JP2000148515A - Memory scheduling method and storage medium storing memory scheduling program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To carry out efficient memory scheduling which does not deteriorate system performance by releasing a physical memory by virtual pages of the virtual address space of a process according to the order of memory scheduling priority and the load state of the memory. SOLUTION: An access frequency managing means 60 checks the virtual address space, checks the frequencies of access to the respective virtual pages, and sets the memory scheduling priority order of a virtual page in an access frequency management area when a physical memory is assigned to the virtual page. Then the mean value of the order of priority of all virtual pages in all processes is found. A memory scheduling means 70 finds the threshold of priority order on the basis of the mean value of the priority order in consideration of the memory load state in operation. Then the priority order of the virtual page is compared with the threshold and when the priority order is less than the threshold, the physical memory is released.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a page conversion algorithm in a virtual storage mechanism, and more particularly, to a memory scheduling method capable of making detailed settings for memory release processing.

[0002]

2. Description of the Related Art In a computer system, when an available memory runs short during the operation of an OS, an operation called aging or paging is performed to release an available memory used by a user process. In the case of an aging system, the virtual space of the user process is checked, and a memory page that has not been referenced for a certain period is released. In the case of a paging system, a memory page is examined, and a memory page that has not been referred to for a certain period of time is released.

That is, in either method, a memory page to be released is selected based on whether or not it has been referenced for a certain period.

Therefore, in the conventional methods, there is a possibility that a memory page is released even when a frequently accessed file happens to be not accessed for a certain period of time.

[0005] Also, since only memory is referred to for a certain period of time as an indicator of memory release, there is a high possibility that a memory page resident in memory does not become a working set of an actual process, and the actual file Memory scheduling depending on the application could not be performed.

In some conventional systems, a function that allows a user to specify memory scheduling in a certain virtual space range is implemented as a system call. For example,
The madvis system call in UNIX corresponds to this. In the madvise system call, it is possible to specify whether access to the specified virtual space range is random or sequential.

In the case of random access, there is a possibility that a memory page once accessed may be referred to again.
Some systems are implemented so that the physical memory allocated to the virtual space range is hardly released. However, since it is a system call, the application must be modified without fail, and it is impossible to control memory scheduling for a commercially available application.

[0008]

In the above-described conventional method, since the memory is released based on whether or not the file has been referred to for a certain period, frequently accessed files may not be accessed for a certain period. Also in such a case, there is a problem that the release of the memory page is executed, and that such a phenomenon occurs frequently, so that the input / output operation to / from the external storage device occurs frequently and the system performance is deteriorated.

[0009] Also, since only whether or not the memory page has been referenced for a certain period of time serves as an indicator of memory release, there is a high possibility that a memory page resident in memory does not become the working set of the actual process, and Memory scheduling depending on the application could not be performed.

Further, in a conventional system in which a function that allows a user to specify a memory scheduling in a certain virtual space range is implemented as a system call, since the system call is required, the application must be modified without fail. There is a problem that it is impossible to control memory scheduling.

A first object of the present invention is to solve the above-mentioned drawbacks of the prior art, and to improve system performance by paging taking into account the priority based on access frequency, the priority for each file, and the priority for each use of a segment. It is an object of the present invention to provide a memory scheduling method and a storage medium for storing a memory scheduling program in a computer system which enables efficient memory scheduling without deterioration.

A second object of the present invention is to provide a memory scheduling method and a storage medium for storing a memory scheduling program in a computer system capable of performing appropriate memory scheduling for each file used for an application and each application. is there.

A third object of the present invention is to provide a memory scheduling method and a storage medium for storing a memory scheduling program in a computer system capable of setting detailed memory scheduling without remodeling an application. is there.

A fourth object of the present invention is to provide a memory scheduling method and a storage medium for storing a memory scheduling program in a computer system which enable a working set of a user process to be determined more accurately.

[0015]

According to the present invention, there is provided a memory scheduling method in a computer system using a virtual storage mechanism, wherein an access frequency managing step for setting a priority order for a virtual page in a virtual address space. And a memory scheduling step of controlling physical memory release processing based on the priority, wherein the access frequency management step is performed for the virtual address space of all processes in the computer system, the virtual address space being associated with the physical memory. A rescheduling priority is set based on an access frequency of a virtual page, a use of data of the virtual page, and a priority set in advance for each file corresponding to the virtual page, and the memory scheduling step includes: The virtual Wherein for each virtual page of less space, based on the load state of the memory scheduling priority and a memory, and executes the release processing of the physical memory.

