Data storage device and storage matrix selection method thereof

Abstract

A storage matrix selection method is used for a data storage medium including a plurality of storage matrices, wherein each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks. The storage matrix selection method comprises the following steps: defining at least one data column with the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted; judging whether there are at least two storage matrices that are uncorrectable according to the distribution of the damaged data columns; and clearing the record of the damaged data columns and disabling the worst storage matrix having the most damaged data columns compared to other storage matrices.

Description

Technical Field

[0001] The present invention relates to a technology for data storage, and in particular to a data storage device and a storage matrix selection method thereof. Background Art

[0002] Data storage devices such as solid state drives (SSD), secure digital (SD) memory cards or portable flash memory are mainly composed of a control unit and a data storage medium. The data storage medium is composed of a plurality of dies enabled by a chip enable (CE) of the control unit. Each die has a plurality of storage planes, and these storage planes have a plurality of data columns for storing data.

[0003] Substandard products are inevitably produced during the manufacturing process of data storage devices. Therefore, before the substandard products are put on the market, it is necessary to screen out all damaged data columns in the data storage medium and record the damaged data columns in a damaged data column summary table. In this way, based on the damaged data column summary table, the control unit can avoid accessing these damaged data columns, thereby avoiding data access errors during operation.

[0004] Data storage devices usually use error correcting codes (ECC) to correct the data they store. However, the number of damaged data columns provided by the hardware is limited by the size of the memory. When the limited number of damaged data columns is evenly distributed to each storage matrix, some storage matrices with more damaged data columns may not be corrected by ECC. Summary of the invention

[0005] The object of the present invention is to provide a storage matrix selection method for a data storage device. The storage matrix selection method of the present invention can effectively use a limited number of damaged data columns, thereby optimizing the distribution of damaged data columns on the storage matrix.

[0006] The storage matrix selection method provided by the present invention is applicable to data storage media. The data storage media includes a plurality of storage matrices, each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks. The storage matrix selection method includes: defining at least one data column with the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted; judging whether there are at least two storage matrices that are uncorrectable according to the distribution of the damaged data columns; and clearing the record of the damaged data columns and disabling the worst storage matrix with the most damaged data columns compared to other storage matrices.

[0007] The data storage device provided by the present invention includes a controller and a data storage medium. The data storage medium includes a plurality of storage matrices, each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks. The controller executes a storage matrix selection method including: defining at least one data column with the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted; judging whether there are at least two storage matrices that are uncorrectable according to the distribution of the damaged data columns; and clearing the record of the damaged data columns and disabling the worst storage matrix with the most damaged data columns compared to other storage matrices.

[0008] In one embodiment of the present invention, the above-mentioned memory matrix selection method is repeatedly executed until at most one memory matrix is ​​determined to be uncorrectable.

[0009] In one embodiment of the present invention, the step of defining at least one data row with the maximum number of error bits in a worst large block as a damaged data row until the number of damaged data rows is exhausted comprises: writing preset data to the data page; reading the written data from the data page; comparing the preset data with the written data to obtain the number of error bits of each data row in each large block, and calculating the number of errors in each large block; selecting a large block with the maximum number of errors as the worst large block; and recording a data row with the maximum number of error bits in the worst large block as the damaged data row.

[0010] In an embodiment of the present invention, the error number is a total number of error bits or an average number of error bits.

[0011] Another storage matrix selection method provided by the present invention is applicable to data storage media. The data storage media includes a plurality of storage matrices, each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks. The storage matrix selection method includes: defining at least one data column with the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted; judging whether there is at least one storage matrix that is uncorrectable based on the distribution of the damaged data columns; and clearing the record of the damaged data columns and disabling the worst storage matrix with the most damaged data columns compared to other storage matrices.

[0012] Another data storage device provided by the present invention includes a controller and a data storage medium. The data storage medium includes a plurality of storage matrices, each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks. The controller performs a storage matrix selection method including: defining at least one data column with the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted; judging whether there is at least one storage matrix that is uncorrectable based on the distribution of the damaged data columns; and clearing the record of the damaged data columns and disabling the worst storage matrix with the most damaged data columns compared to other storage matrices.

