JP2769367B2 - Multiprocessor scheduling method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION [Industrial Application Field] In a multiprocessor operating system in which a plurality of processes are operated in parallel by a multiprocessor system including a plurality of processors, particularly when a high-speed system is constructed or a single processor Program created for a multi-processor system,
The present invention relates to a process scheduling method suitable for operating without any change.

[Related Art] A multi-microprocessor using a plurality of microprocessors has been actively developed. JP-A-56-127261,
A multiprocessor system is one example of this, and shows a system configuration in which a plurality of processors are connected by a bus. A multiprocessor operating system is also being actively developed. Unlike multiprocessor operating systems, unlike single processor systems, multiple processes need to be able to run on different processors simultaneously,
For this, it is necessary to have a strategy for deciding which processor to assign a process to. In this case, the most commonly used method is to allocate the processor to the least loaded processor in order to maximize the utilization efficiency of the processor. This strategy has been widely adopted because it is generally easy to implement. This is because the process scheduler of the operating system operating on each processor may select the next process to be executed in accordance with the priority of the process as in the case of the single processor system. A processor with a heavy load has a long process execution time, and the process scheduler is activated less frequently. Conversely, a processor with a light load has many timings when it enters an idle state having no process to be processed, and therefore, there are also many activation timings of a process scheduler for determining whether to shift to the idle state. For this reason, a processor with a light load is given more opportunities for process scheduling, and the load among the processors automatically changes so as to be even. The above method is a rational process allocation method in a homogeneous multiprocessor system having the same function.

Also, regarding a method of managing and scheduling processes as a group, see “Experience Using Multi-Processor Systems-A Status Report” (“Experience Using Multi-proc
essor Systems-A Status Report ") (ACM Comput
ing Surveys, Vol. 12, No. 2, p. 121 to 165, 1980) is described as group scheduling. This means that, when scheduling processes that can be processed in parallel as a group, these processes are executed on different processors as simultaneously as possible.

[Problem to be Solved by the Invention] A process scheduler having a function of grouping a plurality of processes and executing processes belonging to the same group as much as possible with the same processor is preferable.

The processes are grouped here because of the following user requirements and hardware and operating system construction requirements (note that this group is a group of processes that handle the same data).

In a multiprocessor using a shared bus and a shared memory, each processor has a cache memory to reduce the load on the bus. In such a system, when one processor updates data in the cache in order to guarantee data consistency between caches, the updated result is transferred to a cache having the same data.
If processes that share a large amount of data with each other are executed by different processors, the number of data transfers between caches increases, the cache performance is not exhibited, and the bus load increases. In conventional multiprocessor systems, sufficient consideration has not always been given to this problem.

It is desirable that existing software developed for a single processor be executed on a multiprocessor system without any change. In a multi-programming environment, as is well known, contention for multiple processes can occur. To solve this problem,
When accessing shared data, it is necessary to perform exclusive control using a semaphore mechanism, etc., but the processing overhead of the semaphore mechanism becomes a problem, so that the priority of the process is the same, or a processor interrupt that has a shorter processing time than the semaphore mechanism In many cases, measures such as raising the mask to disable interrupts are taken. These means make use of the fact that, when viewed microscopically on a single processor, the processing on the processor is always performed sequentially. In reality, there are many software programs that do not perform exclusive control on shared data for performance reasons. For this reason, when existing software is executed as it is in a multiprocessor system, there is a problem that normal operation is not necessarily compensated.

The conventional group scheduling is suitable for parallel processing because the processors in the group are operated by different processors, but cannot cope with the above problem.

A prior application related to the present application is Japanese Patent Application Laid-Open No. 1-60092, entitled "Operating System for Multiprocessor," which is the same applicant. This prior application has a process management table, in which a process occupied processor name is recorded, and a process under the control of another processor is intercepted under certain conditions to perform processing in its own management. This prior application does not include the record of the processor used last time or the description of the process management for each group.

SUMMARY OF THE INVENTION An object of the present invention is to provide a process scheduling method capable of improving the performance of a cache memory and reducing the load on a bus by grouping and managing processes and executing processes in the same group on the same processor in a multiprocessor system. Is to provide.

It is another object of the present invention to provide a process scheduling method that can use existing software for a single processor system without any change.

