KR102439286B1 - Stacked hybrid memory device and method for swapping data thereof - Google Patents

Abstract

The present invention includes a control logic layer for controlling an operation of a memory device, a plurality of stacked memory layers, respectively, a plurality of memory layer groups stacked on the control logic layer, and a plurality of stacked memory layers within a plurality of memory layer groups Including a plurality of first TSVs formed through and a plurality of second TSVs formed through a plurality of memory layers of a plurality of memory layer groups stacked from the control logic layer, the swap processing performance is improved, and operation speed is improved It is possible to provide a stacked hybrid memory device capable of improving , and a data swap method thereof.

Description

Stacked hybrid memory device and data swapping method thereof

The present invention relates to a hybrid memory device and a data swap method thereof, and to a stack type hybrid memory device and a data swap method thereof.

With the development of information and communication devices and computer graphics technology, the demand for high-speed, large-capacity and high-performance memory has always increased. In addition, a stacked memory device (also referred to as a 3D memory device) having a structure in which a plurality of memory layers are stacked has been proposed as one of such memory performance improvement techniques.

The stacked memory device has a structure in which a plurality of layers each including at least one memory chip are stacked in a vertical direction, and a plurality of through-electrodes (Through Silicon Via) penetrating the layers in a vertical direction between the stacked plurality of layers. Hereinafter, TSV) is formed and data transmission is performed.

Research on Hybrid Memory Cube (HMC) and High-bandwidth Memory (HBM) to replace volatile memory (VM) such as DRAM using such a stacked memory device is being actively carried out.

Meanwhile, in non-volatile memory, stored data is lost when power is not applied. Recently, as the density of non-volatile memory (NVM) is higher than that of volatile memory, research is underway to replace volatile memory with non-volatile memory. is becoming However, the non-volatile memory also has a limitation in that the write speed is very slow compared to the volatile memory. Therefore, a hybrid memory including both a non-volatile memory and a volatile memory is attracting attention to utilize both the high-density and data storage capacity of the non-volatile memory and the fast write speed of the volatile memory.

1 shows an example of a conventional stacked hybrid memory device.

1 illustrates a stacked memory device in which eight memory layers L0 to L7 are stacked on the control logic layer CL as an example, and the lower four memory layers L0 among the eight memory layers L0 to L7. Each of to L3) is a volatile memory layer (VML) in which a volatile memory is disposed, and each of the remaining four memory layers (L4 to L7) is a non-volatile memory layer (NVML) in which a non-volatile memory is disposed. That is, FIG. 1 shows a stack-type hybrid memory device in which a volatile memory layer and a non-volatile memory layer are stacked together.

In addition, the control logic layer CL and the plurality of memory layers L0 to L7 may transmit data to each other through the TSV. Also, the lowermost control logic layer CL receives data to be stored from an external CPU or GPU through micro bumps formed below, or transmits data stored in the memory layer to an external CPU or GPU. can At this time, the control logic layer CL applies a write command by designating a location in which data applied from the plurality of memory layers L0 to L7 is to be stored, or determines a location where data to be output to the outside is stored, and the plurality of memory layers ( A read command may be applied to the memory layer in which the corresponding data is stored among L0 to L7).

On the other hand, as described above, the hybrid memory is generally configured to use both the advantages of high density and non-volatileness of non-volatile memory and the advantage of fast write speed of volatile memory. The data is first stored in the memory and thereafter, the data stored in the volatile memory is transferred to the non-volatile memory to be stored again.

In this process, data swap frequently occurs between the volatile memory and the non-volatile memory. That is, a situation in which data must be exchanged between the memory layers L0 to L3 of the volatile memory layer VML and the memory layers L4 to L7 of the nonvolatile memory layer NVML frequently occurs.

At this time, as shown in FIG. 1 , in a structure in which the plurality of memory layers L0 to L7 are divided into a volatile memory layer (VML) and a nonvolatile memory layer (NVML) and sequentially stacked, the TSVs (TSV) are formed into the plurality of memory layers L0. ~ L7) are shared and used, so one of the plurality of memory layers (L0 ~ L3) of the volatile memory layer (VML) for exchanging data and a plurality of memory layers of the non-volatile memory layer (NVML) ( One of the memory layers L4 to L7) occupies the TSV (TSV), and the remaining memory layers cannot perform data swap and must wait.