[0013] In one embodiment of the present invention, the above-mentioned storage matrix selection method is repeatedly executed until no storage matrix is ​​determined to be uncorrectable.

[0014] In one embodiment of the present invention, the step of defining at least one data row with the maximum number of error bits in a worst large block as a damaged data row until the number of damaged data rows is exhausted comprises: writing preset data to the data page; reading the written data from the data page; comparing the preset data with the written data to obtain the number of error bits of each data row in each large block, and calculating the number of errors in each large block; selecting a large block with the maximum number of errors as the worst large block; and recording a data row with the maximum number of error bits in the worst large block as the damaged data row.

[0015] In an embodiment of the present invention, the error number is a total number of error bits or an average number of error bits.

[0016] In summary, the storage matrix selection method of the data storage device provided by the present invention can leave one damaged storage matrix, disable other uncorrectable damaged storage matrices, and use all damaged data columns on the remaining damaged storage matrix, or prohibit all uncorrectable damaged storage matrices and distribute a limited number of damaged data columns to other storage matrices, so that the number of damaged data columns can be effectively used and the distribution of damaged data columns can be optimized.

[0017] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following specifically cites a preferred embodiment and describes it in detail with the accompanying drawings as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A data storage device according to an embodiment of the present invention is shown.

[0019] Figure 2A die according to an embodiment of the present invention is shown.

[0020] Figure 3 A memory matrix according to an embodiment of the present invention is shown.

[0021] Figure 4 A die according to another embodiment of the present invention is shown.

[0022] Figure 5 is a flow chart of a storage matrix selection method according to an embodiment of the present invention.

[0023] Figure 6 The flowchart of defining a damaged data column according to an embodiment of the present invention is shown in FIG.

[0024] Figure 7 is a schematic diagram for disabling a damaged storage matrix.

[0025] Figure 8 is a flow chart of a storage matrix selection method according to another embodiment of the present invention. DETAILED DESCRIPTION

[0026] Hereinafter, the present invention will be described in detail by illustrating various embodiments of the present invention by means of drawings. However, the inventive concept may be embodied in many different forms and should not be construed as being limited to the exemplary embodiments described herein. In addition, the same reference numerals may be used to represent similar elements in the drawings.

[0027] Figure 1 , 2, 3 respectively show the data storage device, tube core, and storage matrix of the embodiment of the present invention. The data storage device 1 of the present invention includes a control unit 2 and a data storage medium 3, wherein the data storage medium 3 can be a non-volatile memory, such as flash memory, MRAM (magnetic RAM), FRAM (ferroelectric RAM), PCM (phase change memory), STTRAM (spin transfer torque RAM), ReRAM (resistance RAM) or a memristor capable of storing data for a long time. The data storage medium 3 may include a plurality of tube cores (Dies) D0, D1, D2, ..., D(s-1). The controller 2 uses a chip enable (CE) pin to enable at least one tube core D0, D1, D2, ..., D(s-1) to be accessed. Each tube core D0, D1, D2, ..., D(s-1) includes a plurality of storage matrices (Planes) PL0, PL1, PL2, ..., PL(t-1). Each storage matrix PL0, PL1, PL2, ..., PL(t-1) includes a plurality of blocks B0, B1, B2, ..., B(z-1). Each block B0, B1, B2, ..., B(z-1) includes a plurality of data pages P0, P1, P2, ..., P(n-1). Each data page P0, P1, P2, ..., P(n-1) includes a plurality of data rows and is divided into a plurality of large blocks C0, C1, C2, ..., C(m-1).