According to a second aspect of the present invention, in the access frequency management step, the access frequency, the use of the data, and the file are weighted, and the product of the access frequency and the weight and the priority of the use of the data are set. And the product of the product of the file and the weight, and the product of the priority of the file and the weight are set as the rescheduling priority for the virtual page.

According to a third aspect of the present invention, in the access frequency management step, different weights are set for each of the access frequencies for a plurality of times in the past during a fixed period, and a sum of products of the access frequencies and the corresponding weights is set. And the sum of the product of the priority of use of the data and the weight and the product of the priority and the weight of the file is set as the memory scheduling priority for the virtual page.

According to a fourth aspect of the present invention, in the memory scheduling step, a threshold is set based on an average value of the memory scheduling priorities of all the virtual pages of all the processes and a memory addition state, When the value of the memory scheduling priority of a page is smaller than the threshold value, the physical memory is released.

According to a fifth aspect of the present invention, in the memory scheduling step, when the load state of the memory is high,
A value above the average value of the memory scheduling priorities of all the virtual pages of all the processes,
When the load state is low, a value lower than the average value is set as the threshold value.

According to a sixth aspect of the present invention, in the access frequency managing step, an average value of the memory scheduling priorities for the virtual pages in all processes and all virtual pages is obtained.

According to a seventh aspect of the present invention, in the memory scheduling step, when the value of the memory scheduling priority of the virtual page is smaller than the threshold value, and when the physical memory is unused, the physical scheduling is performed. It is characterized in that memory release processing is performed.

According to a still further aspect of the present invention, in the access frequency managing step, it is determined whether or not the priority of the file is a specific value. Irrespective of the priority determined from, set a special value as the memory scheduling priority, the memory scheduling step,
Even when the value of the memory scheduling priority is smaller than a threshold value, the physical memory is not released if it is the special value.

According to a ninth aspect of the present invention, there is provided a storage medium for storing a memory scheduling program in a computer system using a virtual storage mechanism, wherein the memory scheduling program assigns a priority to a virtual page in a virtual address space. An access frequency management step to be set; and a memory scheduling step to control release processing of a physical memory based on the priority order. The access frequency management step includes a step of setting a physical memory for a virtual address space of all processes in the computer system. Setting a rescheduling priority based on an access frequency of the virtual page associated with the virtual page, a use of data of the virtual page, and a priority set in advance for each file corresponding to the virtual page; Ring step, said each virtual page of the virtual address space of the process, based on the load state of the memory scheduling priority and a memory, and executes the release processing of the physical memory.

[0024]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a memory scheduling system in a computer system using a virtual storage mechanism according to an embodiment of the present invention.

The memory scheduling system according to the present embodiment uses a virtual address space 11 used by the process 10, an access frequency management area 12 for storing a history of access frequencies to the virtual address space 11, and a virtual address space 11. It includes a virtual address management area 13 for separate management and an external storage device 14 containing files used in the process 10.

Here, the virtual address space 11 can be used independently by the process 10. Further, since the memory-mapped file mechanism is employed, the virtual address space 1
In each memory page of No. 1, there is a file that is actually present.

The virtual address space 11 is divided into management units called segments according to its use. Management areas 13a, 13b, 13 in the virtual address management area 13 correspond to the segments 11a, 11b, 11c, respectively.
c exists. Also, each segment is associated with a file 14a, 14b, 14c for writing the contents of the memory when the memory scheduler releases the memory. Although the file management area exists in the external storage device, it resides in the memory during the operation of the system, and the management area of the file corresponding to each segment can be easily referred to during the operation of the system.

The memory scheduling priority for each use of the segment of the virtual address space 11 will be described with reference to FIG. FIG. 2 shows the use of the virtual address space 11 of the process 10 in FIG.

The segment 11a stores a text code used in the process 10. Segment 1
1b stores the initialized data used in the process 20. The uninitialized data (B
SS) is stored. The segment 11d stores a text code of a library to be shared with another process. The segment 11e stores data used in the shared library. The segment 11f is used as a stack of the process 10.

Each management area 13a, 13b, 13 of the virtual address management area 13 corresponds to each segment.
c, 13d, 13e and 13f are provided, in which memory scheduling priorities for each purpose of the segment are stored.