The problem due to TSV occupancy occurs not only in the case of data swapping but also in the case of storing data stored in the volatile memory layer (VML) in the non-volatile memory layer (NVML), and thus the operation speed of the stack type hybrid memory device There are problems that make it difficult to improve.

Korean Patent Publication No. 10-2013-0107282 (published on October 1, 2013)

SUMMARY OF THE INVENTION An object of the present invention is to provide a stack-type hybrid memory device capable of improving an operation speed by improving swap processing performance and a data swap method thereof.

Another object of the present invention is that a plurality of volatile memory layers and a plurality of non-volatile memory layers are alternately stacked according to a predetermined arrangement method, so that a plurality of volatile memory layers and a plurality of non-volatile memory layers can simultaneously write data in parallel or An object of the present invention is to provide a stack-type hybrid memory device capable of performing a swap operation and a data swap method thereof.

According to an embodiment of the present invention for achieving the above object, a stack type hybrid memory device includes a control logic layer for controlling an operation of the memory device; a plurality of memory layer groups each including a plurality of stacked memory layers and stacked on the control logic layer; a plurality of first TSVs formed through a plurality of stacked memory layers in the plurality of memory layer groups; and a plurality of second TSVs formed through the plurality of memory layers of the plurality of memory layer groups stacked from the control logic layer.

Each of the plurality of memory layer groups may transmit data through the plurality of first TSVs when data stored in a plurality of memory layers within the same memory layer group is swapped.

When the plurality of memory layer groups swap data stored in memory layers included in different memory layer groups, data may be transmitted through the plurality of second TSVs.

Each of the plurality of memory layer groups may further include a swap buffer formed in a predetermined memory layer among the plurality of memory layers stacked and included to temporarily store swapped data.

In each of the plurality of memory layer groups, first and second memory layers for swapping data from among the plurality of memory layers select data to be swapped, respectively, and a first memory layer in which the swap buffer is not formed among the first and second memory layers A second memory layer transmits and stores the first selected data to the swap buffer of the first memory layer through the plurality of first TSVs, and the first memory layer stores the plurality of first TSVs without going through the swap buffer data is transferred to and stored in the second memory layer through the memory layer, and data stored in the swap buffer is stored in the first memory layer to perform data swapping.

In each of the plurality of memory layer groups, while the data stored in the swap buffer is stored in the first memory layer, other data selected in the second memory layer is transferred to the first memory layer through the plurality of first TSVs. It can be transferred and stored in the swap buffer.

Each of the plurality of memory layer groups may include at least one volatile memory layer in which a volatile memory is formed and at least one nonvolatile memory layer in which a nonvolatile memory is formed.

The swap buffer may be formed in the at least one non-volatile memory layer.

In a data swap method of a stacked hybrid memory device according to another embodiment of the present invention for achieving the above object, a plurality of memories each including a plurality of memory layers are stacked on a control logic layer for controlling the operation of the memory device. A plurality of first TSVs formed through a layer group and a plurality of memory layers stacked in the plurality of memory layer groups, and a plurality of first TSVs formed through a plurality of memory layers of a plurality of memory layer groups stacked from the control logic layer A data swapping method of a stacked hybrid memory device including a second TSV of selecting data to be swapped in each of the selected first and second memory layers; and if the selected first and second memory layers are memory layers within the same memory layer group, transmitting and storing data selected from each of the first and second memory layers through the plurality of first TSVs in a predetermined order. do.

Accordingly, the stacked hybrid memory device and the data swap method thereof according to an embodiment of the present invention are stacked in a plurality of memory layer group units including a predetermined number of volatile memory layers and at least one nonvolatile memory layer, Allows the memory layer groups of the to transfer data to each other independently. Therefore, a plurality of memory layer groups can simultaneously swap data in parallel or write data stored in the volatile memory layer to the non-volatile memory layer, thereby greatly improving the operation speed.

