JP5077209B2 - Memory management device, memory management method and program - Google Patents

Description

  The present invention relates to a memory management device, a memory management method, and a program, and more particularly to a memory management device, a memory management method, and a program for managing a hierarchical memory.

  There is a memory management technique for managing a hierarchical storage unit including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages, such as a virtual storage or a cache (for example, non-patent) Reference 1).

  In such a memory management technique, as a technique for selecting a page to be swapped out when a page fault occurs, an LRU (Least Recently Used) that selects a page having the longest elapsed time since the last reference was made among the pages of real storage. ), Or a page replacement algorithm such as FIFO (First In First Out) for selecting the oldest existing page from real storage pages.

  For example, the cache control device described in Patent Document 1 holds a plurality of state information for each data block on the cache, and determines a data block to be replaced (replaced) based on the state information.

  In addition, the on-demand paging method described in Patent Document 2 has an in-use main memory management table entry queue provided for each priority, and in-use main memory management corresponding to the priority of a task to which a memory page is allocated. A main memory management table corresponding to the allocated memory page is connected to the table entry queue. In this on-demand paging method, when the allocated page is emptied, the page corresponding to the main memory management table connected to the low-priority in-use main memory management table entry queue is emptied. Selected as a page.

  The information storage device described in Patent Document 3 has a function of determining a recovery priority after a failure based on whether the cache memory block is clean data or dirty data.

  Furthermore, the prefetch cache write-back method and apparatus described in Patent Document 4 has a function of checking a cache entry at a predetermined time interval, and writing back a cache entry that has been dirty and has passed a predetermined time or more. Yes.

Japanese Patent No. 3309425 JP 05-035513 A Japanese Patent Laid-Open No. 08-087450 Japanese Patent Laid-Open No. 10-232827 4. Shuichi Sakai, “Computer Architecture”, March 31, 2004; Cache and virtual memory P70-P90

  However, in the technology based on the devices, methods, and methods described in Patent Documents 1-4 described above, the swap operation when a page fault occurs takes time despite the high priority processing, and the processing is delayed. There was a problem that there was a case.

  For example, application processing for displaying a warning screen of an in-vehicle system has a low operation frequency but a high priority. Data referred to by such processing is often swapped out to the swap area because the reference frequency is low. For this reason, when an application displays a warning screen, a situation frequently occurs in which it is necessary to swap out other data in the real memory, and the write-back operation to the swap area in this swap-out operation is executed. In this case, the processing of the application is frequently delayed because it takes time.

  An object of the present invention is to provide a memory management device, a memory management method, and a program that solve the above-described problems.

The memory management device of the present invention, when a page fault occurs in access to a hierarchical storage means including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages,
Whether the target data of the memory access is priority data or non-priority data is determined based on a priority assigned in advance to the target data and a predetermined threshold, and the target data is the priority data One of the plurality of pages stored in the first storage unit, the page storing the same data as the data stored in the corresponding page in the second storage unit. Select
A page selection unit that outputs identification information of the selected page of the first storage unit;

In the memory management method of the present invention, when a page fault occurs in an access to a hierarchical storage unit including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages,
Whether the target data of the memory access is priority data or non-priority data is determined based on a priority assigned in advance to the target data and a predetermined threshold, and the target data is the priority data One of the plurality of pages stored in the first storage unit, the page storing the same data as the data stored in the corresponding page in the second storage unit. Select
The identification information of the selected page of the first storage unit is output.

The program of the present invention, when a page fault occurs in access to a hierarchical storage means including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages,
Whether the target data of the memory access is priority data or non-priority data is determined based on a priority assigned in advance to the target data and a predetermined threshold, and the target data is the priority data One of the plurality of pages stored in the first storage unit, the page storing the same data as the data stored in the corresponding page in the second storage unit. Select
The identification information of the selected page of the first storage unit is output.

  According to the present invention, it is possible to speed up a swap operation when a page fault occurs in a high priority process.

