KR20220091023A - Hybrid memory device and management method thereof - Google Patents

Abstract

The present invention divides a VM module including a plurality of volatile memory cells, an NVM module including a plurality of non-volatile memory cells, and a program stored in the NVM module into a predetermined segment unit, and is the number of write commands included in each segment. by acquiring and summing the segment counts to calculate a program count that is a write count for the program, selecting a segment whose segment count is equal to or greater than a predetermined second threshold among a plurality of segments of the program in which the program count is greater than or equal to a predetermined first threshold A hybrid memory device that can improve hybrid memory performance by reducing the number of NVM writes and improving write speed, including a memory control module that migrates data stored in the VM module and remaps the data addresses of migrated segments And it can provide a management method thereof.

Description

Hybrid memory device and management method thereof

The present invention relates to a hybrid memory device and a management method thereof, and to a hybrid memory device in consideration of write characteristics and a management method thereof.

With the development of information communication devices and computer graphics technology, the demand for high-speed, large-capacity and high-performance memory has always increased. Accordingly, research to improve memory performance, such as integration improvement and high-speed and low power consumption, has been continuously conducted, but recently, each type of memory has clear advantages and disadvantages, so it is difficult to satisfy all requirements with only a single type of memory.

For example, volatile memory (VM) such as dynamic random access memory (DRAM) has the advantage of fast write speed, but data is lost when power is not supplied. There is a problem in that power consumption increases because it has to perform poetry.

On the other hand, in the case of non-volatile memory (NVM) such as phase-change memory (PCM), it has a higher density than volatile memory, so it is advantageous to implement with a high capacity, and power is not applied. It has the advantage that it can be implemented as a low-power device because it does not consume data and does not consume standby power. However, in the case of NVM, the read speed is similar to that of the VM, but the write speed is significantly slower than that of the VM, which is highly likely to cause a bottleneck. There is a problem in that it is not suitable for application to applications that occur frequently.

Accordingly, recently, research on a hybrid memory including two or more different types of memory devices has been actively conducted. In the case of hybrid memory, it is mostly composed of a combination of VM and NVM, so that the shortcomings of VM and NVM can be complemented with each other.

In other words, hybrid memory takes advantage of NVM's advantages of non-volatility, high density, and low power consumption, but focuses on solving problems caused by slow write speed and limited number of writes by using VM.

Korean Patent Registration No. 10-1713051 (Registered on February 28, 2017)

It is an object of the present invention to provide a hybrid memory device capable of improving performance of a hybrid memory by reducing the number of writes of the NVM and improving the write speed in consideration of the write characteristics of the VM and the NVM, and a method for managing the same.

Another object of the present invention is to divide a program into segments, perform migration by selecting a migration target segment based on the number of write commands in the segmented segment, and adjust the size of a migration unit based on a change in the data address of the migration target segment Accordingly, an object of the present invention is to provide a hybrid memory device capable of improving write efficiency and a method for managing the same.

A hybrid memory device according to an embodiment of the present invention for achieving the above object includes: a VM module including a plurality of volatile memory cells; an NVM module including a plurality of non-volatile memory cells; and divides the program stored in the NVM module into a predetermined segment unit, obtains and adds a segment count that is the number of write commands included in each segment, calculates a program count that is a write count for the program, and the program count is A memory control module that selects a segment having a segment count equal to or greater than a second threshold value from among a plurality of segments of a program that is equal to or greater than a predetermined first threshold value, migrates it with data stored in the VM module, and remaps a data address of the migrated segment includes

The memory control module may include: a VM controller configured to control access to data stored in the VM module; an NVM controller for controlling access to data stored in the NVM module; and a memory manager that selectively migrates a segment of a program stored in the NVM module with data stored in the VM module, and modifies it with an access command to a remapped data address when an access command for the migrated data is applied. can

The memory manager may include: a command analyzing unit that, when an access command including a read or write command is applied, analyzes the applied access command, and divides the read or write command and write request information including program information and data address; After receiving the write request information, the approved program is divided into a plurality of segments, the segment count and the program count are obtained and stored in the write history table, and the segment of the program stored in the NVM module is selected based on the stored write history table a reallocation unit for migrating the corresponding data of the VM module to , and obtaining a migrated and changed data address; The data address in which segment data is stored and the changed data address are remapped and stored, and when the read or write command is applied, the data address specified in the applied read or write command is modified to the remapped data address and the VM controller and a remapping unit outputting to the NVM controller; and a swap buffer for receiving data migrated between the NVM module and the VM module, temporarily storing the data, and transferring the data to a counterpart module.