Still another object of the present invention is to provide a process scheduling method capable of maintaining a uniform work load between processors even when executing a group of processes on the same processor.

Means for Solving the Problems To achieve the above object, a process scheduling method according to the present invention comprises a plurality of processors, wherein the plurality of processors are controlled by a single operating system. In, according to the process management table that manages the process,
A process scheduling method in which the process scheduler of the operating system assigns any one of the processors to a process. The process management table describes a group identifier that allows duplication in the process management table, and stores information of the process having the same group identifier. A group management table is provided, and processor assignment information indicating a processor that executes a process in the group is described in the group management table. The process scheduler refers to the process management table for a process selected to be executed next. Then, by determining the group of the process, and by operating the process in the state of the processor specified by the processor assignment information of the group management table of the group, the process belonging to the same group, It is obtained so as to run on one processor.

When the execution wait time of the process is stored in the process management table, the process scheduler refers to the process management table for the process selected to be executed next. When the processor is operated by an arbitrary processor and the execution waiting time is equal to or less than a certain value, the processor refers to the process management table to determine a group of the process, and specifies a processor designated by the processor assignment information in the group management table of the group. The above process may be operated.

Further, a time during which no process of the group is executed, that is, a group execution waiting time, is stored in the group management table, and the process scheduler refers to the process management table for a process selected to be executed next. Then, the group of the process may be determined, and when the group execution waiting time of the group management table of the group is equal to or longer than a certain value, the processor allocation information of the group management table may be changed.

Further, the process management table stores the execution waiting time of the process and the identifier of the processor that executed the process last time, and the process scheduler refers to the process management table for the process selected to be executed next. When the waiting time is less than a certain value, the processor which executed the process last time or the processor indicated by the processor assignment information in the group management table is executed, and when the waiting time is more than a certain value, the processor can be operated by other processors. You may.

Further, the number of times the assignment of the process to the processor is rejected may be used instead of the execution waiting time of the process.

In addition, the number of times the process was not re-assigned to the processor was used instead of the time during which no process in the group was running, ie, the group execution wait time.

As another process scheduling method of the present invention,
In a multiprocessor system comprising a plurality of processors, wherein the plurality of processors are controlled by a single operating system, a processor that needs to switch a running process is executed according to a process management table that manages the process. A process scheduling method for executing a process scheduler of the operating system and determining a process to be executed next by the processor, wherein a group identifier that allows duplication is described in the process management table, and the group identifier is the same. A group management table for storing information of the process is provided, and processor allocation information indicating a processor that executes a process in the group is described in the group management table.
From among the executable processes, the process specified by the group identifier of the process management table and the processor allocation information of the group management table indicating the processor that is executing the process scheduler is executed next. Is also good.

The process management table stores the execution wait time of the process in the process management table, and the process scheduler refers to the process management table when selecting the next process to be executed. The process may be executable.

Further, the time when no process of the group is executed, that is, the group execution waiting time is stored in the group management table, and the process scheduler selects the process to be executed next when the process scheduler selects the next process to execute. The group of the process is referred to, and when the group execution waiting time of the group management table of the group is equal to or longer than a certain value, the processor allocation information of the group management table is changed so that the process can be executed. Good.

Further, the process management table stores the execution waiting time of the process and the identifier of the processor that executed the process last time, and refers to the process management table when the process scheduler selects the process to be executed next. When the waiting time is equal to or less than a certain value, only when the processor which executed the process last time or the processor indicated by the processor assignment information in the group management table is the same as the processor which is executing the process scheduler, Executable, and when the waiting time is equal to or longer than a certain value, it may be executable.

Further, the number of times the assignment of the process to the processor is rejected may be used instead of the execution waiting time of the process.

Further, the number of times that the process was not re-assigned to the processor was used instead of the time during which no process in the group was running, ie, the group execution wait time.

As another process scheduling method of the present invention,
In a multiprocessor system comprising a plurality of processors, wherein the plurality of processors are controlled by a single operating system, a process scheduler of the operating system may assign a process to one of the processors according to a process management table for managing the process. A process management table that describes a priority of each of the plurality of processors and a group identifier that allows duplication in the process management table, and stores information of the process having the same group identifier. , Processor assignment information for designating a processor that executes a process in a group is provided in the group management table, and a processor in the group management table indicated by the group identifier of the process management table is provided. It may be set higher the priority degree service allocation information corresponding to the processor to instruct.

