JP2001117818A - Memory managing device, managing method, and recording medium recorded with managing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To predict a page to be access in near future. SOLUTION: When an access request is inputted to an address conversion part 2, an access pattern storage part 7 stores the number of the page requested to be accessed, the access frequency, and a history of page number differences. A page preread part 8 extracts a history matching the access history of a currently accessed page from histories in the access pattern storage part 7 and predicts a page to be accessed next from the page number difference corresponding to the page right after the currently accessed page in the extracted history.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a memory management device, a management method, and a recording medium on which a management program is recorded, and more particularly to a memory management for managing a simulation memory mounted on a simulated debugger that simulates and executes software on a host machine. The present invention relates to an apparatus, a management method, and a recording medium that records a management program.

[0002]

2. Description of the Related Art FIG. 16 is a block diagram showing an example of a conventional memory management device. This memory management device manages a simulation memory of a simulator that simulates and executes software. Referring to FIG. 1, a conventional memory management device 100 includes an address conversion unit 101, a host machine memory 102, a cache control unit 103, a disk input / output unit 104, and a disk device 105. . The disk device 105 stores the actual simulation memory required for the simulation.

In such a configuration, when an access to a simulation memory is requested, an address conversion unit 101 converts an access address into a position of the simulation memory, and the cache control unit 103 transmits a request to the host machine memory 1.
02 or the disk device 105
It is determined whether the access is to the upper simulation memory.

[0004] At this time, in the case of accessing the simulation memory on the disk device 105, a transfer process from the disk device 105 to the host machine memory 102 is performed once so that the access on the host machine memory 102 can be performed.

As described above, the read from the simulation memory on the disk device 105 to the host machine memory 102 or the changed host machine memory 102
Is written back to the disk device 105.

[0006]

However, the prior art has the following problems. The first problem is that the actual simulation execution is interrupted when reading from the simulation memory on the disk device 105 to the host machine memory 102. The reason is that in the conventional memory management device 100, the simulation target software cannot predict the page to be accessed in the near future, and cannot transfer the page from the disk device 105 to the host machine memory 102 in advance. The page to be accessed is the host machine memory 1
02, the simulation cannot be performed unless the simulation is interrupted and the simulation memory is transferred from the disk device 105 to the host machine memory 102. Therefore, the conventional memory management device 100
Then, the interruption of the simulation increases.

[0007] The second problem is that there is a lot of useless writing and reading to the disk device 105. The reason is that when a page is rewritten from the host machine memory 102 to the disk device 105, it cannot be determined whether or not there is a page that may be used in the near future. there is a possibility. Therefore, when accessing the rewritten page, it is necessary to perform the read operation again.

On the other hand, Japanese Patent Application Laid-Open No.
230247 (hereinafter referred to as Reference 1),
JP-A-332053 (hereinafter referred to as Reference 2), JP-A-6-289999 (hereinafter referred to as Reference 3), JP-A-8-115259 (hereinafter referred to as Reference 4), and JP-A-8-185272 ( Hereinafter, it is referred to as Document 5.)
Is disclosed.

The technique disclosed in Reference 1 is to prefetch four to seven sectors of the same cylinder and head when a host requests access to two or three sectors of a certain cylinder and head. . In the technique disclosed in Document 2, the state of the memory as a result of the simulation of the CPU and the environment in which the state changes according to the simulation are filed, the contents stored in the file are read out, and the state changes according to the state of the memory and the simulation. Restoring the environment. The technique disclosed in Document 3 is to prefetch and register data of a sector in which the same auxiliary code as the sector requested to be read from the host system is written in a prefetch buffer.
The technique disclosed in Document 4 calculates a difference between a requested address and an address accessed in the past every time a write request is made, and counts numerical information corresponding to block number information that matches the calculated difference. In addition, the number of blocks to be prefetched is determined based on the correspondence between the block number information and the numerical information. The technique disclosed in Reference 5 records the address and length of this command in an address register, searches the address table using the address of this command as a key, and acquires the next access address and length information.