The reallocation unit analyzes the write history table, sets a program having a program count equal to or greater than the first threshold value as a swap candidate program, and selects a segment having a segment count equal to or greater than the second threshold value among a plurality of segments of the set swap candidate program. It can be selected as the migration target segment where the migration is performed.

The reallocation unit transfers the data address for the migration target segment to the NVM controller, so that the migration target segment stored in the NVM module is transferred to the VM module through the swap buffer and the VM controller under the control of the NVM controller. can be saved.

The reassignment unit checks the data address of the migrated segment, and if the data address difference of the sequentially migrated data is less than or equal to a predetermined reference address unit, the reassignment unit counts the accumulated data in the same data scope among the predetermined number of data scopes, and the reference address unit If it exceeds, it is counted again in another data scope, and the address change of sequentially transmitted data is accumulated and recorded in a number of predefined data scopes. It is possible to adjust the migration unit size of migrated data according to the confirmed section by checking whether the recorded size corresponds to any of the unit maintenance section, the unit reduction section, and the unit increase section, which are predetermined based on the currently set migration unit.

When an access command is applied during migration, the memory manager may pause migration and perform an operation corresponding to the access command after completion of transmission of data transmitted to a swap buffer among data of a segment to be migrated.

The memory manager determines whether the data designated by the access command is stored in the swap buffer, and when it is determined that the data is stored in the swap buffer, accesses the data stored in the swap buffer, If it is determined that there is not, access to data stored in the VM module or the NVM module may be performed through the VM controller or the NVM controller.

A hybrid memory management method according to another embodiment of the present invention for achieving the above object is a hybrid memory including a VM module including a plurality of volatile memory cells, an NVM module including a plurality of non-volatile memory cells, and a swap buffer. A device management method, wherein a program stored in the NVM module is divided into a predetermined segment unit, and a segment count that is the number of write commands included in each segment is acquired and summed to calculate a program count that is a write count for the program to do; selecting a segment having a segment count equal to or greater than a second threshold value from among a plurality of segments of the program in which the program count is equal to or greater than a predetermined first threshold value, and migrating it with data stored in the VM module; and remapping the data addresses of the migrated segments.

Accordingly, the hybrid memory device and the management method thereof according to the embodiment of the present invention can improve the performance of the hybrid memory by reducing the number of writes of the NVM and improving the write speed in consideration of the write characteristics of the VM and the NVM. In addition, programs are divided into segments, migration is performed by selecting a migration target segment based on the number of write commands in the segment, and the migration unit size is adjusted based on the change in the data address of the migration target segment to improve write efficiency. can do it

1 shows a schematic structure of a hybrid memory device according to an embodiment of the present invention.
2 is a diagram for explaining a segment count and a program count.
3 to 6 are diagrams for explaining an operation of migrating data by the memory control module of FIG. 1 .
FIG. 7 shows a detailed structure of the memory manager of FIG. 1 .
FIG. 8 shows a detailed structure of the reassignment unit of FIG. 7 .
FIG. 9 is a diagram for explaining a concept in which the migration determining unit of FIG. 8 determines a unit size of migrated data.
10 illustrates a method for managing a hybrid memory device according to an embodiment of the present invention.
11 is a diagram illustrating in detail the migration execution step of FIG. 10 .

In order to fully understand the present invention, the operational advantages of the present invention, and the objects achieved by the practice of the present invention, reference should be made to the accompanying drawings illustrating preferred embodiments of the present invention and the contents described in the accompanying drawings.

Hereinafter, the present invention will be described in detail by describing preferred embodiments of the present invention with reference to the accompanying drawings. However, the present invention may be embodied in various different forms, and is not limited to the described embodiments. In addition, in order to clearly explain the present invention, parts irrelevant to the description are omitted, and the same reference numerals in the drawings indicate the same members.

Throughout the specification, when a part "includes" a certain component, it does not exclude other components, unless otherwise stated, meaning that other components may be further included. In addition, terms such as "...unit", "...group", "module", and "block" described in the specification mean a unit that processes at least one function or operation, which is hardware, software, or hardware. and a combination of software.

1 shows a schematic structure of a hybrid memory device according to an embodiment of the present invention.

Referring to FIG. 1 , the hybrid memory device according to the present embodiment may include a memory control module 100 , a VM module 200 , and an NVM module 300 .

When a read command is applied from an external device such as a processor such as a CPU or GPU, the memory control module 100 acquires data corresponding to the read command among data stored in the VM module 200 or the NVM module 300 to obtain an external device. output as And when a write command and data are applied from an external device, the data applied along with the write command is stored in the VM module 200 or the NVM module 300 .