[Operation] The process scheduling method of the present invention provides a multi-processor system including a plurality of processors.
In a process that can be executed by any of a plurality of processors, the process is intentionally grouped, and the processor that executes the group is limited to a specific processor.

If a process having a large amount of shared data is executed by different processors, communication for cache coherence frequently occurs, and the cache performance is reduced. According to the present invention, such processes can be grouped and executed by the same processor, so that communication for cache matching is not required, and performance can be improved. Furthermore, in order to prevent the load between processors from shifting by executing a group by a specific processor, a process that has been left unexecuted for a certain period of time can be executed by any processor. Load balancing is possible. However, this function can be prohibited by setting a process fixing flag.

Further, according to the present invention, a process group requiring exclusive control can be operated by the same processor, and an existing process for a conventional single processor that does not consider exclusive control on a shared memory is modified. Normal operation can be guaranteed even if used without any problem. This does not mean that other processors are not used. Processes other than the above-mentioned process group can be executed by any processor without such restrictions, and the functions of the multiprocessor are effectively used. can do.

Embodiment An embodiment of the present invention will be described below with reference to FIG.
In FIG. 1, 13-1 to 13-2 are processors, 12-1
12-8 are processes, and 15-1 to 15-3 are groups.

Processors 13-1 to 13-3 are processes 12-1 to 12-
Execute while switching 8. Process 12-1 to 12-3
Are grouped as a group 15-1, processes 12-4 and 12-5 are grouped as a group 15-2, and processes 12-6 to 12-8 are grouped as a group 15-3. The instruction on how to group the processes is given by the user, but may be automatically given by the compiler or the operating system.

Processors that execute processes are determined for each group, and processes belonging to the same group are executed by the same processor. In FIG. 1, the number of groups and the number of processors are equal, but these numbers do not need to be particularly the same, and either number may be larger or smaller.

FIG. 2 shows a system configuration according to the present invention. In FIG. 2, 13-1 to 13-3 are processors, 11-1 to 11-3.
Is a processor number, 12-1 to 12-6 are processes, 14-1
Is a timer, 4 is a connection bus, 5 is a shared memory, 31 is an OS program, 311 is a scheduler, 32-1 to 32-4 are process management tables, 321-1 to 321-4 are processor identifiers, 322-1 to 322 -4, priority; 323-1 to 323-4, execution waiting time; 324-1 to 324-4, counters;
25-4 is a process fixed flag, 236-1 to 326-4 are group identifiers, 33-1 to 33-2 are group management tables,
331-1 to 331-2 are processor identifiers, 332-1 to 332-
2 is an execution waiting time, 333-1 to 333-2 are counters, 334
-1 to 334-2 are group fixed flags.

The processors 13-1 to 13-3 are connected to each other by the connection bus 4 and to the shared memory 5. Each processor has processor numbers 11-1 to 11-3, respectively. In this embodiment, the processors 13-1 to 13-3 have processor numbers "1", "2", and "3", respectively. In this embodiment, the setting of the processor number is performed by a hardware switch. However, as another method, a method of storing the number in a hardware ROM in advance, or a method of setting a program at the time of startup can be considered.

One of the processors 13-1 to 13-3 has a timer 14-1, and a timer interrupt enters the processor at regular intervals.

Each process runs on a processor. The shared memory 5 includes an operating system program 31 and a process scheduler 31 which is a part of the program 31.
1. A process management table 32 (32-1 to 32-) which is a part of a table managed by this operating system.
3) There is a group management table 33 (33-1 to 33-2).

The following information is stored in the process management table 32. First, the processor identifier 21 (321-1 to 31-2) indicating the processor number of the processor to execute the process
21-4). In the embodiment, the process management table 32-
The processor identifier of “1” is “1”, and the process 0 (12-1) corresponding to this process management table has “1” in the processor number 11-1 and must be executed by the processor 13-1. Is shown. If the processor that executes the process is undecided, set this processor identifier to 0. If you do not want to determine the processor to execute and any processor can execute the process, set the processor identifier to a negative value. Keep it.