The present invention provides a look-ahead method that is completely different from the look-ahead methods disclosed in References 1 to 5. Further, means for solving the second problem is not disclosed in Documents 1 to 5.
Therefore, an object of the present invention is to provide a memory management device, a management method, and a recording medium on which a management program is recorded, which can predict a page to be accessed in the near future and can prevent useless writing and reading to a disk device. To provide.

[0011]

According to a first aspect of the present invention, there is provided a first invention including a first storage means and a second storage means, wherein the first storage means is provided in response to an access request to the second storage means. A memory management device that transfers predetermined information of information stored in a second storage unit to the first storage unit and performs information processing based on the information transferred to the first storage unit, The memory management device is the second
History storing means for storing a history of the address and the number of times of access to the storage means, an access history from the latest access request to the second storage means to an access request going back a predetermined number, and the history storing means. Address prefetching means for calculating an address predicted to be requested to access the second storage device next time from the accumulated address and the history of the number of accesses.

Further, a second invention according to the present invention includes a first storage means and a second storage means, wherein the information stored in the second storage means in response to an access request to the second storage means. A memory management method in a memory management device for transferring predetermined information to the first storage means and performing information processing based on the information transferred to the first storage means. A history accumulating step of accumulating a history of the address and the number of times of access to the second storage means, an access history from the latest access request to the second storage means to an access request going back a predetermined number, and the history accumulating means. For calculating an address predicted to be requested to access the second storage unit next time from the address stored in the storage unit and the history of the number of accesses. Characterized in that it comprises a reading step.

[0013] A third invention according to the present invention includes a first storage means and a second storage means, among the information stored in the second storage means in response to an access request to the second storage means. A recording medium for recording a memory management program used in a memory management device for transferring predetermined information to the first storage means and performing information processing based on the information transferred to the first storage means, A history storing step of storing a history of the address and the number of times of access to the second storage means in the recording medium, and a storage of a predetermined number of access requests from the latest access request to the second storage means; Based on the access history and the history of the addresses and the number of times of access stored in the history storage means, it is predicted that the next access request to the second storage means will be made. Memory management program and an address prefetching calculating the less is characterized in that it is recorded.

According to the first to third aspects of the present invention, a page to be accessed in the near future is calculated in order to calculate an address which is predicted to be requested next time on the basis of an address for which access has been requested and a history of the number of accesses. Can be predicted, and useless writing and reading to the disk device can be prevented.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, an outline of the present invention will be described. FIG. 1 is a configuration diagram of an example of a memory management device according to the present invention. In FIG. 1, a simulation memory 6
However, when managing a large-capacity simulation memory that cannot be realized on the host machine memory 3, the management is performed using the disk device 6 in units of a memory of a certain size called a page. When accessing the simulation memory, the simulation memory is once read into the host machine memory 3 and the simulation is executed on the host machine memory 3.
The access to the simulation memory is recorded as an access history in page units, and based on the recorded access history, an access pattern of a page at a specified time interval is determined and stored.

If access to the simulation memory is requested and there is no corresponding page in the host machine memory 3, the page is fetched from the disk device 6. At this time, in order to transfer a page to be predicted next access from the disk device 6 to the memory of the host machine 3 in advance, a page to be used next is determined using the access pattern of the page already stored. Predict, disk device 6
From the host machine memory 3 and from the host machine memory 3 to the disk device 6.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 shows the configuration of the first embodiment of the present invention. Referring to FIG. 1, a memory management device 1 according to the present invention includes an address translation unit 2, a host machine memory 3, a cache control unit 4, a disk input / output unit 5, a disk device 6, an access pattern storage unit. 7 and a page prefetch unit 8.

The outline of these components is as follows. The address converter 2 has a function of converting an address accessed from software to be simulated into a position of a simulation memory on the disk device 6. The host machine memory 3 has a function of simulating access to an actual memory. The cache control unit 4 has a function of allocating a part of the simulation memory to the host machine memory 3 that performs actual memory access. Disk input / output unit 5
Has a function of converting a simulation memory into a position stored on the disk device 6 and a function of inputting / outputting data to / from the actual disk device 6. Disk device 6
Stores the substance of the simulation memory required for the simulation. The access pattern storage unit 7
It has the function of patternizing and accumulating access to the simulation memory. The page prefetch unit 8 has a function of prefetching the simulation memory using the access pattern obtained by the access pattern storage unit 7.