In this embodiment, the memory scheduling priorities of the text section, data section, BSS section, library text section, library data section, and stack section are “Ptext”, “Pdata”, and “Pb”, respectively.
ss "," Pshtxt "," Pshdat "," Ps
tck ". Further, it is also possible to set a use-specific priority unique to each process by using a system call. In that case, the set priority is overwritten in the virtual address management area 13.

The priority of memory scheduling for each file will be described with reference to FIG. FIG. 3 shows the correspondence between the process 10 and the files used by the process 10.

Segment 11a of virtual address space 11
And the segment 11b are used as a text portion and a data portion, respectively, so that these segments correspond to the file 14a which is a load module used when the process 10 is started. Segment 11d and segment 11e are shared library file 14b
It corresponds to. Segment 11c and segment 11f
Is an area uniquely used by the process 10, and since there is no corresponding file, it is associated with the secondary storage file.

The files 14a, 14b, and 14c have the memory scheduling priority "Pl
m, Plib, and Pswap are set in the file management area. It is also possible to set a new priority for a file using a system call, but the priority is effective for all processes using the file.

The access frequency management area will be described with reference to FIG. The virtual address space 11 is associated with a physical memory in a page size unit (virtual page unit) which is a minimum unit when hardware manages a memory. For each virtual page in the virtual address space 11,
An access frequency management area 12 exists.

In FIG. 4, there is a management area 12a corresponding to the virtual page 11a1, a management area 12b corresponding to the virtual page 11a2, and a management area 12c corresponding to the virtual page 11a3. The access frequency management area stores the number of accesses to the corresponding virtual page within a certain period.

In the present embodiment, the number of accesses to the virtual page for the past three times is stored. For example, the number of times of access to the virtual page 11a1 during the last certain period is “F311”, the number of times of access within the previous period is “F312”, and the number of times of access within the previous period is “F313”. is there. The memory scheduling priority calculated using the history of these access counts and the corresponding priority for each file and each application is as follows:
“P31”, which is stored in the access frequency management area 12. If an address solution exception occurs when accessing the virtual page 11a1, the value of "F311" in the corresponding access frequency management area 12a is incremented by extension.

FIG. 5 shows the configuration of means prepared by the system for process and memory scheduling. The access frequency management means 60 operates at regular intervals, and sets the virtual address space 11 of all processes in the system.
Is checked, the access frequency to each virtual page is checked, and the priority in memory scheduling is determined.

The memory scheduling means 70 operates when there is no more free memory in the system, refers to each priority set in the access frequency management area 12 of each process, and releases a memory page. The memory scheduling means 70 terminates the processing when the free memory capacity of the system has been recovered or when the processing has been performed for a preset number of processes.

Process 10, process 20, process 30, process 40, process 50 existing in the system
Are sequentially set in an array (process slot) for managing the processes existing in the system.

The operation of the access frequency management means 60 will be described below with reference to the flowchart of FIG.

The access frequency management means 60 is a system means that is activated at regular intervals. The started access frequency management means 60 first selects a process for checking the access frequency (step 601). In this embodiment, since all processes are checked, the first process in the process slot is selected.

Next, a virtual page whose access frequency is to be checked is set at the head of the virtual address space 11 (step 6).
02). Then, it is checked whether a physical memory is allocated to the virtual page currently checked (step 6).
03).

If no physical memory is allocated in step 603, the virtual page whose access frequency is to be checked is set to the next page (step 604), and the process is repeated from step 603.

In step 603, if a physical memory has been allocated, the memory scheduling priority of the virtual page is set in the access frequency management area 12 (step 605).

The history of the information on the number of accesses is obtained, and the latest information on the number of accesses is cleared to zero (step 60).
6). For example, when step 606 is executed for the access frequency management area 12a in FIG. 4, the value “F311” is set as the number of previous accesses, the value F312 is set as the number of previous accesses, and the latest access number is set. A value “0” (zero clear) is set as the number of times.

Next, it is checked whether the access frequency information has been updated for all the virtual pages (step 60).
7). If there is a virtual page that has not been checked yet, the virtual page is set to the next page and the process is repeated from step 603 (step 604). When all the virtual pages have been checked, a flag indicating that the access frequency management means 60 has performed the process for the process is set (step 608).

It is checked whether or not all the processes have been processed (step 609). If there is a process that has not been processed yet, the next process is selected (step 611), and the processes from step 602 are performed. If the process has been performed for all processes, the average value of the memory scheduling priorities of all virtual pages in all processes is obtained (step 610), and the process ends.