Second, the priority 322 (322) for determining the execution order of the process
-1 to 322-4). The smaller the numerical value of this priority is, the higher the priority of the process is, which indicates that the process must be executed in preference to other processes.

Third is the execution waiting time 323 (323-1 to 323-4). The execution waiting time indicates a time during which the process is in an execution waiting state. When the process is executed, the execution waiting time is returned to zero.

The fourth is a counter 324 (324-1 to 324-4).
This counter is a counter that counts the number of times the execution of the process has been rejected.

Fifth, the process fixing flag 325 (325-1 to 325-
4). This contains information on whether to fix the process to the processor. If the process is fixed, the process is not executed by a processor other than the processor indicated by the processor identifier 321.

Sixth is a group identifier 326 (326-1 to 326-4). Here, information indicating which group the process belongs to is entered. The group identifier and the group management table have a one-to-one correspondence.

The following information is described in the group management table. First, the processor identifier. This indicates the processor number that executes the process belonging to the group.

The second is the execution waiting time 322 (322-1 to 322-2). The execution waiting time is a time during which none of the processes belonging to the group is being executed by the processor and any of the processes is in the execution waiting state. If any process in the group is executed, the execution waiting time is returned to zero.

Third is a counter. This counter counts the number of times that execution of a process belonging to the group was rejected.
It is cleared to 0 when a process in the group is executed.

Fourth is a group fixing flag. When this flag is set, the processor that executes the group is fixed.

FIG. 3 shows the operating system program 31.
3 shows an algorithm of the process scheduler 311 which is a part of the process scheduler 311.

First, from the processes waiting to be executed,
Select the process to be executed next according to the priority (31
1-1). The group identifier 326 is read from the process management table of the selected process. The group management table is searched based on the group identifier (311-2).

Processor identifier 331 in this group management table
Is read, and the following processing is performed according to this value (311−
3).

If the value is a positive value, the processor executing the scheduler program 311 reads its own processor number from 11 (11-1 to 11-3), and reads the value of the positive processor identifier 331 and Compare (311-5). If they do not match, the process with the next highest priority is selected (311-1). If they match, the selected process is assigned to its own processor (311-6).

When the processor identifier in the group management table is 0, the value of the processor identifier is changed to the value of its own processor number (11) (311-4), and the selected process is assigned to its own processor (311-6). ).

When the processor identifier is a negative value, the selected process is unconditionally assigned to its own processor (31
1-6).

FIG. 4 shows a program 31 of the operating system.
15 shows another embodiment of the algorithm of the process scheduler 311 which is a part of the embodiment.

First, from the processes waiting to be executed,
Select the process to be executed next according to the priority (31
1-1). Process management table 32 of the selected process
Reads the processor identifier 321 and compares it with its own processor number 11 (311-7). If the values are equal, the selected process is assigned to its own processor (311-6).

If not equal, it is determined whether or not the process fixing flag 325 is set (311-8). If so, re-select the process with the next highest priority (311
-1).

If not, check whether the execution waiting time 323 of the process management table 32 has exceeded the threshold value (311
-9).

If it exceeds, the processor identifier 321 of the process management table 32 is set to its own processor number (311−
14) Assign the selected process to yourself (311−
6).

If it does not exceed, the group management table is searched from the group identifier 326 in the process management table 32, and it is checked whether the processor identifier in the table is equal to its own processor number (311-10). If they are the same, execute 311-14 again.

If they are not equal, the group fixing flag 334 in the group management table 33 is checked (311-11).

When the flag is set, the process with the next highest priority is checked (311-1).

When the flag is not set, it is checked whether the execution waiting time 332 of the group management table 33 has exceeded the threshold value. If not, the process with the next highest priority is checked (311-1).

If it exceeds, the processor identifier 331 of the group management table 33 and the processor identifier 321 of the processor management table 32 are set to their own processor number 11 (3
11-13, 311-14) and execute the selected process (311
-6).

FIG. 5 shows a program 31 of the operating system.
15 shows another embodiment of the algorithm of the process scheduler 311 which is a part of the embodiment.

First, from the processes waiting to be executed,
Select the process to be executed next according to the priority (31
1-1). Process management table 32 of the selected process
Then, the processor identifier 321 is read and compared with its own processor number 11 (311-7). If the values are equal, the selected process is assigned to its own processor (311-6).