Also, in the present embodiment, the memory control module 100 manages data stored in the VM module 200 and the NVM module 300 . The memory control module 100 may perform migration for exchanging data between the NVM module 300 and the VM module 200 . The memory control module 100 selects data stored in the NVM module 300 to migrate to the VM module 200 and swaps corresponding data stored in the VM module 200 to the NVM module 300 to migrate can do it In the present embodiment, the memory control module 100 may analyze data stored in the NVM module 300 to select and migrate data to be stored in the VM module 200 based on the number of write commands.

The memory control module 100 may include a memory manager 110 , a VM controller 120 , and an NVM controller 130 .

The memory manager 110 determines a module in which data corresponding to a read command applied from the outside is stored among the VM module 200 or the NVM module 300 , and when it is determined that the data is stored in the VM module 200 , the VM controller A read command is transmitted to 120 , and when it is determined that data is stored in the NVM module 300 , a read command is transmitted to the NVM controller 130 . Then, the VM controller 120 or the VM controller 120 receives the corresponding data and outputs it to the outside.

Also, the memory manager 110 transmits the write command and data to be stored to the VM controller 120 or the VM controller 120 according to the write command, so that the data is stored in the VM module 200 or the NVM module 300 . . Here, the data applied along with the write command is basically stored in the NVM module 300, but if the data to be written is previously migrated from the NVM module 300 to the VM module 200 and stored, the VM module 200 It can be updated by overwriting the data stored in the .

And in this embodiment, the memory manager 110 selects data to be migrated to the VM module 200 from among the data stored in the NVM module 300 , and displays the selection result to the VM controller 120 and the NVM controller 130 . to each of the NVM module 300 and the VM module 200 to mutually swap selected data and corresponding data.

The memory manager 110 is a swap buffer (not shown) for temporarily storing data transferred from the VM controller 120 and the NVM controller 130 so that data migration can be performed between the VM module 200 and the NVM module 300 . city) may be included. A detailed operation of performing migration by the memory manager 110 will be described later.

The VM controller 120 controls the VM module 200 according to the management of the memory manager 110 . The VM controller 120 searches for and acquires data stored in the VM module 200 in response to a read command applied through the memory manager 110 , and transmits the acquired data to the memory manager 110 . And in response to the write command, the applied data is stored in a predetermined memory address of the VM module 200 . In this case, the VM controller 120 may include a buffer 121 for temporarily storing data obtained from the VM module 200 or data transmitted from the memory manager 110 to be stored in the VM module 200 .

In addition, the VM controller 120 refreshes the data stored in the VM module 200 at a predetermined cycle to prevent the data stored in the VM module 200 from being lost, and is applied as a hybrid memory device under the management of the memory manager 110 . Data stored in the VM module 200 may be received before the power is turned off and transferred to the NVM controller 130 through the swap buffer of the memory manager 110 .

The NVM controller 130 controls the NVM module 300 according to the management of the memory manager 110 . Similar to the VM controller 120 , the NVM controller 130 searches for and acquires data stored in the NVM module 300 in response to a read command applied through the memory manager 110 , and transfers the acquired data to the memory manager 110 . ), and stores data applied to a predetermined memory address of the NVM module 300 in response to a write command. In addition, the NVM controller 130 may also include the NVM buffer 131 to temporarily store data acquired from the NVM module 300 or data transmitted from the memory manager 110 to be stored in the NVM module 300 .

Unlike the VM controller 120 , the NVM controller 130 does not need to perform a refresh operation on data stored in the NVM module 300 . However, the NVM controller 130 may receive data stored in the NVM module 300 selected by the memory manager 110 and transmit it to the VM controller 120 through the swap buffer.

The VM module 200 includes a plurality of VM cells and stores data while power is applied. The VM module 200 receives and stores or refreshes data under the control of the VM controller 120 of the memory control module 100 , and transmits the stored data to the VM controller 120 .

The NVM module 300 may include a plurality of NVM cells to always store data regardless of whether power is supplied or not. The NVM module 300 receives and stores data under the control of the NVM controller 130 of the memory control module 100 , and transmits the stored data to the NVM controller 130 .

In the present embodiment, the memory control module 100 classifies and manages data for a program (or application) in segments. Here, a segment represents a data unit selected for migration. That is, as shown in FIG. 1 , a plurality of programs (Program 0 to Program 3) each consisting of a plurality of segments (Segment 0 to Segment 3) may be stored in the NVM module 300 , and the memory control module 100 ) may be migrated to the VM module 200 by selecting segments from each of a plurality of programs (Program 0 to Program 3) stored in the NVM module 300 . In FIG. 1 , as an example, in the NVM module 300 , one segment ((segment 3), (segment 0)) in the 0th and first programs (Program 0, Program 1) and 2 in the second program (Program 2) respectively It is assumed that the segment 1 and segment 2 are selected and migrated to the VM module 200 .