The method of determining the memory scheduling priority in the present embodiment will be described with reference to the flowchart of FIG. FIG. 7 is a flow chart showing step 60 in FIG.
5 is described. In the present embodiment, a value is set as a weight for determining which of the access frequency, the priority for each file, and the priority for each use of the segment is to be prioritized. Here, the weight of the access frequency is set to “Wa”, the weight of the file is set to “Wf,” and the weight for each use of the segment is set to “Ws”.

First, the access frequency management area 12 corresponding to the currently examined virtual page is obtained (step 70).
1). The number of accesses in the latest period stored therein is “F311”, the number of accesses in the previous period is “F312”, and the number of accesses in the previous period is “F313”. Since the latest access frequency information should be given more importance, the weight for the past access frequency is set small. In the present embodiment, the weight of the access frequency “F311” is “Wa”, the weight of the access frequency “F312” is “１ ／ Wa”, and the weight of the access frequency “F313” is “1/4 Wa”. The sum of the product of the number of times and the weight is set as the priority based on the access frequency (step 702).

Next, the virtual address management area 13 for the virtual page is obtained (step 703), and the priority order for each segment application is read. The product of the value and the weight "Ws" is set as the priority for each segment application (step 70).
4).

Next, the management area of the file corresponding to the virtual page is obtained (step 705), and the priority set for each file is read. The product of the value and the weight “Wf” is set as a priority for each file (step 706).

The sum of the priorities determined in steps 702, 704, and 706 is set as the memory scheduling priority of the virtual page in the access frequency area 12 (step 707).

The operation of the memory scheduling means 70 will be described with reference to the flowchart of FIG.

When the free memory of the system decreases and the memory scheduling means 70 is started, first, a process for releasing the physical memory is selected (step 80).
1). When the memory scheduling means 70 is activated last time, the process managed in the process slot next to the last selected process lot is selected. However, a process in which the flag indicating that the access frequency management means 60 has performed the process in step 608 of FIG. 6 is not selected.

The threshold value of the memory scheduling priority is determined (step 802). Here, based on the average value of the priorities determined by the access frequency management means 60, a value is determined in consideration of the memory load state during operation. Here, for example, the memory load status is set to "M
1 ”,“ M2 ”, and“ M3 ”. When the memory load status is“ M1 ”, the memory load is not so high, so the threshold is set to １ ／ of the average value of the priority order and“ M2 ”. In this case, the average value of the priorities is set. In the case of "M3", the memory load condition is extremely high, so that the average value of the priorities is set to twice.

Next, the first virtual page in the virtual address space of the selected process is selected (Step 803).
It is checked whether a physical memory is allocated to the virtual page (step 804).

If no physical memory has been allocated,
The next virtual page is selected (step 805), and the process of step 804 is performed.

If a physical memory has been allocated, the priority set in the access frequency management area corresponding to the virtual page is compared with the threshold determined in step 802 (step 806).

If the priority value is equal to or larger than the threshold value in step 806, the physical memory corresponding to the virtual page is not released, and the next virtual page is selected (step 805), and step 804 is executed.

If the value of the priority is smaller than the threshold, the correspondence between the virtual page and the physical memory is released (step 807).

Next, it is checked whether another process refers to the corresponding physical memory (step 8).
08) If not referenced, the physical memory is released (step 809).

It is checked whether or not processing has been performed on all virtual pages of the process (step 810). If there is a virtual page that has not been processed yet, the same processing is performed on the next virtual page (step 805). . If the processing for all virtual pages has been completed, it is expected that some physical memory has been released, so it is checked whether or not free memory has been recovered (step 81).
1).

If the memory load of the system is low, the processing of the memory scheduling means 70 ends. Also, when the processing is completed for the preset number of processes (step 812), the processing of the memory scheduling means 70 is terminated.

If the amount of free memory in the system has not recovered and the processing for the set number of processes has not been completed, the next process to be processed is selected (step 813).
Specifically, the process managed in the slot next to the process slot that managed the process that performed the process last is selected. If the process managed in the last slot of the process slot has been selected, the process managed in the first slot of the process slot is selected. However, as in step 801, a process that has never been processed by the access frequency management means is not selected.

Making the contents of a specific file always resident in the memory is realized by the following method. FIG. 9 is a flowchart in which a step for making the contents of a specific file always reside in the memory is added.
And FIG. 9 and 10, steps common to those in FIGS. 7 and 8 are denoted by the same step numbers. That is, in step 901 of FIG. 9, it is determined whether or not the priority set in the file is a specific value. If there is a special value, the value of the priority determined from the use of the segment and the access frequency Is ignored, and a special value, for example, the value “−1” is set as a priority in the access frequency management area (step 902). If the value is not a special value, the priority in step 707 described above is set.