If not equal, it is determined whether or not the process fixing flag 325 is set (311-8). When standing, the counter 324 of the process management table 32 and the counter 333 of the group management table 33 are incremented (311-17, 311-18), and the process with the next highest priority is selected again (311-1). .

When the process fixing flag is not set, it is checked whether the counter 325 of the process management table 32 has exceeded the threshold value (311-15).

If it exceeds, the processor identifier 321 of the process management table 32 is set to its own processor number (311−
14) Assign the selected process to yourself (311−
6).

If not exceeded, the group management table is searched from the group identifier 326 in the process management table 32, and it is checked whether the processor identifier in the group management table is equal to its own processor number (311-10). If they are the same, execute 311-14 again.

If they are not equal, the group fixing flag 334 in the group management table 33 is checked (311-11). If the flag is set, execute 311-17.

When the flag is not set, it is checked whether the counter 333 of the group management table 33 has exceeded the threshold value. If not, execute 311-17.

If it exceeds, the processor identifier 331 of the group management table 33 and the processor identifier 321 of the processor management table 32 are set to their own processor number 11 (3
11-13, 311-14) and execute the selected process (311
-6).

FIG. 6 describes in detail the method of assigning the selected process of 311-6 in FIGS. 3, 4 and 5 to its own processor.

First, the register of the process that has been executed is saved (312-1). This is so that the process can be resumed from the point where it was interrupted when the order of execution is returned to the process later.

Next, the execution wait time 323 of the process management table 32 of the selected process is set to 0 (312-2).

Similarly, the execution waiting time 332 of the group management table 33 of the group to which the selected process belongs, the counter 324 of the process management table 32, and the counter 333 of the group management table are also set to 0 (respectively, 312-3, 312-4, 312-
5).

Since each process has an independent user space, the user space is switched to that of the selected process (312-6).

Finally, the value of the register saved when the process was interrupted last time is restored (312-7). Since the register to be saved also includes the program counter, the processing is restarted from the place where it was interrupted last time.

Reference numeral 313 in FIG. 7 shows an interrupt processing routine sent by the timer of FIG. 14-1 to the processor. Timer interrupts occur at 10 ms intervals.

When a timer interrupt occurs, the processor first checks the entire process management table 32, and for a process that is in an execution waiting state, the execution waiting time 323
Is incremented (313-1).

Next, the entire group management table 33 is checked, and the waiting time 333 of the group that does not include the running process but includes the process waiting to be executed is incremented (313-2).

Another embodiment of the system configuration diagram of the present invention shown in FIG.
Shown in the figure. In FIG. 8, 13-1 to 13-3 are processors, 11-1 to 11-3 are processor numbers, and 12-1 to 12-6.
Is a process, 14-2 is a timer, 4 is a connection bus, 5 is a shared memory, 31 is an OS program, 311 is a scheduler, 32
-1 to 32-4 are process management tables, 321-1 to 321-
4 is a processor identifier, 325-1 to 325-4 are process fixing flags, 326-1 to 326-4 are group identifiers, and 327-
1 to 327-4 are CPU time, 328-1 to 328-4,329-1 to 329
-4,330-1 to 330-4 are priorities, 33-1 to 33-2 are group management tables, 331-1 to 331-2 are processor identifiers, and 334-1 to 334-4 are group fixed flags.

The processors 13-1 to 13-3 are connected to each other and to the shared memory 5 by a connection bus 4. Each processor has processor numbers 11-1 to 11-3, respectively. In this embodiment, the processors 13-1 to 13-3 have processor numbers "1", "2", and "3", respectively.

One of the processors 13-1 to 13-3 has a timer 14-2, and a timer interrupt enters the processor at regular intervals. In the embodiment, two types of interrupts are provided every 10 ms and every 1 s.

Each process runs on a processor. The shared memory 5 includes an operating system program 31 and a process scheduler 31 which is a part of the program 31.
4. Process management table 32 (32-1 to 32-) which is a part of the table managed by this operating system
3) There is a group management table 33 (33-1 to 33-2).