1 shows an example of a conventional stacked hybrid memory device.
2 illustrates an example of a stack-type hybrid memory device according to an embodiment of the present invention.
3 shows an example of a detailed configuration of a memory layer.
FIG. 4 is a diagram for explaining a concept in which a plurality of memory layer groups simultaneously perform data swap in parallel in the stacked hybrid memory device of FIG. 2 .
5 to 9 are diagrams for explaining a data swap method between memory layers.

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 several 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.

2 shows an example of a stack-type hybrid memory device according to an embodiment of the present invention, and FIG. 3 shows an example of a detailed configuration of a memory layer.

FIG. 4 is a diagram for explaining a concept in which a plurality of memory layer groups simultaneously perform data swap in parallel in the stacked hybrid memory device of FIG. 2 .

Referring to FIG. 2 , in the present embodiment, the stacked hybrid memory device also has a control logic layer CL and a plurality of memory layers L0 stacked on the control logic layer CL like the stacked hybrid semiconductor memory device of FIG. 1 . ~ L7).

The control logic layer CL controls operations of the plurality of stacked memory layers L0 to L7. As described above, the control logic layer CL may transmit or receive data to an external CPU or GPU through micro bumps or the like, and control the operation of a plurality of memory layers L0 to L7 through a plurality of TSVs. Accordingly, data stored in the plurality of memory layers L0 to L7 may be applied, or data to be stored in the plurality of memory layers L0 to L7 may be transmitted. In addition, the control logic layer CL may select a memory layer in which data to be swapped is stored from among a plurality of stacked memory layers L0 to L7, and may control the selected memory layer to swap corresponding data.

However, in the case of the stacked hybrid semiconductor memory device of FIG. 1 , memory layers L0 to L3 corresponding to the volatile memory layer VML among the plurality of memory layers L0 to L7 are all disposed below, and the nonvolatile memory layer The memory layers L4 to L7 corresponding to (NVML) are stacked on top of the volatile memory layer VML, whereas, as shown in FIG. 2 , in the stacked hybrid semiconductor memory device according to the present embodiment, the volatile memory layer ((L0 to L2), (L4 to L6)) and the nonvolatile memory layers L3 and L7 may be alternately arranged in a predetermined order.

In particular, in the present embodiment, at least one volatile memory layer ((L0 to L2), (L4 to L6)) and at least one nonvolatile memory layer (L3, L7) constitute the memory layer groups MLG1 and MLG2. And control in units of a plurality of memory layer groups (MLG1, MLG2) each including at least one volatile memory layer ((L0 ~ L2), (L4 ~ L6)) and at least one nonvolatile memory layer (L3, L7) It is stacked on the logic layer CL.

In addition, a plurality of TSVs for transferring data between the stacked memory layers are formed between the plurality of memory layers L0 to L7. However, in the conventional stack-type hybrid memory device shown in FIG. 1 , a plurality of TSVs basically pass through a plurality of memory layers from the control logic layer CL located at the bottom to the memory layer L7 located at the top. is formed by

On the other hand, in the stacked hybrid memory device of the present embodiment, a plurality of TSVs are electrically connected from the control logic layer CL to the uppermost memory layer L7, similar to the TSVs of FIG. 1 , a plurality of second TSVs (TSV2). and a plurality of first TSVs TSV1 electrically connecting memory layers positioned inside each of the plurality of memory layer groups MLG1 and MLG2.

Meanwhile, each of the plurality of memory layers L0 to L7 includes a bank area BA and a plurality of banks in which a plurality of banks each including at least one memory cell array MCA are disposed. ) may include a plurality of channel areas CA including an interconnect area IA in which a plurality of TSVs for transmitting data stored in the control logic layer CL or another memory layer are formed. Here, as shown in FIG. 2 , the channel regions CA at positions corresponding to each other in each of the plurality of memory layers L0 to L7 form a channel through which data can be transmitted using a common TSV. can

In addition, at least one memory cell array included in each of a plurality of banks includes a plurality of word lines (not shown) constituting a row and a plurality of bit lines (not shown) constituting a column. A plurality of memory cells MC for storing data are disposed at positions where a plurality of word lines and a plurality of bit lines cross each other. In addition, a row buffer RB for temporarily storing data stored in memory cells of a selected row in the memory cell array or data to be stored in memory cells of a selected row is further disposed in the plurality of banks.