  Next, a first embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing the configuration of the present embodiment. Referring to FIG. 1, the present embodiment includes a control unit 100, a physical memory 200, and a disk device 300. The control unit 100 is also called a memory management device. The physical memory 200 and the disk device 300 are also generally called a first storage unit and a second storage unit, respectively. Further, the physical memory 200 and the disk device 300 are configured as hierarchical storage means as described below.

  The control unit 100 includes a task management table 110, a page table 120, a memory management unit 130, a page selection unit 140, a swap processing unit 150, a swap area management unit 160, and a disk device interface unit 170.

  The physical memory 200 includes a physical page of a predetermined size. The physical page of the physical memory 200 is generally called a page of the first storage unit.

  The disk device 300 includes a swap area 310 composed of a plurality of virtual pages of a predetermined size. The swap area 310 is an area on the disk device 300 in which the data contents of the physical page on the physical memory 200 can be saved when the capacity of the physical memory 200 is insufficient. The virtual page of the disk device 300 is generally called a page of the second storage means.

  The task management table 110 is a table for managing task priorities, and includes at least one task entry 119 including a task identifier 111 as a task identifier and a corresponding task priority 112 as shown in FIG. .

  As shown in FIG. 4, the page table 120 includes a plurality of page entries 129 including a physical page number 121, an allocation flag 122, a reference frequency 123, a change flag 124, and a virtual page number 125.

  The physical page number 121 is assigned to each physical page of the physical memory 200 so that each physical page can be uniquely identified. The physical page number 121 is also generally called page identification information.

  The allocation flag 122 is a flag indicating whether or not the physical page of the physical memory 200 corresponding to the physical page number 121 has been allocated to the virtual page. “1” indicates “already allocated”, and “0 ”Indicates“ not assigned ”.

  The reference frequency 123 indicates the frequency with which the physical page of the physical memory 200 corresponding to the physical page number 121 is referenced. The reference frequency 123 indicates that the higher the value is, the higher the reference frequency is, and the lower the value is, the lower the reference frequency is.

  The change flag 124 is a flag indicating whether or not the contents of the physical page of the physical memory 200 corresponding to the physical page number 121 have been changed. “1” indicates “changed”, and “0”. Indicates “not changed”. Note that the content of the physical page of the physical memory 200 corresponding to the physical page number 121 has been changed because the data stored in the physical page of the physical memory 200 is in the swap area 310 of the disk device 300. It is also included that the data stored in the virtual page is changed.

  The virtual page number 125 is assigned to each virtual page included in the swap area 310 so that each virtual page can be uniquely identified. The virtual page number 125 indicates the number of the virtual page to which the physical page of the physical memory 200 corresponding to the physical page number 121 is allocated. The virtual page number 125 is valid information when the corresponding allocation flag 122 is “1”.

  The memory management unit 130 converts the virtual address of the virtual page into the physical address of the physical memory 200 based on an instruction from a processor (not shown) or a DMA (Direct Memory Access) control unit (not shown), and transfers the physical address to the physical memory 200. Handle access. When the memory management unit 130 detects that the virtual page to be accessed does not exist in the physical memory 200, that is, detects a page fault, the memory management unit 130 outputs a page fault occurrence notification. In addition, the memory management unit 130 has a means for calculating the frequency of referring to each physical page of the physical memory 200 using, for example, an NFU (Not Frequently Used) algorithm or an aging algorithm improved from the NFU algorithm. The referred frequency is stored in the reference frequency 123 of the page table 120.

  The page selection unit 140 has a means for holding a predetermined threshold 141, and refers to the task management table 110 and the page table 120 in response to the input of the page fault occurrence notification, and determines a virtual page to be swapped out. .

  The swap processing unit 150 performs swap-in and swap-out of the page selected by the page selection unit 140. Hereinafter, an example of specific means for realizing the swap-in and swap-out processing will be described.

  First, the swap-in process will be described. The swap processing unit 150 refers to the page table 120 corresponding to the input of the virtual page number 125, and acquires the physical page number 121 of the page entry 129 whose allocation flag 122 is “0”.