FIG. 2 is a diagram for explaining a segment count and a program count, and FIGS. 3 to 6 are diagrams for explaining an operation of migrating data by the memory control module of FIG. 1 .

The memory manager 110 of the memory control module 100 shown in FIG. 1 counts the number of write commands included in each segment of the programs (Program A and Program C) stored in the NVM module 300 . At this time, when a program to be written (Program A, Program C) is applied from the outside, the memory manager 110 stores the applied program in the NVM module 300 while counting the number of write commands in units of each segment, and the program is If stored, the program count may be obtained by summing the counted segment counts. Here, the number of write commands counted for each segment is referred to as a segment count. Then, the accumulated sum of segment counts for a plurality of segments included in each program data (Program A, Program C) is calculated as the program count.

Referring to FIG. 2 , in program A, segment counts for four segments are (51, 20, 32, 24), respectively, and the program count is calculated as 127. And, in the C program (Program C), the segment count for the four segments is (49, 38, 77, 47), respectively, and the program count is calculated as 211.

When data is changed in a program stored in the NVM module 300 , that is, when a change occurs in a segment, the memory manager 110 may update the segment count according to the changed segment and reflect this in the program count.

FIG. 3 shows an initial state of the NVM module 300 , in which programs (Program A to Program C) are not substantially stored in the NVM module 300 , indicating a state in which all program counts are 0.

4 shows a state in which programs (Program A to Program C) are stored in the NVM module 300, the memory manager 110 checks the number of writes for each of a plurality of segments for each program (Program A to Program C). A count is acquired, and a program count is calculated based on the acquired segment count. Here, it is assumed that the program counts for each program (Program A to Program C) are 127, 79, and 211, respectively.

When the program count is calculated, the memory manager 110 searches for a program having a program count equal to or greater than a predetermined first threshold (here, 100 for example) among the calculated program counts. When a program (Program A, Program C) whose program count is equal to or greater than the first threshold is searched for, the searched programs (Program A, Program C) are selected as a swap candidate program.

Then, a segment in which the segment count of each segment included in the selected swap candidate programs (Program A, Program C) is equal to or less than the first threshold value is greater than or equal to a second threshold (here, 50 for example) is searched for, and the segment When a segment whose count is equal to or greater than the second threshold is found, the found segment is selected as a migration target.

In FIG. 5, it is assumed that the segment count of each segment in the swap candidate programs (Program A, Program C) is the same as in FIG. , in the C program (Program C), the second segment having a segment count of 77 is selected as a migration target. Here, as an example, a case in which one segment is selected as a migration target in the programs (Program A, Program C) is illustrated, but the number of segments selected as a migration target in each program (Program A, Program C) is not limited. . That is, all segments having a segment count equal to or greater than the second threshold in the selected swap candidate programs Program A and Program C may be selected as migration target segments. However, in the program (Program B) that is not selected as the swap candidate programs (Program A, Program C), even if all segments having the segment count equal to or greater than the second threshold exist, the memory manager 110 determines the unselected program (Program B). Since the segment count is not considered, it is not selected as a migration target segment.

In addition, when a segment to be migrated is selected, the memory manager 110 performs the migration between the VM controller 120 and the NVM controller 130 so that the segment selected in the NVM module 300 is migrated to the VM module 200 as shown in FIG. 6 . ) to migrate the migration target segment selected by the NVM module 300 to the VM module 200 . At this time, the migration target segment selected by the NVM module 300 is transferred to the swap buffer of the memory manager 110 through the buffer 131 of the NVM controller 130, and the migration target segment transferred to the swap buffer is transferred to the VM controller ( It may be applied to and stored in the VM module 200 through the buffer 121 of the 120 .

Also, the memory manager 110 may change the unit size of the migrated data when the selected migration target segment is migrated.

FIG. 7 shows a detailed structure of the memory manager of FIG. 1 .

Referring to FIG. 7 , the memory manager 110 according to the present embodiment may include a command analysis unit 111 , a reallocation unit 112 , a remapping unit 113 , and a swap buffer 114 .

The command analysis unit 111 analyzes a command such as read or write applied from an external device, and transmits the read request and the write request to the remapping unit 113 , and transmits the write request information to the reassignment unit 112 . . Here, the write request information may include a data address and program information.