Next, the overall operation of this embodiment will be described in detail with reference to FIGS. 2 and 6 are flowcharts showing the operation of the memory management device, FIG. 3 is a schematic diagram for explaining extraction of an access pattern, FIG. 4 is a schematic diagram for explaining storage of an access pattern, and FIG. 5 is an access pattern. And FIG. 7 is a schematic diagram for explaining an access pattern search.

First, when an access request to the simulation memory is made (step A1 in FIG. 2), the address is converted into a position in the simulation memory by the address conversion unit 2 (step A2 in FIG. 2). Next, the access pattern is extracted and stored from the page accessed by the access pattern storage unit 7 (step A3). In parallel with this, the cache control unit 4 determines whether the simulation memory exists on the host machine memory 3 or on the disk device 6 (step A4). (If "Yes" in step A5), the position of the simulation memory is converted to a position on the host machine memory 3 and an actual access is performed (step A6).

On the other hand, if it does not exist in the host machine memory 3 ("NO" in step A5), the disk input / output unit 5 converts it to a storage position on the disk device 6 which matches the address of the simulation memory. (Step A7). Further, the contents of the simulation memory at the corresponding address from the disk device 6 are transferred to the host machine memory 3 through the cache control unit 4 (step A).
8) The position of the simulation memory is converted to a position on the host machine memory 3 and an actual access is performed (step A6). In parallel with this operation, the page prefetch unit 8
The next page is predicted from the currently accessed page, and the page is prefetched (step A9).

Next, details of the page prefetching procedure will be described. Referring to FIG. 3, the access pattern storage unit 7 extracts as a history the page number for which an access request has been made, the number of times the page has been accessed, and the difference between the accessed pages. That is, the last access page number and the number of times the page has been accessed, the previous access page number and the number of times the page has been accessed, the current access page number and the number of times the page has been accessed are extracted as the page access history (FIG. 3). (A) and the difference between the accessed pages, that is, the difference between the page number two times before and the previous page number, the difference between the previous page number and the current page number, and the like are extracted as the page access history (FIG. 3 (B)). The configuration of the simulation memory space of the disk device 6 is shown on the right side of FIG. In FIG. 3B, the vertical axis indicates the access page number, and the horizontal axis indicates the elapsed time. That is, FIG.
(B) shows, as an example, that the access page number of the last access page is the largest, the previous access page number is the smallest, and the current access page number is in the middle, and the length of the arrow in the elapsed time direction, that is, The length of the elapsed time indicates the number of times each page number has been accessed.

Further, the access pattern accumulating section 7 divides the access history into fixed sections as shown in FIG. 4, and stores the divided access histories as access patterns 1, 2,. FIG. 4 (A)
In the access history, access pattern 1,
4B, the content stored as pattern 1 is shown in FIG. 4 (B), and the content stored as pattern 2 is shown in FIG. 4 (C).

Further, an access pattern as shown in FIG. 5A is stored in the access pattern storage unit 7, and when the storage location is full, the one which is used less frequently is selected as a discard target. To add a new access pattern. That is, in the example shown in FIG.
Is used less frequently, so this is targeted for disposal.
As shown in FIG. 5B, the update pattern is stored in the storage location of this pattern 3.

Next, the operation of the page prefetch section 8 will be described in detail with reference to FIGS. First, an access pattern is created from the access history of the currently accessed page (step A11 in FIG. 6). FIG. 7B shows the current access pattern. The thick line portion of the current access pattern indicates the access history up to the current time, and the tip of the arrow indicates the current time. The part ahead of the current time is the part of the prediction page.