[0028] The storage matrix selection method disclosed in the present invention is described as follows. Figure 4The die of another embodiment of the present invention is shown. In order to effectively select a storage matrix, preferably, the die D0 in the data storage medium 3 is selected as a sample die for executing the storage matrix selection method of the present invention, and the present invention is not limited thereto. In this embodiment, the die D0 includes four storage matrices PL0, PL1, PL2, PL3. Each storage matrix PL0, PL1, PL2, PL3 includes a plurality of blocks B0, B1, B2, ..., B(z-1). Each block B0, B1, B2, ..., B(z-1) includes a plurality of data pages P0, P1, P2, ..., P(n-1). Each data page P0, P1, P2, ..., P(n-1) includes a plurality of data columns and is divided into eight large blocks, wherein the data pages P0, P1, P2, ..., P(n-1) in the storage matrix PL0 are divided into eight large blocks C0, C1, C2, C3, C4, C5, C6, C7, and the data pages P0, P1, P2, ..., P(n-1) in the storage matrix PL1 are divided into eight large blocks C8, C9, C10, C11, C12, C13, C14, C15, C16, C17. 2, C13, C14, C15, the data pages P0, P1, P2, ..., P(n-1) in the storage matrix PL2 are divided into eight large blocks C16, C17, C18, C19, C20, C21, C22, C23, and the data pages P0, P1, P2, ..., P(n-1) in the storage matrix PL3 are divided into eight large blocks C24, C25, C26, C27, C28, C29, C30, C31.

[0029] Figure 5 FIG. 1 is a flow chart of a storage matrix selection method according to an embodiment of the present invention. Figure 5 As shown, the storage matrix selection method of the present invention comprises step S1: the control unit 2 defines at least one data column with the maximum number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted.

[0030] In step S1, the control unit 2 defines the number of damaged data columns, such as Figure 6The flowchart of defining damaged data columns according to an embodiment of the present invention is shown. In step S11, the control unit 2 writes preset data to the data pages P0, P1, P2, ..., P(n-1) of the storage matrix PL0, PL1, PL2, PL3. The preset data can be random or specific data. In step S13, the control unit 2 reads the write data from the data pages P0, P1, P2, ..., P(n-1) of the storage matrix PL0, PL1, PL2, PL3. In step S15, the control unit 2 compares the preset data with the write data to obtain the number of error bits of each data column in each large block C0, C1, C2, ..., C31, and calculates the number of errors in each large block C0, C1, C2, ..., C31. When a bit in the written data is not equal to the corresponding bit in the preset data, the bit is determined to be an error bit, and the error number of each large block C0, C1, C2, ..., C31 is obtained by summing up the number of error bits appearing on each data column in each large block C0, C1, C2, ..., C31. In step S17, the control unit 2 selects the large block with the largest error number as the worst large block, wherein the error number is the total number of error bits or the average value of error bits in each large block C0, C1, C2, ..., C31. In step S19, the control unit 2 records the data column with the largest number of error bits in the worst large block as the damaged data column. Then, the control unit 2 repeatedly counts the damaged data columns into the damaged data column summary table until the number of damaged data columns is exhausted (for example, the damaged data column summary table can record up to 512 damaged data columns). Finally, the control unit 2 calculates the total number of error bits of the data columns in each large block C0, C1, C2, ..., C31, the average error rate of each data column in each large block C0, C1, C2, ..., C31, the position of the damaged data column in each large block C0, C1, C2, ..., C31, and the size of each large block C0, C1, C2, ..., C31. Calculating the average error rate of the data columns and the size of the large blocks C0, C1, C2, ..., C31 is a technique well known to those skilled in the art and will not be described here.

[0031] like Figure 5 As shown, the storage matrix selection method of the present invention includes step S3: the control unit 2 determines whether at least two storage matrices PL0, PL1, PL2, PL3 are uncorrectable according to the distribution of the damaged data columns.

[0032] In step S3, the control unit 2 first determines whether each large block C0, C1, C2, ..., C31 can be corrected by ECC based on the average error rate of each data column in each large block, the position of the damaged data column in each large block, and the size of each large block. When the number of errors in the large block in the storage matrix exceeds the correction capability of ECC, this means that this storage matrix is ​​uncorrectable. Then, the control unit 2 determines whether at least two of the storage matrices PL0, PL1, PL2, and PL3 are uncorrectable. In addition, the distribution of the damaged data column includes the number, position, and error rate of the damaged data column in each storage matrix PL0, PL1, PL2, and PL3.