The following information is stored in the process management table 32. First, the processor identifier 21 (321-1 to 31-2) indicating the processor number of the processor to execute the process
21-4). In the embodiment, the process management table 32-
The processor identifier of “1” is “1”, and the process 0 (12-1) corresponding to this process management table has “1” in the processor number 11-1 and must be executed by the processor 13-1. Is shown. If the processor that executes the process is undecided, the processor identifier is set to 0. If the processor to be executed is not particularly determined and any processor can execute the process, the processor identifier is set to a negative value.

Second, the process fixing flag 325 (325-1 to 325-
4). This contains information on whether to fix the process to the processor. If the process is fixed, the process is described in the processor identifier 321.

Third is a group identifier 326 (326-1 to 326-4). Here, information indicating which group the process belongs to is entered. The group identifier and the group management table have a one-to-one correspondence.

Fourth is CPU time. This is the CPU time used by the process and is used to calculate the priority. However,
Since this CPU time is reduced to 1/4 every second, it is not an accurate CPU time.

Fifth, priorities 328 to 330 for determining the execution order of the process
It is. Each process has a plurality of priorities because each processor has a priority. The smaller the numerical value of this priority is, the higher the priority of the process is, which indicates that the process must be executed in preference to other processes having a higher numerical value of priority.

In the group management table, the processor identifier 331 (3
31-1 to 331-2), in which the number of a processor that executes a process belonging to the group is described.

FIG. 9 shows the scheduler program 31 in FIG.
4 shows an algorithm. The processor executing the scheduler program sequentially checks the priority of the process management table corresponding to itself. Then, the process with the highest priority (the numerical value of the priority is the smallest) is selected (314-1). However, at this time, a process in which the process fixing flag 325 of the process management table 32 is set and the process identifier 321 is not its own processor number is excluded.

Next, the selected process is assigned to its own processor (314-2).

Reference numeral 315 in FIG. 10 denotes a processing routine for an interrupt generated every 10 ms by the timer 14-2. Here, the CPU of the process management table 32 of the process executed by each processor is described.
Time 327 is incremented.

Reference numeral 316 in FIG. 11 denotes a processing routine for an interruption every 1 s generated by the timer 14-2. Here, the priorities 328 to 330 of each process management table 32 are recalculated. 31
Processing is performed for each process in 6-1.
In step 3, the processing is performed for each processor. 316−
In 2, the CPU time 327 is reduced to 1/4 every second so as not to become too large.

The priority is calculated in 316-4 to 316-7. The priority is basically equal to the CPU time of the process, but 20 is added to the priority of the processor not described in the processor identifier 321 of the process management table 32. Also,
Also, 20 is added to the priority of the processor not described in the processor identifier of the group management table 33.

As a result, the priority of the processor described in the processor identifier of the process management table or the group management table is increased, and the process is easily executed.

FIG. 12 shows a case where a program developed for a single processor is executed by a multiprocessor of a conventional scheduling method. In FIG. 12, 12-1 to 12-1
12-2 is a process, 41-1 and 41-2 are interrupt prohibition processes, 42-1 and 42-2 are resource use processes, and 43-1 and 43-2 are interrupt permission processes.

The single processor executes one process at a time. Therefore, many programs that are supposed to be executed by a single processor are used by simply prohibiting interrupts even when using shared resources. This is because the processing is faster than using locked resources. As long as interrupts are disabled in a single processor, only one process is instantaneously executed, so that multiple processes do not use resources at the same time.

However, in a multiprocessor, as shown in FIG.
Even if interrupts are disabled in 41-1 and 41-2, 42-1 and 42-2
, Multiple processes may use resources at the same time.

FIG. 13 shows a process using the same resources, showing a case where a program for a single processor is executed by a multiprocessor using the scheduling method of the present invention.
If 12-1 and 12-2 are put into the same group and executed by the processor 0, the use of the resources of 42-1 and 42-2 is always executed in series, and there is no conflict between resource use.

In recent years, instead of a conventional process that individually has a program and data, a thread and a light weight process that share a program and data with each other have attracted attention.

Since threads that share programs and data have a large amount of shared data, the improvement in cache performance when grouping is further enhanced.

Also, when threads that share programs and data are grouped and executed by one processor, switching between programs and data is often unnecessary, and the processing time of the OS can be shortened.

Therefore, the present invention is effective even when applied to such a thread or a light weight process.