Next, a search is made for a pattern whose created pattern matches the access pattern already stored in the access pattern storage section 7 (step A12 in FIG. 6).
FIG. 7A shows an example in which the access pattern matches the creation pattern among the access patterns already stored. On this occasion,
A width is provided for the value of the stored access pattern, and it is determined that they match if the value is within the range of the width. Then, when they match (in the case of “Yes” in step A13 of FIG. 6), a location that matches the current access position is determined by the number of accesses from the found pattern, and the next page to be accessed is determined by the access pattern (FIG. 6). The page number of the next page to be read is obtained using the difference between the page numbers stored in the access patterns 4 (B) and (C) (step A14 in FIG. 6). In order to read the predicted page, the free space of the host machine memory 3 is determined (step A1 in FIG. 6).
5) If there is no free space in the host machine memory 3 (“No” in step A15 in FIG. 6), the disk device 6 is transferred from the host machine memory 3 via the disk input / output unit 5.
(Step A16 in FIG. 6). On the other hand, if there is a vacancy (in the case of “present” in step A15 in FIG. 6), the page to be prefetched is moved from the disk device 6 to the host machine memory 3 via the disk input / output unit 5 (step A17 in FIG. 6). . When the access pattern does not match the stored access pattern (in the case of "none" in step A13 of FIG. 6), the access pattern storage unit 7 additionally stores the access pattern as an access pattern at regular intervals based on the current access history (FIG. 6). Step A18 in FIG. 6).

Next, a specific example will be described. FIG. 8 is a schematic diagram illustrating an example of a page access history, FIG. 9 is a schematic diagram illustrating an example of an access pattern to be stored, FIG. 10 is a schematic diagram illustrating an example of calculating a page number for performing page prefetching, and FIG. It is a schematic diagram which shows the specific example of page prefetch.

First, as shown in FIG. 8, the page number requested from the first (page number 10) to the sixth (page number 18) of the simulation memory space in order from the address translation unit 2 and the number of access times of the page. , And the value of the page number difference (the difference between the previously accessed page number and the currently accessed page number and stored as a positive or negative value) are stored in the access pattern storage unit 7 as a page access history. Is done. Further, as shown in FIG. 9, the access pattern storage unit 7 stores
The generated plurality of access patterns (three patterns of patterns 1 to 3 in the example of FIG. 9) are stored separately for each pattern.

Next, as shown in FIG. 10, the page look-ahead section 8 compares the current access history with a plurality of stored access patterns, and finds out the accessed page number or the value of the page number difference. An access pattern having the same or an approximate value is searched. FIG.
In the example of (A), the page number of the current access history is 1
3 and 27, and the page number difference is 0 and 14,
A pattern 1 having page numbers 10 and 25 close to this number and a page number difference close to this page number difference of 0 and 15 is searched as an access pattern having an approximate value (FIG. 10). (See pattern 1 in (B)).

Next, a page number predicted to be accessed in the future is calculated from the page number difference stored in the access pattern 1. Referring to pattern 1 in FIG. 10B, page number 3 is stored after page number 25, and the difference is −22. Therefore, the page prefetch unit 8 adds -22 to the currently accessed 27 pages (see FIG. 10A), and sets the numerical value 5 obtained as a result of the addition as the next predicted page. By the same calculation, 15, 17, and 20 are obtained as subsequent prediction pages.

The predicted page number calculated here (FIG. 11)
(A) is read from the simulation memory on the disk device 6 to the host machine memory 3 via the cache control unit 4 via the disk input / output unit 5 as shown in FIG. It is. As a result of these operations,
A page predicted to be accessed next to the currently accessed page can be stored in the host machine memory 3 in advance.

Next, a second embodiment will be described. FIG. 12 is a configuration diagram of the second embodiment. In addition,
In the figure, the same components as those in the configuration of the first embodiment (see FIG. 1) are denoted by the same reference numerals, and description thereof will be omitted. Referring to FIG. 12, the memory management unit 10 includes an address translation unit 2, a host machine memory 3, a cache control unit 4, a disk input / output unit 5, a disk device 6,
It includes a write-back page determination unit 9 in addition to the access pattern storage unit 7 and the page pre-reading unit 8.

