specified using the time generated by the timer 14 has been disclosed, the present invention is not limited to this.
A numerical value generated by the random number generating means is used to specify the entry of the search table 10 that is the access destination, or a counter value that uses the number of entries as the remainder system to specify the entry of the search table 10 that is the access destination. The configuration may be adopted.

[0032]

As described above, according to the present invention,
Processing for measuring and evaluating the CPU usage amount when the CPU resources are grouped and the CPU resources belonging to each CPU resource group are dispatched according to the CPU allocation amount assigned to the CPU resource group. It is possible to execute the dispatch processing without performing the above-mentioned processing and without requiring special judgment materials such as queue movement and dispatch prohibition flag. from now on,
Compared with the prior art, the operating system overhead in dispatching CPU resources can be greatly reduced.

[Brief description of drawings]

FIG. 1 is a principle configuration diagram of the present invention.

FIG. 2 is an example of the present invention.

FIG. 3 is an example of execution processing of an execution body creation program.

FIG. 4 is a processing flow executed by a table setting program.

FIG. 5 is a processing flow executed by a dispatch control program.

[Explanation of symbols]

1 Data processing device 2 table means 3 generation means 4 issuing means 5 Execution body 6 cues 7 group representative table

─────────────────────────────────────────────────── ─── Continuation of front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06F 9/46

Claims (4)

(57) [Claims] 1. A group the executing entity that uses the CPU, the data process unit that executes a distribution control of the executable units to be assigned to the CPU, is composed of a plurality of entries, the CPU allocation having on the SL group If the number of the entries and table means employing a configuration that is assigned to the upper SL group, the new execution body by specifying the group is created in response
Newly created at the end of the queue for the specified group
Queue the executed body and specify the above group
If a new executable is created without
New at the end of the queue of the group to which the requesting execution body belongs
Queuing for queuing newly created executables
Means, and generating means for numerical information associated sequentially generated regularly or irregularly to the identifier of an entry in the table means, when there is a dispatch request for execution body, numerical generated by the generating means Information a data processing apparatus that comprises an issuing unit for dispatching the execution body being queued, and wherein the entry over the table means associated with the. 2. The data processing apparatus according to claim 1, wherein the generating means generates time information associated with an entry identifier by using time information generated by a timer means. Processing equipment. 3. The data processing device according to claim 1, wherein the generating means generates the numerical value information associated with the entry identifier by using the random numerical value information generated by the random number generating means. And data processing device. 4. The data processing apparatus according to claim 1, wherein the generating means generates numerical information associated with an entry identifier by using a counter value having the number of entries as a remainder system. And data processing device.