[0033] like Figure 5 As shown, the storage matrix selection method of the present invention includes step S5: when at least two storage matrices PL0, PL1, PL2, PL3 are determined to be uncorrectable, the control unit 2 clears the records of the damaged data columns and disables the worst storage matrix with the most damaged data columns compared to other storage matrices.

[0034] In step S5, when multiple storage matrices are uncorrectable, the control unit 2 counts the total number of damaged data columns in each uncorrectable storage matrix. For example, assume that the storage matrices PL0 and PL3 are correctable, the storage matrices PL1 and PL2 are uncorrectable, and the number of damaged data columns in the storage matrix PL1 is greater than the number of damaged data columns in the storage matrix PL2. The control unit 2 first determines whether there are at least two storage matrices PL0, PL1, PL2, and PL3 that are uncorrectable. When the control unit 2 determines that the storage matrices PL1 and PL2 are uncorrectable and determines that the uncorrectable storage matrix PL1 is the worst storage matrix based on the number of damaged data columns and should be disabled first, the control unit 2 eliminates the record of the damaged data column in the damaged data column summary table and disables the uncorrectable storage matrix PL1, such as Figure 7 Schematic diagram for disabling a damaged memory matrix is ​​shown.

[0035] like Figure 5As shown, the storage matrix selection method of the present invention can be repeatedly executed until the control unit 2 determines that at most one storage matrix PL0, PL1, PL2, PL3 is uncorrectable. For example, after the control unit 2 eliminates the record of the damaged data column in the damaged data column summary table and disables the uncorrectable storage matrix PL1, the control unit 2 again defines at least one data column with the maximum number of error bits in a worst block as a damaged data column in the storage matrices PL0, PL2, PL3 until the number of damaged data columns is exhausted. Then, the control unit 2 determines whether there are at least two storage matrices PL0, PL2, PL3 that are uncorrectable. When the control unit 2 determines that only the storage matrix PL2 is uncorrectable, the control unit 2 records the damaged data column of the storage matrix PL2 in the damaged data column summary table until the number of damaged data columns is exhausted, and attempts to correct the storage matrix PL2 through ECC, so that the storage matrix PL2 still has a chance to be used. Since the storage matrices PL0 and PL3 are correctable and the storage matrix PL2 may be used, the storage matrix selection method of the present invention can effectively use a limited number of damaged data columns and can optimize the distribution of damaged data columns on the storage matrix.

[0036] In another embodiment, it is assumed that the storage matrices PL0, PL1, PL2, PL3 after step S1 are correctable. The control unit 2 first determines whether there are at least two storage matrices PL0, PL1, PL2, PL3 that are uncorrectable. When the control unit 2 determines that the storage matrices PL0, PL1, PL2, PL3 are correctable, the control unit 2 records the damaged data columns of the storage matrices PL0, PL1, PL2, PL3 into the damaged data column summary table, so that the control unit 2 can avoid accessing these damaged data columns according to the damaged data summary table. In this way, except for the damaged data columns, all data columns can be used to store data. Therefore, the data storage medium 3 of the data storage device 1 of the present invention has an optimized data storage capacity.

[0037] Figure 8is a flow chart of a storage matrix selection method according to another embodiment of the present invention. In this embodiment, the storage matrix selection method of the present invention comprises step S1: the control unit 2 defines at least one data column with the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns is exhausted; step S3': the control unit 2 determines whether at least one storage matrix PL0, PL1, PL2, PL3 is uncorrectable according to the distribution of the damaged data columns; step S5': the control unit 2 clears the records of the damaged data columns and disables the worst storage matrix with the most damaged data columns compared to other storage matrices. It can be noted that the difference between the present storage matrix selection method and the previous storage matrix selection method is that the control unit 2 in the present storage matrix selection method determines whether at least one storage matrix is ​​uncorrectable in step S3', the control unit 2 disables at least one uncorrectable storage matrix when determining that at least one storage matrix is ​​uncorrectable in step S5', and the storage matrix selection method of the present invention can be repeatedly performed until the control unit 2 determines that no storage matrix is ​​uncorrectable.