The function of the write-back page determination unit 9 added in the second embodiment is that when writing back from the host machine memory 3 to the disk device 6, a page which is likely to be accessed soon is calculated by the page prefetch unit 8. Based on the number, a page not close to that value is selected and written back to the disk device 6 via the disk input / output unit 5.

The operation of the write-back page determination section 9 is shown in FIG.
This will be described in detail with reference to FIG. FIG. 13 is a flowchart showing the operation of the second embodiment. First, a search is performed to determine whether the page to be written back to the disk device 6 is on the currently used access pattern (step A2 in FIG. 13).
1). If the access pattern does not exist (see FIG. 1)
In the case of "No" in step A22 of step 3), the corresponding page is written back to the disk device 6 using the disk input / output unit 5 as it is (step A25 in FIG. 13). On the other hand, when the corresponding page exists on the access pattern (in the case of “present” in step A22 in FIG. 13), the page is left in the host machine memory 3 (step A23 in FIG. 13). Further, when there is another page that is a write-back candidate (in the case of “Yes” in step A24 of FIG. 13), step A21
Step A23 is repeated. On the other hand, when there is no write-back candidate (in the case of “none” in step A24 of FIG. 13), the oldest page on the host machine memory 3 is written back to the disk device 6 (step A26 of FIG. 13). According to the second embodiment, access to the disk device 6 can be effectively used.

Next, a specific example of the operation of the second embodiment is shown in FIG. FIG. 14 is a schematic diagram showing a specific example of the write-back page determination. Referring to the figure, page numbers 100, 5, 60,
When 20, 3, and 200 are registered (FIG. 14A)
), The write-back page determination unit 9 determines the candidate page to be prefetched from the page numbers included in the list and the access patterns created by the access pattern storage unit 7 and the page prefetch unit 8. It is compared whether they are the same (see the upper diagram in FIG. 14B). And
If the page numbers are the same, the cache control unit 4 is instructed to register the same page number in the list held in the host machine memory 3. In the example of FIG. 14, the page numbers 5 and 20 registered in the write-back list (see the diagram on the left side of FIG. 14A) are registered as predicted page numbers in the current pattern (FIG. 14B ), The page numbers 5 and 20 are registered in the host machine memory 3 (see the figure on the right side of FIG. 14A).

On the other hand, if the page is not the same as the candidate page to be pre-read, the page is left as it is on the list to be rewritten. Also, when the holding list of the host machine memory 3 is full,
The oldest page (for example, page numbers 13 and 30 in the diagram on the right side of FIG. 14A) is added to the list of pages to be written out, and the data is written back from the host machine memory 3 to the disk device 6.

Next, a third embodiment will be described. The third embodiment relates to a recording medium on which a memory management program is recorded. FIG. 15 is a configuration diagram of an example of a recording medium and a driving device that drives the recording medium. Referring to FIG. 1, the driving device is a CPU (Centr).
al Processing Unit) 21, an input unit 22, and a storage unit 23. And
The CPU 21 controls the memory management devices 1 and 10 described above. On the other hand, the recording medium 24 has the above-described FIGS.
The memory management program for executing the processing shown in the flowchart of FIG.

Next, the operation of the driving device will be described. First, a program load instruction is input from the input unit 22 to C
When input to the PU 21, the CPU 21 reads the memory management program recorded on the recording medium 24 and writes the read memory management program into the storage unit 23. next,
When a program execution instruction is input to the CPU 21 from the input unit 22, the CPU 21 reads a memory management program from the storage unit 23 and controls the memory management devices 1 and 10 according to the program. Since the contents of the control have already been described, the description is omitted.

[0039]

According to the first aspect of the present invention, the first aspect
A second storage unit configured to transfer predetermined information of information stored in the second storage unit to the first storage unit in response to an access request to the second storage unit; 1. A memory management device for performing information processing based on information transferred to one storage means, wherein the memory management device stores a history of an address for which an access request has been made to the second storage means and a history of the number of accesses. The storage means, the access history from the latest access request to the second storage means to an access request which has been advanced a predetermined number of times and the history of the addresses and access counts stored in the history storage means, Address prefetch means for calculating an address predicted to be requested to be accessed, so that a page to be accessed in the near future can be predicted, and It is possible to prevent unnecessary writing and reading of the disk device.