  Subsequently, the swap processing unit 150 sets the allocation flag 122 of this page entry 129 to “1”, the reference frequency 123 to “0”, and the change flag 124 to “0”. The virtual page number 125 of the stored virtual page is stored in this page entry 129. Subsequently, the swap processing unit 150 sends a request to transfer the swap-in target data stored in the virtual page to the physical page of the physical memory 200 corresponding to the physical page number 121 acquired previously. And the disk device interface unit 170.

  The memory management unit 130 and the disk device interface unit 170 that have input the transfer request control the physical memory 200 and the disk device 300, respectively, and transfer the swap-in target data. The transfer realizing means (not shown) may be routed through the memory management unit 130 and the disk device interface unit 170, or may be directly exchanged between the physical memory 200 and the disk device 300. Good.

  Next, the swap processing unit 150 receives a transfer completion notification from the memory management unit 130 and completes the swap-in process.

  Next, the swap-out process will be described. The swap processing unit 150 refers to the page entry 129 of the page table 120 including the physical page number 121 specified in response to the input of the physical page number 121 specifying the physical page of the physical memory 200 to be swapped, and changes it. Check the flag 124. When the change flag 124 is “0”, since the physical page of the physical memory 200 is a clean page, it is not necessary to write the data of this page from the physical memory 200 to the swap area 310. Therefore, the swap processing unit 150 sets the allocation flag 122 of the page entry 129 corresponding to the data to be swapped out to “0”, and completes the swap-out process.

  When the change flag 124 is “1”, the physical page of the physical memory 200 is a dirty page, and thus data of this page needs to be written from the physical memory 200 to the swap area 310. Therefore, the swap processing unit 150 refers to the page entry 129 of the page table 120 including the designated physical page number 121 and acquires the corresponding virtual page number 125. Subsequently, the swap processing unit 150 converts the data to be swapped out stored in the physical page of the physical memory 200 corresponding to the designated physical page number 121 into the virtual page corresponding to the virtual page number 125 acquired first. The transfer request is output to the memory management unit 130 and the disk device interface unit 170.

  The memory management unit 130 and the disk device interface unit 170 that have input the transfer request control the physical memory 200 and the disk device 300, respectively, and transfer data to be swapped out. The transfer realizing means (not shown) may be routed through the memory management unit 130 and the disk device interface unit 170, or may be directly exchanged between the physical memory 200 and the disk device 300. Good.

  Upon receiving a transfer completion notification from the disk device interface unit 170, the swap processing unit 150 sets the allocation flag 122 of the page entry 129 corresponding to the data to be swapped out to “0”, and completes the swap-out processing.

  Note that swap-in and swap-out may be generally called page-in and page-out, respectively.

  The swap area management unit 160 periodically refers to the page table 120 to detect a physical page in the physical memory 200 with the change flag 124 being “1” and the lowest reference frequency 123, and this detected physical page Are saved in the swap area 310.

  Next, the operation of the present embodiment will be described in detail with reference to the drawings.

  FIG. 5 is a flowchart showing an operation when a page fault is detected.

  First, the control unit 100 starts this operation when the virtual page to be accessed does not exist in the physical memory 200, that is, when a page fault is detected (S600). When the memory management unit 130 detects a page fault, the memory management unit 130 outputs a page fault occurrence notification to the page selection unit 140 (S602).

  Next, the page selection unit 140 acquires from the operating system (not shown) the task identifier 111 and the virtual page number 125 that become a page fault in response to the input of the page fault occurrence notification, and refers to the task management table 110 to refer to this The task priority 112 corresponding to the task that has become the page fault is acquired (S604).

  Next, the page selection unit 140 refers to the page table 120 and acquires the physical page number 121 with the lowest reference frequency 123 (S606).

  Next, the page selection unit 140 compares the task priority 112 acquired in S604 with the threshold 141 (S610). As a result of the comparison, if the task priority 112 acquired in S604 is lower than the threshold 141 (NO in S610), the process proceeds to S620. If the task priority 112 acquired in S604 exceeds the threshold 141 (YES in S610), the process proceeds to S614.