A program stored in the hybrid memory device may be executed by an external processor or the like, and data change occurs by a write command during execution, and the changed data must be stored in the hybrid device again. Accordingly, when the data changed by the write command is applied, the command analysis unit 111 transmits write request information including the address of the changed data and program information and segment information to the reassignment unit 112, so that the previously stored data is changed. to be updated with data. In this embodiment, the command analysis unit 111 transmits the write request information to the reassignment unit 112 to the segment migrated to the VM module 200 by the program count and segment count in the program stored in the NVM module 300. This is so that information about the data is transmitted to the remapping unit 113 so that the remapping unit 113 can write and update data normally.

If the applied program is a new program that is not stored in the previous VM module 200 or the NVM module 300, the reassignment unit 112 divides the applied program into segments and counts the number of each segment including a write command. to obtain a segment count, and calculate and store the program count based on the obtained segment count. In addition, the reassignment unit 112 sets a location in which a plurality of segments of a program are to be stored among the VM module 200 or the NVM module 300 according to the calculated program count and segment count.

The reassignment unit 112 may allocate addresses so that segments of the program to be applied are basically stored in the NVM module 300 . However, the reallocation unit 112 selects a swap candidate program based on the stored program count and the segment count, selects a migration target segment from the selected swap candidate program, and stores the selected migration target segment in the VM module 200 . Addresses can be assigned as much as possible.

That is, addresses may be assigned to the segments of the unselected program and the segments other than the migration target segment in the swap candidate program to be stored in the NVM module 300 . In some cases, the reassignment unit 112 allocates and stores addresses so that all segments for the new program are stored in the NVM module 300 , and then selects a migration target segment from among the program segments stored in the NVM module 300 . It may be migrated to the VM module 200 .

After the VM controller 120 or NVM controller 130 stores the approved program segment, the reallocation unit 112 may obtain the stored data address by being authorized by the VM controller 120 or the NVM controller 130 . .

In addition, the reassignment unit 112 analyzes the program count and segment count for the program previously stored in the NVM module 300 to selectively migrate segments of the program stored in the NVM module 300 to the VM module 300 . have. As described above, when a new program is written, the reassignment unit 112 obtains and stores the program count and segment count in advance. Recalculate, update, and save. That is, the write history of segments of each program is stored in advance. Then, a migration target segment is selected from among the program segments according to the stored write history and migrated to the VM module 300 . At this time, the reallocation unit 112 transfers the data address of the segment to be migrated from the NVM module 300 and the data address of the segment migrated to the VM module 200 to the remapping unit 113 so that the remapping unit 113 is Changes in the migrated data address are saved.

Also, the reallocation unit 112 may analyze the data address of the migration target segment to adjust the unit size of the migrated data. A detailed description of how the reallocation unit 112 adjusts the unit size of migrated data will be described later.

The remapping unit 113 receives and stores the data address changed by migration from the reassignment unit 112, and when a read command or a write command is applied from the command analysis unit 111, a read command and a read command based on the stored data address The write command is modified and transmitted to the VM controller 120 or the NVM controller 130 .

When the data address for the migrated data is applied by the reallocation unit 112 , the remapping unit 113 stores the existing data address and the migrated data address in the remapping table. Thereafter, when a read or write command is applied from the command analysis unit 111 , it is determined whether a memory address reallocated for the corresponding data exists in the remapping table, and the memory address reallocated by the reallocation unit 112 exists. Then, the VM controller 120 or the NVM controller 130 modifies the read or write command into a command for the reallocated memory address and transmits it to the VM controller 120 or the NVM controller 130, thereby causing the reallocated memory address Enables data to be read from or written to.

The swap buffer 114 migrates data stored in the VM module 200 to the NVM module 300 or migrates data stored in the NVM module 300 to the VM module 200, the VM controller 120 and NVM Data transferred from the buffers 121 and 131 of the controller 130 is temporarily stored and then transferred to the counterpart buffers 121 and 131 to prevent data collisions during migration.

FIG. 8 shows the detailed structure of the reassignment unit of FIG. 7 , and FIG. 9 is a diagram for explaining the concept of determining the unit size of the migrated data by the migration determining unit of FIG. 8 .

Referring to FIG. 8 , the reallocation unit 112 may include a write request information analyzer 1121 , a write history table 1122 , a migration determiner 1123 , and a migration request unit 1124 .

The write request information analysis unit 1121 analyzes the write request information applied from the command analysis unit 111 . The write request information analyzer 1121 checks the program information, segment information, and data address from the write request information, and transmits it to the write history table 1122 .

The write history table 1122 divides a program into segments from the program information and segment information checked by the write request information analysis unit 1121, and counts the number of write commands included in each segment to obtain a segment count, and obtain Calculate the program count based on the segment count. Then, the calculated program count and segment count are stored in the write history table 1122 .

The migration determiner 1123 sets a swap candidate program from the program count stored in the write history table 1122 and selects a migration target segment from the segment counts of segments included in the set swap candidate program.