Specifically, the first effect is that the simulation speed can be improved. The reason for this is that the read-ahead from the disk device that stores the simulated memory to the cache memory is performed in parallel with the actual simulation, thereby reducing the transfer waiting time from the disk device to the memory of the host machine and reducing the simulation time. This is because interruption can be reduced.

The second effect is that it is possible to grasp how the software to be simulated operates. The reason is that since the pattern of the accessed page is stored, the execution history of the software to be simulated and the access history to the simulated memory can be extracted. This makes it possible to trace the movement of the software to be simulated in a certain section.

A third effect is that access to the disk device can be used effectively. The reason for this is to write back to the disk device while avoiding pages that may be accessed shortly.

Further, the second and third inventions according to the present invention have the same effect as the first invention.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an example of a memory management device according to the present invention.

FIG. 2 is a flowchart illustrating an operation of the memory management device.

FIG. 3 is a schematic diagram for explaining extraction of an access pattern.

FIG. 4 is a schematic diagram for explaining storage of an access pattern.

FIG. 5 is a schematic diagram for explaining updating of an access pattern.

FIG. 6 is a flowchart illustrating an operation of the memory management device.

FIG. 7 is a schematic diagram for explaining search of an access pattern.

FIG. 8 is a schematic diagram illustrating an example of a page access history.

FIG. 9 is a schematic diagram showing an example of an access pattern to be stored.

FIG. 10 is a schematic diagram illustrating an example of calculating a page number for performing page prefetching.

FIG. 11 is a schematic diagram showing a specific example of page prefetching.

FIG. 12 is a configuration diagram of a second embodiment.

FIG. 13 is a flowchart showing an operation of the second embodiment.

FIG. 14 is a schematic diagram showing a specific example of a write-back page determination.

FIG. 15 is a configuration diagram of an example of a recording medium and a driving device that drives the recording medium.

FIG. 16 is a configuration diagram of an example of a conventional memory management device.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 memory management device 2 address conversion unit 3 host machine memory 4 cache control unit 5 disk input / output unit 6 disk device 7 access pattern storage unit 8 page read-ahead unit 9 write-back page determination unit 24 recording medium

Claims (27)