  Note that the comparison between the task priority 112 and the threshold 141 is based on the priority corresponding to the target data and a predetermined threshold whether the access target data is priority data or non-priority data. It is also included in the determination.

  In S614, the page selection unit 140 checks the state of the change flag 124 corresponding to the physical page number 121 acquired in S606 and S616 described later (S614). As a result of this confirmation, if the state of the change flag 124 is 0 (NO in S614), the process proceeds to S620. If the state of the change flag 124 is 1 (YES in S614), the process proceeds to S616.

  In S616, the page selection unit 140 refers to the page table 120 and refers to the page entry 129 except for the page entry 129 corresponding to the physical page number 121 already acquired in S606 or S616. The physical page number 121 corresponding to the page entry 129 having the lowest frequency 123 is acquired. If all the physical page numbers 121 are already acquired, the acquired state is canceled for all the physical page numbers 121, and the physical page number 121 with the smallest reference frequency 123 is selected from all the page entries 129. Obtain (S616). Then, the process returns to S614. When all the physical page numbers 121 are already acquired, the acquired state is canceled for all the physical page numbers 121, and the process of S614 and S616 is continued to perform a clean process described later. The page entry 129 in which the state of the change flag 124 is “0” can be detected by the conversion process.

  Next, the page selection unit 140 outputs the acquired physical page number 121 and virtual page number 125 to the swap processing unit 150 (S620).

  Next, in response to the input of the physical page number 121 and the virtual page number 125, the swap processing unit 150 sets the data stored in the physical page of the physical memory 200 corresponding to the physical page number 121 to the virtual page number 125. Swap out to the virtual page of the corresponding swap area 310 (S622). As described above, when the physical page of the physical memory 200 in which the data to be swapped out is stored is a clean page, compared with the case where this is a dirty page, the physical memory 200 to the disk device 300 is compared. The processing time is shortened by the fact that data transfer to is unnecessary. On the other hand, if the task priority 112 is higher than the predetermined threshold 141, the status of the change flag 124 is confirmed in S614, and the change flag 124 is “0”, that is, a clean page as the physical page number 121 of the physical memory 200 to be swapped. The one that is is selected. Accordingly, when the task priority 112 is higher than the predetermined threshold 141, swap-out processing with a reduced processing time is executed.

  Next, the swap processing unit 150 swaps in the virtual page data corresponding to the virtual page number 125 input in Step 622 to the physical page of the physical memory 200 swapped out in S622 (S624).

  FIG. 6 is a flowchart showing the operation of the cleaning process. In the cleaning process, the data of the dirty page (change flag 124 is “1”) in the physical memory 200 is transferred to the swap area 310 of the disk device 300, and the physical page in the target physical memory 200 is transferred to the clean page (change flag 124). Is “0”).

  First, the control unit 100 starts this operation in response to a cleaning process start instruction output at a predetermined time interval from a timing unit (not shown) (S700).

  The swap area management unit 160 refers to the page table 120 in response to the input of the cleaning process start instruction, and detects the page entry 129 having the lowest reference frequency 123 (S702).

  Next, the swap area management unit 160 checks the state of the change flag 124 of the detected page entry 129 (S704). As a result of this confirmation, if the state of the change flag 124 is “1” (YES in S704), the process proceeds to S710. If the state of the change flag 124 is “0” (NO in S704), the process proceeds to S706.

  In S706, the swap area management unit 160 checks whether all the page entries 129 in the page table 120 have already been detected in S702 or S708 described later (S706). As a result of this confirmation, if all page entries 129 have already been detected (YES in S706), the process ends. If there is an undetected page entry 129 (NO in S706), the process proceeds to S708. Note that all the page entries 129 in the page table 120 have already been detected in S702 or S708 described later mean that all physical pages are in a clean page state.

  In S708, the swap area management unit 160 refers to the page table 120, and detects the page entry 129 having the lowest reference frequency 123 from the page entries 129 excluding the page entry 129 already detected in S702 or S706. (S708). Then, the process returns to S704.