Then, in step 1001 of FIG. 10, it is determined in addition to whether the priority is smaller than the threshold and whether the value of the priority is "-1". If there is, in step 805, the processing of the next page is performed without releasing the physical memory. By doing so, the contents of a specific file can always be made to reside in the memory.

When calculating a priority order based on the number of accesses in step 702 in FIG. 7, a histogram of the number of accesses to the virtual page is created for each process, and means for transferring the contents to the application is provided. Thus, the application creator can grasp the usage status of the virtual address space of the process.

The memory scheduling method described above
As well as being realized in hardware, it can also be realized in software by a memory scheduling program for realizing each of the functions described above, which is recorded on a magnetic disk, a semiconductor memory, or another recording medium of a computer system. is there. The memory scheduling program is read from the recording medium into the data processing unit (CPU) of the computer system, and controls the operation of the data processing unit to realize the functions of the above-described units and execute the memory scheduling process.
That is, the processing shown in FIGS.

It should be noted that the present invention is not limited to the above-described embodiment, but can be implemented with various modifications within the scope of the technical idea.

[0071]

As described above, according to the memory scheduling of the present invention, firstly, paging is performed by taking into account the priority based on the access frequency, the priority for each file, and the priority for each use of the segment. More detailed and efficient memory scheduling without deteriorating system performance is enabled.

Second, since the priority for memory scheduling can be set for each file, it becomes possible to set appropriate memory scheduling for each file used in the application.

Third, priorities for memory scheduling can be set in a file management area, the frequency of file access is examined at the system design stage, and priorities are given to the files before the application operates. By setting the order, it is possible to apply a preset memory scheduling to an application using the file, so that it is possible to set a detailed memory scheduling policy without modifying the application.

Fourth, since the history of the frequency of access to the memory pages used by the user process is retained, the working set of the user process can be obtained more accurately.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a memory scheduling system according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a use of a virtual address space in the embodiment of the present invention.

FIG. 3 is a diagram showing a correspondence between a process and a file used by the process in the embodiment of the present invention.

FIG. 4 is a block diagram showing a correspondence between an access frequency management area and a virtual address space according to the embodiment of the present invention.

FIG. 5 is a block diagram showing a relationship between an access frequency management unit, a process, and a memory scheduling unit according to the embodiment of the present invention.

FIG. 6 is a flowchart illustrating an operation of an access frequency management unit according to the embodiment of the present invention.

FIG. 7 is a flowchart illustrating a determination of a priority order in an access frequency management area according to the embodiment of the present invention.

FIG. 8 is a flowchart showing the operation of the memory scheduling means in the embodiment of the present invention.

FIG. 9 is a flowchart illustrating another method of determining a priority in an access frequency management area according to the embodiment of the present invention.

FIG. 10 is a flowchart showing the operation of the memory scheduling means in the embodiment of the present invention.

[Explanation of symbols]

 10, 20, 30, 40, 50 process 11 virtual address space 11a to 11c segment 12 access frequency management area 13 virtual address management area 13a to 13c management area 14 external storage device 14a to 14c file 60 access frequency management means 70 memory scheduling means

Claims (16)