[Effects of the Invention] The process scheduling method according to the present invention provides a multi-processor system including a plurality of processors.
In a process that can be executed by any of a plurality of processors, the processes are intentionally grouped, and the processor that executes the group is limited to a specific processor.

If a process having a large amount of shared data is executed by different processors, communication for cache coherence frequently occurs, and the cache performance is reduced. According to the present invention, such processes can be grouped and executed by the same processor, so that communication for cache matching is not required, and performance can be improved. Furthermore, in order to prevent the load between the processors from shifting due to the execution of the group by a specific processor, for a process located without being executed for a certain period of time, it is possible to execute on any processor, Load balancing is possible. However, this function can be prohibited by setting a process fixing flag.

Further, according to the present invention, a process group that requires exclusive control can be operated by the same processor, and the existing process for a conventional single processor that does not consider the exclusive control for the shared memory can be modified. Normal operation can be guaranteed even if used without. This does not mean that other processors are not used. Processes other than the above-mentioned process group can be executed by any processor without such restrictions, and the functions of the multiprocessor are effectively used. can do.

[Brief description of the drawings]

1 is a schematic diagram of process scheduling according to an embodiment of the present invention, FIG. 2 is a system configuration diagram, FIG. 3 is an algorithm diagram of a scheduler, FIG. 4 is another embodiment of a scheduler algorithm, FIG. FIG. 6 is another embodiment of the algorithm of the scheduler, FIG. 6 is an algorithm diagram for allocating the selected process to its own processor, FIG. 7 is a diagram of a timer interrupt processing routine, and FIG. 8 is another embodiment of the system configuration. , FIG. 9 is an algorithm diagram of the scheduler, FIG. 10 is a processing routine of a timer interrupt every 10 ms, FIG. 11 is a processing routine of a timer interrupt every 1 second, and FIG. Diagram when executing a program for
FIG. 13 is an explanatory diagram when a program for a single processor is executed by the scheduling method of the present invention. 4 connection bus, 5 shared memory, 11 processor number, 12 process, 13 processor, 14 timer, 15 group, 31 OS program, 32 process management table , 33 ... Group management table, 31
1 ... scheduler, 321 ... processor identifier, 322 ...
… Priority, 323 …… Execution waiting time, 324 …… Counter, 32
5: Processor fixed flag, 326: Group identifier,
327: CPU time, 328 to 330: priority, 331: processor identifier, 332: execution wait time, 333: counter,
334: Group fixed flag, 41: Interrupt prohibition processing, 42: Resource use processing, 43: Interrupt permission processing

──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Yoshiki Kobayashi 4026 Kuji-cho, Hitachi City, Ibaraki Prefecture Inside Hitachi, Ltd. Hitachi Research Laboratory Co., Ltd. (72) Inventor Tomoaki Nakamura 5-2-1 Omika-cho, Hitachi City, Ibaraki Prefecture Co., Ltd. (56) References JP-A-62-75739 (JP, A) JP-A-55-83951 (JP, A) JP-A-62-187958 (JP, A) JP-A-2-122365 ( JP, A) JP-A-2-242434 (JP, A) JP-A-2-238556 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06F 15/163

Claims (16)