  Next, the swap area management unit 160 acquires the physical page number 121 and the virtual page number 125 of the page entry 129 that has been confirmed in S704 that the status of the change flag 124 is “1”. Subsequently, the swap area management unit 160 requests to transfer the data stored in the physical page to be cleaned corresponding to the acquired physical page number 121 to the virtual page corresponding to the acquired virtual page number 125. Is output to the memory management unit 130 and the disk device interface unit 170 (S710). The memory management unit 130 and the disk device interface unit 170 that have input the transfer request control the physical memory 200 and the disk device 300 to transfer the target data. The transfer realizing means (not shown) may be routed through the memory management unit 130 and the disk device interface unit 170, or may be directly exchanged between the physical memory 200 and the disk device 300. good.

  Upon receiving a transfer completion notification from the disk device interface unit 170, the swap area management unit 160 sets the change flag 124 of the page entry 129 corresponding to the page to be cleaned, to “0”, and completes the cleaning process. (S712).

  FIG. 2 shows a basic configuration of the present embodiment. Referring to FIG. 2, the basic configuration of the present invention includes a memory management device 400 including a page selection unit 140. The page selection unit 140 holds a predetermined threshold 141.

When a page fault occurs in a memory access to a hierarchical storage unit including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages, the page selection unit 140 Whether access target data is priority data or non-priority data is determined based on a predetermined priority corresponding to each target data and a threshold value 141, and the target data is priority data When the determination is made, the first data stored in the page of the first storage unit among the pages of the first storage unit is the same as the data stored in the corresponding page of the second storage unit Select one of the storage pages of
The identification information of the page of the selected first storage unit is output.

  The first effect of the present embodiment described above is that it is possible to speed up the swap operation when a page fault occurs in a process with a high priority, and this swap particularly for a task with a high priority. It is possible to speed up the processing of the operation.

  The reason is that, when a page fault occurs during access to the first storage unit, whether the access target data is priority data or non-priority data is determined in advance as the priority of the task that uses the target data. This is because the physical page number of the clean page is output when it is determined based on the determined threshold value and it is determined that the target data is priority data.

  The second effect of the present embodiment is that the swap operation when a page fault occurs in a high priority process can be speeded up. In particular, the search for the physical memory to be allocated is speeded up in the process of the swap operation. It is possible to reduce the operation of writing data already stored in the allocated physical memory to the swap area.

  The reason is that at each predetermined timing, dirty pages on the physical memory are detected in ascending order of reference frequency value, and the data stored in the dirty pages is written to the swap area to make a clean page. Because it was.

  Next, a second embodiment of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 7, the configuration of the second exemplary embodiment of the present invention is different from the first exemplary embodiment in that the control unit 100 includes an object management table 180 and a library management table 190.

  FIG. 8 shows the structure of the object management table 180. The object management table 180 is a table for managing the priority of an object, and includes at least one object entry 189 including an object identifier 181 that is an object identifier and a corresponding object priority 182. Any one or more of the object identifiers 181 may be a pseudo object identifier 181 provided for control by priority based on factors other than the object. The pseudo object identifier 181 indicates the priority based on a factor that determines that the priority detected by an operating system (not shown) or an application (not shown) should be increased or decreased. used.

  FIG. 9 is a diagram showing the structure of the library management table 190. The library management table 190 is a table for managing the library priority, and includes at least one library entry 199 including a library identifier 191 that is a library identifier and a corresponding library priority 192. Any one or more of the library identifiers 191 may be a pseudo library identifier 191 provided for control by priority based on factors other than the library. The pseudo object identifier 181 indicates the priority based on a factor that determines that the priority detected by an operating system (not shown) or an application (not shown) should be increased or decreased. used.

  In the second embodiment of the present invention, the operation of the page selection unit 140 is different from that of the first embodiment.

  The page selection unit 140 according to the present embodiment has a means for holding a predetermined threshold 141, and the task management table 110, the object management table 180, and the library management table 190 correspond to the input of the page fault occurrence notification. A virtual page to be swapped out is determined with reference to at least one of the pages and the page table 120.

  The flowchart showing the operation when a page fault is detected in the present embodiment is the same as that in the first embodiment shown in FIG. 5, but the operations in S604 and S610 are different.