The migration request unit 1124 transmits the data address of the migration target segment to the VM controller 120 and the NVM controller 130 so that the migration target segment is migrated to the VM module 200 or the NVM module 300 .

The migration request unit 1124 may receive the migrated data address from the VM controller 120 and the NVM controller 130 , store it in the write history table 1122 , and transmit it to the remapping unit 113 . In this case, the migration request unit 1124 may adjust the unit size of the migrated data based on the data address of the data included in the segment.

Referring to FIG. 9 , the migration request unit 1124 may include a plurality of data scopes Scope0 to Scope5. Here, the number of data scopes Scope0 to Scope5 may be variously adjusted. Here, the data scope is a kind of virtual counter to determine whether data is sequentially applied. indicates the count.

Here, the total size of each of the plurality of data scopes (Scope0 to Scope5) may be set to indicate the maximum migrateable data migration unit size (eg, 128Kbyte). For example, in FIG. 9 , each area of each data scope (Scope0 to Scope5) is set to represent a minimum data migration unit size of 4Kbye, and each of the plurality of data scopes (Scope0 to Scope5) may be divided into 32 areas. However, since the migration unit size is generally changed in the order of a power of 2, each region may be set to represent a migration unit size that increases from 4Kbye to a power of two. That is, each area of the data scope is sequentially 4, 8, 16, 32 ... , it may be set to indicate the order of 128 Kbytes.

Also, the migration request unit 1124 may set a first reference migration unit having a smaller value and a second reference migration unit having a larger value based on the currently set migration unit. That is, the section between the first reference migration unit and the second reference migration unit is set as the unit maintenance section, centered on the current migration unit, the section less than the first standard migration unit is set as the unit reduction section, and more than the first standard migration unit Set the section to the unit increment section.

Thereafter, if the address difference between sequentially accessed data addresses is less than or equal to the currently set preset reference address unit, the migration request unit 1124 sequentially performs each area in one data scope among a plurality of data scopes Scope0 to Scope5. fill If the data address difference continues to be less than or equal to the reference address unit, the same data scope continues to fill the region, whereas if the data address difference exceeds the reference address unit, each region is sequentially filled again in the next data scope.

The migration request unit 1124 repeats the above process, and when each area for all data scopes Scope0 to Scope5 is filled, it can cycle again to sequentially fill the data scopes Scope0 to Scope5.

In addition, the migration request unit 1124 calculates the size of the area filled in a plurality of data scopes (Scope0 to Scope5) during a predetermined period, and is more than half of the plurality of data scopes (Scope0 to Scope5), that is, more than half of the data scopes. A section in which the size of the filled region is included among the unit maintenance section, the unit reduction section, and the unit increase section is determined. If it is determined to be included in the unit maintenance section, the currently set migration unit is maintained, whereas if it is determined to be included in the unit reduction section or the unit increase section, the set migration unit is increased or decreased by a predetermined unit.

Here, when the migration request unit 1124 varies the migration unit, it increases or decreases the size of data to be migrated at a time in response to the size of continuous data in the segment to be migrated, thereby reducing the number of times of migration itself as well as the VM module ( 200) and the NVM module 300 to reduce the number of writes to improve the write speed as well as to improve the durability of the NVM module 300.

10 is a diagram illustrating a management method of a hybrid memory device according to an embodiment of the present invention, and FIG. 11 is a diagram illustrating the migration execution step of FIG. 10 in detail.

Referring to FIG. 10 , in the method of managing a hybrid memory device, it is first determined whether migration should be performed on data of a currently stored program ( S10 ). Here, the migration may be performed to swap data stored in the NVM module 300 with data stored in the VM module 200 . When the data stored in the NVM module 300 is migrated with the VM module 200 , after the program stored in the NVM module 300 is read, the read program may be executed. In addition, migration may be performed even when a new program is applied to and stored in the hybrid memory device by a write command. And when it is determined that migration is to be performed, the hybrid memory device performs migration in a predetermined manner (S20).

Referring to FIG. 11 , in the migration step, first, the number of write commands in each of a plurality of segments constituting each of a plurality of programs is counted to obtain a segment count, and the obtained segment counts are summed to obtain a program for each program A count is calculated (S21).

Then, a program whose calculated program count is equal to or greater than a predetermined first threshold is determined (S22). If a program whose program count is equal to or greater than the first threshold is identified, the checked program is set as a swap candidate program (S23). Thereafter, among a plurality of segments of the set swap candidate program, a segment having a segment count equal to or greater than a second threshold is determined (S24). When a segment having a segment count equal to or greater than the second threshold is determined, the determined segment is selected as a migration target segment ( S25 ).