(57) [Claims] In a multiprocessor system comprising a plurality of processors, wherein the plurality of processors are controlled by a single operating system, a process scheduler of the operating system operates according to a process management table for managing processes. A process scheduling method for assigning any one of the processors to the group management server, wherein a group identifier allowing duplication is described in the process management table, and a group management table storing information of the process having the same group identifier is provided. The table describes processor allocation information indicating the processors that execute the processes in the group. The process scheduler refers to the process management table for the process selected to be executed next. Determine the group of the process,
A multiprocessor scheduling method, characterized in that processes belonging to the same group are executed by the same processor by operating the process in a processor designated by the processor assignment information of the group management table of the group. 2. The process management table stores an execution waiting time of the process, and for the process selected by the process scheduler to be executed next, the execution waiting time is equal to or more than a certain value with reference to the process management table. When the execution wait time is less than a certain value, the process management table is referred to to determine the group of the process, and the process is specified by the processor allocation information in the group management table of the group. 2. The multiprocessor scheduling method according to claim 1, wherein the process is operated on the selected processor. 3. The group management table stores a time during which no process of the group is being executed, that is, a group execution waiting time, and the process scheduler stores
For the process selected to be executed next, the group of the process is determined by referring to the process management table. When the group execution waiting time in the group management table of the group is equal to or greater than a certain value, the group management table of the group management table is determined. A multiprocessor scheduling method characterized in that processor assignment information can be changed. 4. The process management table stores an execution waiting time of the process and an identifier of a processor executed last time, and the process scheduler stores the process management table for a process selected to be executed next. For reference, when the waiting time is equal to or less than a certain value, the process is executed by the processor which executed the previous process or the processor indicated by the processor assignment information in the group management table. 3. The multiprocessor scheduling method according to claim 1, wherein the method is enabled. 5. The multiprocessor scheduling method according to claim 2, wherein the number of times the assignment of the process to the processor is rejected is used instead of the execution waiting time of the process. 6. The method according to claim 3, wherein a time during which no process in the group is executed, that is, the number of times the assignment of the process to the processor is rejected is used in place of the group execution waiting time. Multiprocessor scheduling method. 7. A multi-processor system comprising a plurality of processors, wherein the plurality of processors are controlled by a single operating system, the processor which needs to switch a running process is provided.
A process scheduling method for executing a process scheduler of the operating system according to a process management table for managing a process and determining a process to be executed next by the processor, wherein a group identifier that allows duplication is stored in the process management table. Write,
A group management table for recording information of the process having the same group identifier; providing processor assignment information for instructing a processor to execute a process in the group in the group management table; And executing a process next designated by the group identifier of the process management table, wherein the processor assignment information of the group management table indicates the processor executing the process scheduler. Processor scheduling method. 8. The process management table stores the execution waiting time of the process, and the process scheduler
8. The multiprocessor scheduling method according to claim 7, wherein when the process to be executed next is selected, the process management table is referred to, and when the execution waiting time is equal to or more than a certain value, the process is executable. 9. The group management table stores a time during which no process of the group is being executed, that is, a group execution waiting time, and the process scheduler stores
When selecting the next process to be executed, the group of the process is determined by referring to the process management table. When the group execution waiting time of the group management table of the group is equal to or longer than a certain value, the group management table 9. The multiprocessor scheduling method according to claim 7, wherein the processor allocation information is changed so that the process can be executed. 10. The process management table stores, in the process management table, an execution waiting time of the process and an identifier of a processor in which the process has been executed last, and when the process scheduler selects a process to be executed next, the process management is performed. Referring to the table, when the waiting time is equal to or less than a certain value, the processor that executed the process last time or the processor indicated by the processor assignment information in the group management table is the same as the processor that is executing the process scheduler. 8. The multiprocessor scheduling method according to claim 7, wherein the method is executable only when the waiting time is longer than a certain value. 11. The multiprocessor scheduling method according to claim 8, wherein the number of times the assignment of the process to the processor is rejected is used instead of the execution waiting time of the process. 12. The method according to claim 3, wherein the time when no process in the group is executed is used, that is, the number of times the process is rejected to be assigned to the processor is used instead of the group execution waiting time. 6. The multiprocessor scheduling method of 6. 13. A multi-processor system comprising a plurality of processors, wherein the plurality of processors are controlled by a single operating system. In the process scheduling method, any one of the processors is assigned a priority, a priority of each of the plurality of processors and a group identifier that permits duplication are described in the process management table, and information of the process having the same group identifier is stored. Set up a group management table,
The group management table is provided with processor assignment information indicating processors executing processes in the group, and the priority corresponding to the processor indicated by the processor assignment information in the group management table indicated by the group identifier of the process management table is set. A multiprocessor scheduling method characterized by setting a high value. 14. A method according to claim 1, wherein a code indicating that the process can be executed by any processor is provided as one of the codes of the process assignment information.
14. The multiprocessor scheduling method according to any one of items 13 to 13. 15. A method according to claim 1, wherein a flag indicating that the process can be executed only by a specific processor is provided in one or both of the process management table and the group management table. A multiprocessor scheduling method according to any one of the preceding claims. 16. The process according to claim 1, wherein exclusive control of the resources is simplified by giving the same group identifier to processes sharing the resources. Multiprocessor scheduling method.