[0038] For example, it is assumed that after step S1, the storage matrices PL0 and PL3 are correctable, the storage matrices PL1 and PL2 are uncorrectable, and the number of damaged data columns in the storage matrix PL1 is greater than the number of damaged data columns in the storage matrix PL2. The control unit 2 first determines whether there is at least one storage matrix PL0, PL1, PL2, and PL3 that is uncorrectable. When the control unit 2 determines that the storage matrices PL1 and PL2 are uncorrectable and determines that the uncorrectable storage matrix PL1 is the worst storage matrix based on the number of damaged data columns and should be disabled first, the control unit 2 eliminates the record of the damaged data column in the damaged data column summary table and disables the uncorrectable storage matrix PL1. After the control unit 2 eliminates the record of the damaged data column in the damaged data column summary table and disables the uncorrectable storage matrix PL1, the control unit 2 again defines at least one data column with the maximum number of error bits in a worst block in the storage matrices PL0, PL2, and PL3 as a damaged data column until the number of damaged data columns is used up. The control unit 2 determines whether there is at least one storage matrix PL0, PL2, and PL3 that is uncorrectable. When the control unit 2 determines that the storage matrix PL2 is uncorrectable, the control unit 2 deletes the record of the damaged data column in the damaged data column summary table and disables the uncorrectable storage matrix PL2. After the control unit 2 deletes the record of the damaged data column in the damaged data column summary table and disables the uncorrectable storage matrix PL2, the control unit 2 again defines at least one data column with the maximum number of error bits in a worst block in the storage matrices PL0 and PL3 as a damaged data column until the number of damaged data columns is used up. Next, the control unit 2 determines whether there is at least one storage matrix PL0 and PL3 that is uncorrectable. When the control unit 2 determines that the storage matrix PL0 and PL3 are correctable, the control unit 2 records the damaged data columns of the storage matrix PL0 and PL3 in the damaged data column summary table, so that the control unit 2 can avoid accessing these damaged data columns according to the damaged data summary table. In this way, except for the damaged data columns, all data columns can be used to store data. Therefore, the data storage medium 3 of the data storage device 1 of the present invention has an optimized data storage capacity.

[0039] In summary, the storage matrix selection method of the present invention can not only define at least one data column with the largest number of error bits in the worst large block as a damaged data column, and record the damaged data column in the damaged data column summary table, but also determine whether there is a storage matrix that is uncorrectable, and disable the damaged storage matrix. Therefore, the control unit of the data storage device of the present invention can avoid accessing damaged data columns and damaged storage matrices. In this way, the storage matrix selection method for a data storage device of the present invention can effectively use a limited number of damaged data columns, thereby optimizing the distribution of damaged data columns on the storage matrix.

[0040] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technician familiar with this profession can make some changes or modifications to equivalent embodiments of equivalent changes by using the methods and technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A storage matrix selection method, characterized in that: A data storage medium including a plurality of storage matrices, wherein each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks, wherein the storage matrix selection method includes: defining at least one data column having the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column total table is exhausted, wherein the worst large block is the large block having the largest number of errors; Determining whether at least two storage matrices are uncorrectable based on the distribution of the plurality of damaged data columns; and The records of the plurality of damaged data columns in the damaged data column summary table are cleared, and when it is determined that the at least two memory matrices are uncorrectable, the worst memory matrix which is uncorrectable compared to other memory matrices and has the most damaged data columns is disabled.

2. The storage matrix selection method according to claim 1, characterized in that: The memory matrix selection method is repeatedly executed until at most one memory matrix is ​​determined to be uncorrectable.

3. The storage matrix selection method according to claim 1, characterized in that: The step of defining at least one data column having the maximum number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column summary table is exhausted comprises: Writing preset data to the plurality of data pages; reading write data from the plurality of data pages; comparing the preset data with the written data to obtain the number of error bits of each data column in each large block, and calculating the number of errors in each large block; Selecting a chunk with the largest number of errors as the worst chunk; and A data row having a maximum number of error bits in the worst block is recorded as the damaged data row.