Then, migration of the migration target segment is started (S26). When migration starts, the address difference of data sequentially transferred from the segment is checked, and if the identified address difference is less than or equal to a predetermined reference address unit, it is accumulated in the same data scope, and when it exceeds the reference address unit, it is accumulated in another data scope In this way, address changes of sequentially transferred data are accumulated and recorded in a predetermined number of data scopes. In addition, the size accumulated in multiple data scopes is checked for each predetermined section, and the size accumulated in more than half of the data scopes among the multiple data scopes is a preset unit maintenance section and unit reduction section based on the currently set migration unit, and The migration unit is changed by checking which section of the unit increase section corresponds to (S27).

Referring back to FIG. 10 , it is determined whether a data access command such as a read or write command is applied ( S30 ). If it is determined that the access command is approved, it is determined whether data migration is currently being performed between the VM module 200 and the NVM module 300 ( S40 ). If it is determined that migration is being performed, the migration is paused after waiting until the data currently being migrated is stored in the swap buffer 114 (S50).

Then, it is determined whether the data to which the access command is applied is stored in the swap buffer 114 (S60). When it is determined that data is stored in the swap buffer 114, access is performed to the swap buffer 114 (S70). That is, data is read or written. And the migration is continued (S80).

However, if it is determined that it is not stored in the swap buffer 114 , the data stored in the VM module 200 or the NVM module 300 is accessed through the VM controller 120 or the NVM controller 130 ( S90 ). Also, even when migration is not currently being performed, data stored in the VM module 200 or the NVM module 300 is accessed through the VM controller 120 or the NVM controller 130 ( S90 ).

The method according to the present invention may be implemented as a computer program stored in a medium for execution by a computer. Here, the computer-readable medium may be any available medium that can be accessed by a computer, and may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, and read dedicated memory), RAM (Random Access Memory), CD (Compact Disk)-ROM, DVD (Digital Video Disk)-ROM, magnetic tape, floppy disk, optical data storage, and the like.

Although the present invention has been described with reference to the embodiment shown in the drawings, which is only exemplary, those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom.

Accordingly, the true technical protection scope of the present invention should be defined by the technical spirit of the appended claims.

100: memory control module 200: VM module
300: NVM module 110: memory manager
120: VM Controller 121: VM Buffer
130: NVM Controller 131: NVM Buffer
111: command analysis unit 112: reassignment unit
113: remapping unit 114: swap buffer
1121: light request information analysis unit 1122: light history table
1123: migration decision unit 1124: migration request unit

Claims (17)