In FIG. 2 , one of a plurality of volatile memory layers ((L0 to L2), (L4 to L6)) is shown as an example of the memory layers L0 to L7, and in the case of the nonvolatile memory layers L3 and L7, each A plurality of dies including a plurality of memory blocks including a plurality of memory cells may be configured to replace a bank, but there is a similarity in the rough configuration. In addition, a volatile memory may be formed in some regions of the non-volatile memory layers L3 and L7.

2 and 4 illustrate the first and second TSVs TSV1 and TSV2 formed in one channel for convenience of explanation, the first and second TSVs TSV1 and TSV2 are all channels, that is, each memory layer. It can be formed at a designated location in the entire area of .

In the stack-type hybrid memory device of the present embodiment described above, in the stack-type hybrid memory device according to the present embodiment, a plurality of volatile memory layers (L0 to L2, L4 to L6) and a plurality of non-volatile memory layers (L3, L7) are a plurality is divided into memory layer groups MLG1 and MLG2 of As shown, this is to solve the problem that some of the memory layers occupy the TSV while performing data swap, so that other memory layers cannot transmit data through the TSV and have to wait.

That is, in the present embodiment, since a plurality of first TSVs TSV1 are separately formed in each memory layer group MLG1 and MLG2, a plurality of memory layers (L0 to L3) of each memory layer group MLG1 and MLG2, (L4 to L7)) may perform data swapping by using the plurality of first TSVs TSV1 formed in the corresponding memory layer groups MLG1 and MLG2 irrespective of data transmission states of other memory layer groups.

Accordingly, in the stacked hybrid memory device of the present embodiment, each of the memory layer groups MLG1 and MLG2 is formed in parallel between the volatile memory layers ((L0 to L2), (L4 to L6) and the nonvolatile memory layers L3 and L7 at the same time. of data swapping can be performed. In FIG. 4 , as an example of this, while one volatile memory layer L0 and a nonvolatile memory layer L3 of the first memory layer group MLG1 perform data swapping, one of the second memory layer group MLG2 It indicates that the volatile memory layer (L0) and the nonvolatile memory layer (L3) can perform data swap together.

In this case, the plurality of memory layer groups MLG1 and MLG2 select first and second memory layers in which data to be swapped from among the plurality of memory layers are stored according to the control of the control logic layer, and in each of the selected first and second memory layers Data to be swapped is selected, and if the selected first and second memory layers are memory layers within the same memory layer group, a plurality of first TSVs (TSV1) are selected in a predetermined order for data selected from each of the first and second memory layers. A data swap operation can be performed by transmitting through and storing the transmitted data. At this time, different memory layer groups can also swap data between internal memory layers.

Accordingly, it is possible to significantly improve performance by improving the operation speed of the stacked hybrid memory device.

However, if the selected first and second memory layers are memory layers in different memory layer groups, the data selected in each of the first and second memory layers are transmitted through a plurality of second TSVs (TSV2) in a predetermined order and transmitted. A data swap operation can be performed by saving data.

5 to 9 are diagrams for explaining a data swap method between memory layers.

5 to 9 illustrate data swapping between two memory layers L1 and L2. Although data swapping between the same volatile memory layer is illustrated here for convenience of description, data swapping between a nonvolatile memory layer and a volatile memory layer may also be performed in the same manner.

In addition, swap buffers SB0 and SB1 for data swap only in the interconnect area IA of one of the two memory layers L1 and L2 in which data swap is performed (here, for example, the first memory layer L1). A case in which is further formed is shown. Here, the swap buffers SB0 and SB1 are divided into two to correspond to the size of the swap data that can be transmitted at one time through the plurality of first TSVs TSV1, but the swap buffer may be implemented as a single buffer device. In this case, the size of the swap data that can be transmitted at once is smaller than the size of the row buffer.