[Claims] 1. A memory scheduling method in a computer system using a virtual storage mechanism, comprising: an access frequency management step of setting a priority for a virtual page in a virtual address space; and a physical memory release processing based on the priority. And a memory scheduling step of controlling the access frequency.The access frequency management step includes: for a virtual address space of all processes in a computer system, an access frequency of the virtual page associated with a physical memory; A rescheduling priority is set based on a priority set in advance for each file corresponding to a use and the virtual page. The memory scheduling step includes the step of setting the memo for each virtual page in the virtual address space of the process. Based on the load state of the scheduling priority and a memory, the memory scheduling method in a computer system and executes the release processing of the physical memory. 2. The access frequency management step, wherein the access frequency, the use of the data, and the file are weighted, and the product of the access frequency and the weight, the priority of the use of the data, the product of the weight, and 2. The memory scheduling method according to claim 1, wherein a sum of a product of a file priority and a weight is used as the rescheduling priority for the virtual page. 3. The access frequency management step sets weights having different values for each of the plurality of access frequencies in the past multiple times in a certain period, and calculates a sum of a product of the access frequency and the corresponding weight, and a use of the data. 3. The memory scheduling method in a computer system according to claim 2, wherein the sum of the product of the priority and the weight of the file and the product of the priority and the weight of the file is used as the memory scheduling priority for the virtual page. 4. The memory scheduling step sets a threshold value based on an average value of the memory scheduling priorities of all the virtual pages of all the processes and a memory addition state, and sets the memory scheduling priority of the virtual pages. 4. A process for releasing the physical memory when the value of the order is smaller than the threshold value.
13. A memory scheduling method in the computer system according to item 5. 5. The memory scheduling step, wherein when the load state of the memory is high, a value exceeding an average value of the memory scheduling priorities of the virtual pages of all the processes is determined. 5. The method according to claim 4, wherein a value less than the average value is set as the threshold value. 6. The computer system according to claim 4, wherein in the access frequency management step, an average value of the memory scheduling priorities for the virtual pages in all processes and all virtual pages is obtained. Memory scheduling method. 7. The memory scheduling step, wherein when the value of the memory scheduling priority of the virtual page is smaller than the threshold, and when the physical memory is not used, the physical memory is released. 7. The memory scheduling method in a computer system according to claim 4, wherein: 8. The access frequency management step determines whether or not the priority of the file is a specific value. If there is a special value, the priority is determined regardless of the use of the data and the priority determined from the access frequency. Instead, a special value is set as the memory scheduling priority, and the memory scheduling step includes the step of releasing the physical memory if the value of the memory scheduling priority is the special value even when the value is smaller than the threshold value. 8. The memory scheduling method in a computer system according to claim 4, wherein the memory scheduling is not performed. 9. A storage medium for storing a memory scheduling program in a computer system using a virtual storage mechanism, the memory scheduling program comprising: an access frequency management step of setting a priority order for a virtual page in a virtual address space. And a memory scheduling step of controlling physical memory release processing based on the priority order. The access frequency management step is a virtual address space of all processes in the computer system, which is associated with the physical memory. The access frequency of the virtual page, the rescheduling priority is set based on the use of the data of the virtual page and the priority set in advance for each file corresponding to the virtual page, the memory scheduling step, A storage medium for storing a memory scheduling program for executing a physical memory release process based on the memory scheduling priority and a memory load state for each virtual page in the virtual address space of the process. . 10. The access frequency management step, wherein the access frequency, the use of the data, and the file are weighted, and the product of the access frequency and the weight, the priority of the use of the data, the product of the weight, and The storage medium according to claim 9, wherein a sum of a product of a file priority and a weight is set as the rescheduling priority for the virtual page. 11. The access frequency management step sets different weights for each of the plurality of access frequencies for a plurality of times in the past during a certain period, and calculates a total sum of a product of the access frequency and the corresponding weight, and an application of the data. 11. The storage for storing the memory scheduling program according to claim 10, wherein the sum of the product of the priority and the weight of the file and the product of the priority and the weight of the file is used as the memory scheduling priority for the virtual page. Medium. 12. The memory scheduling step, wherein a threshold value is set based on an average value of the memory scheduling priorities of all the virtual pages of all the processes and a memory addition state, and the memory scheduling priority of the virtual pages is set. 2. The method according to claim 1, wherein when the rank value is smaller than the threshold value, the physical memory is released.
A storage medium for storing the memory scheduling program according to claim 1. 13. The memory scheduling step, wherein, when the load state of the memory is high, a value exceeding an average value of the memory scheduling priorities of the virtual pages of all the processes is determined. 13. The storage medium storing the memory scheduling program according to claim 12, wherein a value less than the average value is set as the threshold value. 14. The memory scheduling according to claim 12, wherein, in the access frequency management step, an average value of the memory scheduling priorities for the virtual pages in all processes and all virtual pages is obtained. Storage medium for storing programs. 15. The memory scheduling step, wherein when the value of the memory scheduling priority of the virtual page is smaller than the threshold, and when the physical memory is not used, a release process of the physical memory is performed. A storage medium for storing the memory scheduling program according to claim 12. 16. The access frequency management step determines whether or not the priority of the file is a specific value, and if there is a special value, the priority is determined regardless of the use of the data and the priority determined from the access frequency. Instead, a special value is set as the memory scheduling priority, and the memory scheduling step includes the step of releasing the physical memory if the value of the memory scheduling priority is the special value even when the value is smaller than the threshold value. 13. The method according to claim 12, wherein the step is not performed.
A storage medium storing the memory scheduling program according to claim 15.