4. A storage matrix selection method, characterized in that: A data storage medium including a plurality of storage matrices, wherein each storage matrix includes a plurality of blocks, each block includes a plurality of data pages, and each data page includes a plurality of data columns and is divided into a plurality of large blocks, wherein the storage matrix selection method includes: defining at least one data column having the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column total table is exhausted, wherein the worst large block is the large block having the largest number of errors; Determining whether at least one storage matrix is ​​uncorrectable according to the distribution of the plurality of damaged data columns; and The records of the plurality of damaged data columns in the damaged data column summary table are cleared, and when it is determined that the at least one memory matrix is ​​uncorrectable, the worst memory matrix which is uncorrectable compared to other memory matrices and has the most damaged data columns is disabled.

5. The storage matrix selection method according to claim 4, characterized in that: The step of defining at least one data column having the maximum number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column summary table is exhausted comprises: Writing preset data to the plurality of data pages; reading write data from the plurality of data pages; comparing the preset data with the written data to obtain the number of error bits of each data column in each large block, and calculating the number of errors in each large block; Selecting a chunk with the largest number of errors as the worst chunk; and A data row having a maximum number of error bits in the worst block is recorded as the damaged data row.

6. A data storage device, characterized in that: The invention comprises a controller and a data storage medium, wherein the data storage medium comprises a plurality of storage matrices, each storage matrix comprises a plurality of blocks, each block comprises a plurality of data pages, and each data page comprises a plurality of data columns and is divided into a plurality of large blocks, and the controller performs a storage matrix selection method comprising: defining at least one data column having the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column total table is exhausted, wherein the worst large block is the large block having the largest number of errors; Determining whether at least two storage matrices are uncorrectable based on the distribution of the plurality of damaged data columns; and The records of the plurality of damaged data columns in the damaged data column summary table are cleared, and when it is determined that the at least two memory matrices are uncorrectable, the worst memory matrix which is uncorrectable compared to other memory matrices and has the most damaged data columns is disabled.

7. The data storage device according to claim 6, wherein: The memory matrix selection method is repeatedly executed until at most one memory matrix is ​​determined to be uncorrectable.

8. The data storage device according to claim 6, wherein: The step of defining at least one data column having the maximum number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column summary table is exhausted comprises: Writing preset data to the plurality of data pages; reading write data from the plurality of data pages; comparing the preset data with the written data to obtain the number of error bits of each data column in each large block, and calculating the number of errors in each large block; Selecting a chunk with the largest number of errors as the worst chunk; and A data row having a maximum number of error bits in the worst block is recorded as the damaged data row.

9. A data storage device, characterized in that: The invention comprises a controller and a data storage medium, wherein the data storage medium comprises a plurality of storage matrices, each storage matrix comprises a plurality of blocks, each block comprises a plurality of data pages, and each data page comprises a plurality of data columns and is divided into a plurality of large blocks, and the controller performs a storage matrix selection method comprising: defining at least one data column having the largest number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column total table is exhausted, wherein the worst large block is the large block having the largest number of errors; Determining whether at least one storage matrix is ​​uncorrectable according to the distribution of the plurality of damaged data columns; and The records of the plurality of damaged data columns in the damaged data column summary table are cleared, and when it is determined that the at least one memory matrix is ​​uncorrectable, the worst memory matrix which is uncorrectable compared to other memory matrices and has the most damaged data columns is disabled.

10. The data storage device according to claim 9, characterized in that The step of defining at least one data column having the maximum number of error bits in a worst large block as a damaged data column until the number of damaged data columns in the damaged data column summary table is exhausted comprises: Writing preset data to the plurality of data pages; reading write data from the plurality of data pages; comparing the preset data with the written data to obtain the number of error bits of each data column in each large block, and calculating the number of errors in each large block; Selecting a chunk with the largest number of errors as the worst chunk; and A data row having a maximum number of error bits in the worst block is recorded as the bad data row.

Images