In addition, although the swap buffers SB0 and SB1 may be provided in each of the plurality of memory layers L0 to L7, only one of the memory layers L1 and L2 in which data is to be swapped is the swap buffer SB0. , SB1) can be provided, data swapping can be performed, so there is an advantage in that manufacturing cost can be reduced. In particular, in the stack-type hybrid memory device, as described above, data swap is rarely required between the volatile memory layers. In reality, data swap can be viewed as occurring between the volatile memory layer and the non-volatile memory layer, so the swap buffer (SB0, SB1) may be formed in only one of the volatile memory layers ((L0 to L2), (L4 to L6)) or the nonvolatile memory layers L3 and L7 in the memory layer groups MLG1 and MLG2.

In addition, since the data swapping method between memory layers of FIGS. 5 to 9 only describes data swapping between two memory layers L1 and L2, the two memory layers L1 and L2 must be the same memory layer group MLG1, It does not need to be a memory layer disposed in MLG2). However, when data swap is required in other memory layers while data swap is performed by two memory layers, the memory layers for swapping data must be memory layers arranged in the same memory layer group (MLG1, MLG2). .

6 to 9, the lower part indicates the time required for each step in data swapping.

5 to 9, first, as indicated by ① in FIG. 5, in the data swap operation between the two memory layers L1 and L2, the memory cells of the bank in each of the two memory layers L1 and L2 A row in the array MCA in which data to be swapped is stored is selected ACT, and data of the selected row is transferred to a corresponding row buffer RB, respectively. Here, the time when a row is selected and data of the selected row is transferred to and stored in the row buffer RB is denoted by tRCD.

And as indicated by ② in FIG. 6 , data stored in the row buffer RB of the second memory layer L2 in which the swap buffers SB0 and SB1 are not formed is transferred to the first memory layer through the first TSV TSV1 . It is applied to the swap buffer SB0 of (L1) and stored. Since the data in the row buffer RB of the second memory layer L2 has moved to the swap buffer SB0, the data stored in the row buffer RB of the first memory layer L1 is as indicated by ③ in FIG. 7 . , is directly transferred to and stored in the row buffer RB of the second memory layer L2 without going through the swap buffers SB0 and SB1.

Thereafter, as indicated by ④ in FIG. 8 , data stored in the swap buffer SB0 is transferred to and stored in the row buffer RB of the first memory layer L1, and in this case, as indicated by ⑤, the second memory Untransmitted data stored in the row buffer RB of the layer L2 may be applied to and stored in another swap buffer SB1 of the first memory layer L1 through the first TSV TSV1 . That is, the data stored in the swap buffer SB0 is stored in the row buffer RB of the first memory layer L1 and the data stored in the row buffer RB of the second memory layer L2 is stored in the swap buffer SB1. An operation that is applied and stored may be performed together. And the operation indicated by ⑤ is the same operation as ② of FIG. 6 , and thereafter, as shown in FIG. 9 , all data stored in the row buffers RB of the first and second memory layers L1 and L2 are swapped. The operations of FIGS. 6 to 8 are repeatedly performed until the

As a result, except for the first time, the data stored in the row buffer RB of the second memory layer L2 is transferred to the swap buffers SB0 and SB1 of the first memory layer L1 through the first TSV TSV1. The applied and stored operation and the previously applied and stored data of the swap buffers SB0 and SB1 are transferred to and stored in the row buffer RB of the first memory layer L1 can be performed together, thereby greatly increasing the data swap time can be reduced.

In addition, since the data swap operation can be independently performed in each of the plurality of memory layer groups MLG1 and MLG2, the data swap time can be further reduced, thereby improving the performance of the stack type hybrid memory device.

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.

CL: control logic layer L0 to L7: memory layer
VML: Volatile Memory Layer NVML: Non-Volatile Memory Layer
TSV1, TSV2: first and second TSV CA: channel region
BA: Bank Area IA: Interconnect Area
MCA: memory cell array RB: row buffer
SB0, SB1: swap buffer