[Claims] A first storage unit and a second storage unit;
In response to an access request to the second storage means, predetermined information among information stored in the second storage means is transferred to the first storage means, and based on the information transferred to the first storage means, A memory management device for performing information processing, wherein a history accumulating unit accumulating a history of an address and an access count of an access request to the second storage unit; Address prefetch means for calculating an address predicted to be requested next time to the second storage means from an access history up to a number of access requests up to the previous number and a history of addresses and access times stored in the history storage means; A memory management device comprising: And a prediction information transfer means for reading information corresponding to the predicted address calculated by the address prefetch means from the second storage means and writing the read information to the first storage means. Claim 1 characterized by the following:
A memory management device as described. 3. The memory management device according to claim 1, wherein the address stored in said history storage means includes address difference information between a current address and a previous address. 4. The history of addresses and the number of accesses stored in the history storage means is divided and stored at regular intervals, and the address prefetching means stores the divided history and the access history to the second storage means. 4. The memory management device according to claim 1, wherein a predicted address is calculated by collating the memory address. 5. The apparatus according to claim 4, wherein the address prefetching means extracts a history that matches the access history to the second storage means from the histories divided by the history accumulating means by the collation. A memory management device as described. 6. The address prefetching unit calculates a predicted address from difference information included in a history that matches the access history to the second storage unit among the histories divided by the history accumulation unit. Claim 4
Or the memory management device according to 5. 7. A write-back address determining means for determining which address information of the information written in said first storage means is to be written back to said second storage means, Selecting an address which does not correspond to the predicted address calculated by the address prefetch means as a write-back address.
7. The memory management device according to any one of claims 6 to 6. 8. The memory management device according to claim 1, wherein the address is set for each page. 9. The memory management device according to claim 1, wherein the information stored in said second storage means is information for simulating and executing software. 10. A storage device comprising: a first storage unit and a second storage unit, wherein predetermined information of information stored in the second storage unit in response to an access request to the second storage unit is stored in the first storage unit. A memory management method in a memory management device for performing information processing based on the information transferred to the first storage means, wherein an address and an access count for which an access request has been made to the second storage means A history accumulating step of accumulating the history of the second storage means; and an access history from the latest access request to the second storage means to an access request that goes back a predetermined number of times and a history of addresses and access counts accumulated in the history accumulation means. Address prefetching step of calculating an address predicted to be requested next time for access to the storage means. Method. 11. A prediction information transfer step of reading information corresponding to the predicted address calculated in the address prefetch step from the second storage means and writing the read information to the first storage means. 11. The memory management method according to claim 10, wherein: 12. An address stored in the history storing step includes address difference information between a current address and a previous address.
Or the memory management method according to 11. 13. The history of addresses and access counts accumulated in the history accumulation step is divided and accumulated at regular intervals, and in the address prefetching step, the divided history and the access to the second storage means are stored. 13. The memory management method according to claim 10, wherein a predicted address is calculated by collating with a history. 14. The method according to claim 1, wherein in the address prefetching step, a history that matches the access history to the second storage means is extracted from the history divided in the history accumulation step by the collation. 14. The memory management method according to claim 13. 15. In the address prefetching step, a predicted address is calculated from difference information included in a history that matches the access history to the second storage unit among the histories divided in the history accumulation step. The memory management method according to claim 13 or 14, wherein 16. A write-back address determining step of determining which address information of the information written in said first storage means is to be written back to said second storage means, 16. The memory management method according to claim 10, wherein an address that does not correspond to the predicted address calculated in the address prefetching step is selected as a write-back address. 17. The memory management method according to claim 10, wherein the address is set for each page. 18. The memory management method according to claim 10, wherein the information stored in said second storage means is information for simulating and executing software. 19. A storage device, comprising: a first storage unit and a second storage unit, wherein predetermined information of information stored in the second storage unit in response to an access request to the second storage unit is stored in the first storage unit. And a memory management program for use in a memory management device for performing information processing based on the information transferred to the first storage means, wherein an access request is issued to the second storage means. A history storing step of storing a history of the addresses and the number of times of access, a history of access from the latest access request to the second storage means to an access request going back a predetermined number, and the addresses and accesses stored in the history storage means. An address prefetching step of calculating an address predicted to be requested to access the second storage means next time from the history of the number of times. Recording medium recording a non-memory management program. 20. A method according to claim 20, further comprising the step of reading information corresponding to the predicted address calculated in said address prefetching step from said second storage means and writing the read information into said first storage means. 20. The recording medium according to claim 19, wherein: 21. The address stored in the history storing step includes information on a difference between a current address and a previous address.
Or the recording medium of 20. 22. The history of addresses and access counts accumulated in the history accumulation step is divided and accumulated at regular intervals, and in the address prefetching step, the divided history and the access to the second storage means are stored. 22. The recording medium according to claim 19, wherein a predicted address is calculated by collating with a history. 23. The apparatus according to claim 23, wherein in the address prefetching step, a history matching the access history to the second storage means is extracted from the history divided in the history accumulation step by the collation. 23. The recording medium according to 22. 24. The address prefetching step, wherein a predicted address is calculated from difference information included in a history that matches the access history to the second storage unit among the histories divided in the history accumulation step. 24. The recording medium according to claim 22, wherein: 25. A write-back address determining step of determining which address information of the information written in the first storage means is to be written back to the second storage means, 25. The recording medium according to claim 19, wherein an address that does not correspond to the predicted address calculated in the address prefetching step is selected as a write-back address. 26. The recording medium according to claim 19, wherein said address is set in page units. 27. The recording medium according to claim 19, wherein the information stored in said second storage means is information for simulating and executing software.