a VM module including a plurality of volatile memory cells;
an NVM module including a plurality of non-volatile memory cells; and
The program stored in the NVM module is divided into a predetermined segment unit, a segment count that is the number of write commands included in each segment is acquired and summed to calculate a program count that is a write count for the program, and the program count is A memory control module that selects a segment whose segment count is equal to or greater than a predetermined second threshold value among a plurality of segments of the program that is equal to or greater than a specified first threshold, migrates it with data stored in the VM module, and remaps the data address of the migrated segment; A hybrid memory device comprising The method of claim 1 , wherein the memory control module comprises:
a VM controller for controlling access to data stored in the VM module;
an NVM controller for controlling access to data stored in the NVM module; and
A hybrid comprising a memory manager that selectively migrates a segment of a program stored in the NVM module with data stored in the VM module, and modifies it with an access command to a remapped data address when an access command for the migrated data is applied memory device. The method of claim 2 , wherein the memory manager
a command analysis unit that, when an access command including a read or write command is applied, analyzes the applied access command, and divides the read or write command into write request information including program information and data address;
After receiving the write request information, the approved program is divided into a plurality of segments, the segment count and the program count are obtained and stored in the write history table, and the segment of the program stored in the NVM module is selected based on the stored write history table a reassignment unit for migrating the data corresponding to the VM module and acquiring the migrated and changed data address;
The data address in which segment data is stored and the changed data address are remapped and stored, and when the read or write command is applied, the data address specified in the applied read or write command is modified to the remapped data address and the VM controller and a remapping unit outputting to the NVM controller; and
and a swap buffer for receiving data migrated between the NVM module and the VM module, temporarily storing the data, and transferring the data to a counterpart module. The method of claim 3, wherein the reallocation unit
By analyzing the write history table, a program having a program count equal to or greater than the first threshold value is set as a swap candidate program, and a segment having a segment count equal to or greater than the second threshold value among a plurality of segments of the set swap candidate program is migrated. A hybrid memory device that you select as a migration target segment. 5. The method of claim 4, wherein the reallocation unit
A hybrid that transmits a data address for the migration target segment to the NVM controller so that the migration target segment stored in the NVM module is stored in the VM module through the swap buffer and the VM controller under the control of the NVM controller memory device. The method of claim 5, wherein the reallocation unit
Check the data address of the migrated segment. If the data address difference of sequentially migrated data is less than or equal to a predetermined reference address unit, the accumulated count is counted in the same data scope among the predetermined number of data scopes. If the reference address unit is exceeded, another Accumulated counting again in the data scope, the address change of sequentially transmitted data is accumulated and recorded in a number of predefined data scopes, the size accumulated and recorded in the plurality of data scopes is checked, and the size accumulated and recorded in more than half of the data scopes is A hybrid memory device that adjusts the migration unit size of migrated data according to the confirmed section by checking which section corresponds to a predetermined unit maintenance section, a unit reduction section, and a unit increase section based on the currently set migration unit. 4. The method of claim 3, wherein the memory manager
A hybrid memory device that, when an access command is applied during migration, pauses migration and performs an operation corresponding to the access command after the transfer of data transferred to the swap buffer among data of the migrated segment is completed. The method of claim 7, wherein the memory manager
It is determined whether the data specified by the access command is stored in the swap buffer, and when it is determined that the data is stored in the swap buffer, access is performed to the data stored in the swap buffer, and when it is determined that the data is not stored in the swap buffer , a hybrid memory device that accesses data stored in the VM module or the NVM module through the VM controller or the NVM controller. The method of claim 7, wherein the memory manager
A hybrid memory device that continues a paused migration once an access is made. A method for managing a hybrid memory device comprising: a VM module including a plurality of volatile memory cells; an NVM module including a plurality of non-volatile memory cells; and a swap buffer, the method comprising:
dividing the program stored in the NVM module into a predetermined segment unit, acquiring a segment count that is the number of write commands included in each segment, and summing them to calculate a program count that is a write count for the program;
selecting a segment having a segment count equal to or greater than a second threshold value from among a plurality of segments of the program in which the program count is equal to or greater than a predetermined first threshold value, and migrating the segment with data stored in the VM module; and
A method for managing a hybrid memory device, comprising remapping data addresses of migrated segments. 11. The method of claim 10, wherein calculating the program count comprises:
when an access command including a read or write command is applied, analyzing the applied access command and dividing the read or write command into write request information including program information and data address; and
and dividing the approved program into a plurality of segments by receiving the write request information, and storing the segment count and the program count in a write history table. The method of claim 11, wherein the migrating
analyzing the write history table and setting a program having a program count equal to or greater than the first threshold value as a swap candidate program; and
A hybrid memory device comprising: selecting a segment having a segment count equal to or greater than the second threshold value among a plurality of segments of a set swap candidate program as a migration target segment to be migrated and migrating the selected segment with corresponding data of the VM module management method. The method of claim 12, wherein the remapping comprises:
acquiring a migrated and changed data address; and
A method for managing a hybrid memory device, comprising the step of remapping and storing a data address in which segment data is stored and the changed data address. The method of claim 13 , wherein the managing method of the hybrid memory device comprises:
When an access command for migrated data is applied, the method of managing a hybrid memory device further comprising the step of modifying the remapped data address with an access command. 15. The method of claim 14, wherein the migrating step
A method for managing a hybrid memory device, comprising accessing the migration target segment stored in the NVM module according to a data address for the migration target segment, and storing the accessed migration target segment in the VM module through the swap buffer. 16. The method of claim 15, wherein the migrating step
Check the data address of the migrated segment. If the data address difference of sequentially migrated data is less than or equal to a predetermined reference address unit, the accumulated count is counted in the same data scope among the predetermined number of data scopes. If the reference address unit is exceeded, another accumulatively counting again in the data scope and accumulating and recording address changes of sequentially transferred data in a plurality of predetermined data scopes; and
Check the accumulated size recorded in multiple data scopes, and check whether the accumulated recorded size in more than half of the data scopes corresponds to the preset unit maintenance section, unit decrease section, and unit increase section based on the currently set migration unit. The management method of a hybrid memory device further comprising the step of adjusting the size of the migration unit of the data to be migrated according to the identified section. The method of claim 16 , wherein the managing method of the hybrid memory device comprises:
when an access command is applied during migration execution, after completion of transmission of data transmitted to a swap buffer among data of a segment to be migrated, stopping migration and performing an operation corresponding to the access command;
determining whether the data specified by the access command is stored in the swap buffer;
if it is determined that the swap buffer is stored, accessing the data stored in the swap buffer;
performing access to data stored in the VM module or the NVM module if it is determined that the swap buffer is not stored; and
When the access is performed, the method of managing a hybrid memory device further comprising the step of continuing the paused migration.

Images