  In S604 of the present embodiment, the page selection unit 140 acquires at least one of the task priority 112, the object priority 182 and the library priority 192 in response to the input of the page fault occurrence notification. The acquisition of the task priority 112 is as described in the first embodiment. The acquisition of the object priority 182 will be described. The page selection unit 140 acquires the object identifier 181 and the virtual page number 125 that have become page faults from an operating system (not shown), and refers to the object management table 180 to determine the page faults. The object priority 182 corresponding to the lost object is acquired. The acquisition of the library priority 192 will be described. The page selection unit 140 acquires a library identifier 191 and a virtual page number 125 that have become page faults from an operating system (not shown), and refers to the library management table 190 to determine the page faults. The library priority 192 corresponding to the library that has become is acquired.

  In S610 of this embodiment, the page selection unit 140 compares the threshold 141 with at least one of the task priority 112, the object priority 182 and the library priority 192 acquired in S604, resulting in a page fault. It is determined whether the data corresponding to the virtual page is priority data or non-priority data. For example, the page selection unit 140 may make a determination based on a result of selecting and comparing any one of the task priority 112, the object priority 182 and the library priority 192, or the task priority 112. Alternatively, it may be determined by combining the results of selecting and comparing any two or more of the object priority 182 and the library priority 192. Further, the page selection unit 140 compares, for example, an average value, a total value, or a weighted value calculated by performing operations on the task priority 112, the object priority 182 and the library priority 192 with the threshold 141. It may be a target. In addition, the page selection unit 140 may perform a logical sum, a logical product, or perform other operations, for example, when determining by combining the comparison results. May be. If it is determined that the data is non-priority data (NO in S610), the process proceeds to S620. If it is determined that the data is priority data (YES in S610), the process proceeds to S614.

  The flowchart showing the operation of the cleaning process in the present embodiment is the same as that in the first embodiment shown in FIG.

  The effect of the present embodiment described above is that the swap operation when a page fault occurs in a high priority process can be speeded up. In particular, the precise and easy selection of the target for speeding up the process of the swap operation is possible. It is a point that can be made possible.

  The reason for this is that when a page fault occurs during access to the first storage unit, whether the access target data is priority data or non-priority data is determined based on the priority and target of the object that is the target data. This is because the determination is made based on the priority of the library including the data.

  Each component described in each of the above embodiments may cause a computer to execute predetermined processing by a program, for example.

  Each component described in each of the above embodiments does not necessarily have to be individually independent, and a plurality of components are realized as one module, or one component is realized as a plurality of modules. Alternatively, a configuration in which a certain component is a part of another component, a part of a certain component overlaps with a part of another component, or the like may be employed.

  Further, in each of the embodiments described above, a plurality of operations are described in order in the form of a flowchart, but the description order does not limit the order in which the plurality of operations are executed. For this reason, when each embodiment is implemented, the order of the plurality of operations can be changed within a range that does not hinder the contents.

  Furthermore, in each embodiment described above, a plurality of operations are not limited to being executed at different timings. For example, another operation may occur during the execution of a certain operation, or part or all of the execution timing of a certain operation and the execution timing of another operation may overlap.

  Further, in each of the embodiments described above, it is described that a certain operation becomes a trigger for another operation, but the description does not limit all relationships between the certain operation and the other operations. For this reason, when each embodiment is implemented, the relationship between the plurality of operations can be changed within a range that does not hinder the contents. The specific description of each operation of each component does not limit each operation of each component. For this reason, each specific operation of each component may be changed within a range that does not hinder the functional, performance, and other characteristics in implementing each embodiment.

  In addition, each component in each embodiment described above may be realized by hardware, software, or a mixture of hardware and software, if necessary. May be.

  Further, the physical configuration of each component is not limited to the description in the above embodiment, and may exist independently, may exist in combination, or may be separated. It may be configured.

  The present invention can be applied to memory management of various information processing apparatuses and various communication processing apparatuses.