Claims (15)

a control logic layer that controls the operation of the memory device;
a plurality of memory layer groups each including a plurality of stacked memory layers and stacked on the control logic layer;
a plurality of first TSVs formed by dividing each of the plurality of memory layer groups to penetrate a plurality of memory layers included in each memory layer group; and
a plurality of second TSVs formed through a plurality of memory layers stacked from the control logic layer without distinguishing the plurality of memory layer groups;
The plurality of memory layer groups are
When data stored in a plurality of memory layers in the same memory layer group is swapped, data is transmitted through the plurality of first TSVs;
A stack type hybrid memory device that transmits data through the plurality of second TSVs when data stored in memory layers included in different memory layer groups are swapped. delete delete The method of claim 1 , wherein each of the plurality of memory layer groups comprises:
The stacked hybrid memory device further comprising: a plurality of swap buffers formed in a predetermined memory layer among the plurality of memory layers stacked and included to temporarily store swapped data. 5. The method of claim 4, wherein each of the plurality of memory layer groups comprises:
The first and second memory layers for swapping data among the plurality of memory layers select data to be swapped, respectively, and the second memory layer in which the plurality of swap buffers are not formed among the first and second memory layers is first selected data is transferred to and stored in one of the plurality of swap buffers of the first memory layer through the plurality of first TSVs, wherein the first memory layer does not go through the plurality of swap buffers but through the plurality of first TSVs The stack-type hybrid memory device is transferred to and stored in the second memory layer, and data stored in one of the plurality of swap buffers is stored in the first memory layer to perform data swapping. 6. The method of claim 5, wherein each of the plurality of memory layer groups comprises:
While data stored in one of the plurality of swap buffers is stored in the first memory layer, other data selected in the second memory layer is transferred to another swap buffer of the first memory layer through the plurality of first TSVs A stacked hybrid memory device that is stored and stored. 7. The method of claim 6, wherein each of the plurality of memory layer groups comprises:
A stacked hybrid memory device comprising: at least one volatile memory layer in which a volatile memory is formed; and at least one nonvolatile memory layer in which a non-volatile memory is formed. 8. The method of claim 7, wherein the plurality of swap buffers are
A stacked hybrid memory device formed on the at least one non-volatile memory layer. A plurality of memory layer groups stacked on a control logic layer for controlling an operation of a memory device, each including a plurality of memory layers, and a plurality of memories included in each memory layer group by distinguishing each of the plurality of memory layer groups A stacked hybrid including a plurality of first TSVs formed through the layers and a plurality of second TSVs formed through a plurality of memory layers stacked from the control logic layer without distinguishing the plurality of memory layer groups A method for swapping data in a memory device, the method comprising:
selecting first and second memory layers in which data to be swapped is stored from among a plurality of memory layers;
selecting data to be swapped in each of the selected first and second memory layers; and
If the selected first and second memory layers are memory layers within the same memory layer group, transmitting and storing data selected from each of the first and second memory layers through the plurality of first TSVs in a predetermined order; Data swapping method of stacked hybrid memory device. 10. The method of claim 9, wherein the data swap method
If the selected first and second memory layers are memory layers in different memory layer groups, transmitting and storing data selected in each of the first and second memory layers through the plurality of second TSVs in a predetermined order is further performed. A method of swapping data in a stackable hybrid memory device comprising: 11. The method of claim 10, wherein the first memory layer comprises:
A data swap method in a stacked hybrid memory device in which a plurality of swap buffers in which swapped data is temporarily stored are formed. The method of claim 11, wherein the step of transmitting and storing the plurality of first TSVs comprises:
transmitting and storing data selected in the second memory layer to one of the plurality of swap buffers of the first memory layer through the plurality of first TSVs;
transmitting and storing data of the first memory layer to the second memory layer through the plurality of first TSVs without going through the plurality of swap buffers; and
and storing data stored in one of the plurality of swap buffers in the first memory layer. The method of claim 12, wherein the step of transmitting and storing the plurality of first TSVs comprises:
While data stored in one of the plurality of swap buffers is stored in the first memory layer, other data selected in the second memory layer is transferred to another swap buffer of the first memory layer through the plurality of first TSVs Data swapping method of a stacked hybrid memory device further comprising the step of being stored. The method of claim 11 , wherein each of the plurality of memory layer groups comprises:
A data swapping method of a stack type hybrid memory device, comprising at least one volatile memory layer having a volatile memory formed thereon and at least one nonvolatile memory layer having a nonvolatile memory formed thereon. 15. The method of claim 14, wherein the plurality of swap buffers
A data swapping method of a stacked hybrid memory device formed on the at least one non-volatile memory layer.

Images