It is a block diagram which shows the structure of the 1st Embodiment of this invention. It is a block diagram which shows the basic composition in the 1st Embodiment of this invention. It is a figure which shows the structure of the task management table in the 1st and 2nd embodiment of this invention. It is a figure which shows the structure of the page table in the 1st and 2nd embodiment of this invention. It is a flowchart which shows the operation | movement at the time of detecting the page fault in the 1st and 2nd embodiment of this invention. It is a flowchart which shows the operation | movement of the cleaning process in the 1st and 2nd embodiment of this invention. It is a block diagram which shows the structure of the 2nd Embodiment of this invention. It is a figure which shows the structure of the object management table in the 2nd Embodiment of this invention. It is a figure which shows the structure of the library management table in the 2nd Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Control part 110 Task management table 111 Task identifier 112 Priority 119 Task entry 120 Page table 121 Physical page number 122 Allocation flag 123 Reference frequency 124 Change flag 125 Virtual page number 129 Page entry 130 Memory management part 140 Page selection part 141 Threshold 150 Swap processing unit 160 Swap area management unit 170 Disk device interface unit 180 Object management table 181 Object identifier 182 Priority 189 Object entry 190 Library management table 191 Library identifier 192 Priority 199 Library entry 200 Physical memory 300 Disk device 310 Swap area

Claims (6)

When a page fault occurs in a memory access to a hierarchical storage unit including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages,
Whether the target data of the memory access to the first storage unit is priority data or non-priority data, the priority of the task, the priority of the object, and the priority of the library that are pre-assigned to the target data When the determination is made based on any two or more and a predetermined threshold, and the target data is determined to be the priority data, the second of the plurality of pages of the first storage unit Select one of the pages storing the same data as the data stored in the corresponding page of the storage unit,
A memory management device comprising: a page selection unit that outputs identification information of the selected page of the first storage unit. For each page of the first storage unit, the correspondence between the data stored in the page and the second storage unit in ascending order of the reference frequency value corresponding to the page at each predetermined timing. A dirty page that is a page of the first storage unit that is different from the data stored in the page to be
The swap area management unit that outputs an instruction to write data stored in the dirty page to a corresponding page of the second storage unit when the dirty page is detected. Memory management device. When a page fault occurs in a memory access to a hierarchical storage unit including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages,
Whether the target data of the memory access to the first storage unit is priority data or non-priority data, the priority of the task, the priority of the object, and the priority of the library that are pre-assigned to the target data When the determination is made based on any two or more and a predetermined threshold, and the target data is determined to be the priority data, the second of the plurality of pages of the first storage unit Select one of the pages storing the same data as the data stored in the corresponding page of the storage unit,
Outputting identification information of the selected page of the first storage unit. A memory management method, comprising: For each page of the first storage unit, the correspondence between the data stored in the page and the second storage unit in ascending order of the reference frequency value corresponding to the page at each predetermined timing. A dirty page that is a page of the first storage unit that is different from the data stored in the page to be
Memory management method according to claim 3 Symbol mounting and outputs an instruction to write the data stored in the dirty pages when detecting the dirty pages to page corresponding to the second storage unit. When a page fault occurs in a memory access to a hierarchical storage unit including a first storage unit including a plurality of pages and a second storage unit including a plurality of pages,
Whether the target data of the memory access to the first storage unit is priority data or non-priority data, the priority of the task, the priority of the object, and the priority of the library that are pre-assigned to the target data When the determination is made based on any two or more and a predetermined threshold, and the target data is determined to be the priority data, the second of the plurality of pages of the first storage unit Select one of the pages storing the same data as the data stored in the corresponding page of the storage unit,
A program for causing a computer to execute a process of outputting identification information of the selected page of the first storage unit. For each page of the first storage unit, the correspondence between the data stored in the page and the second storage unit in ascending order of the reference frequency value corresponding to the page at each predetermined timing. A dirty page that is a page of the first storage unit that is different from the data stored in the page to be
5. Symbol mounting, characterized in that to execute a process of outputting an instruction to write the data stored in the dirty pages when detecting the dirty pages to the second corresponding page of the storage unit to the computer Program.

Images