WO2011124185A2 - 内存检测方法和内存检测装置 - Google Patents

内存检测方法和内存检测装置 Download PDF

Info

Publication number
WO2011124185A2
WO2011124185A2 PCT/CN2011/074099 CN2011074099W WO2011124185A2 WO 2011124185 A2 WO2011124185 A2 WO 2011124185A2 CN 2011074099 W CN2011074099 W CN 2011074099W WO 2011124185 A2 WO2011124185 A2 WO 2011124185A2
Authority
WO
WIPO (PCT)
Prior art keywords
memory
basic unit
write operation
unit
basic
Prior art date
Application number
PCT/CN2011/074099
Other languages
English (en)
French (fr)
Other versions
WO2011124185A3 (zh
Inventor
叶荣标
张志龙
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to CN201180001054.4A priority Critical patent/CN102893263B/zh
Priority to PCT/CN2011/074099 priority patent/WO2011124185A2/zh
Publication of WO2011124185A2 publication Critical patent/WO2011124185A2/zh
Publication of WO2011124185A3 publication Critical patent/WO2011124185A3/zh

Links

Classifications

    • GPHYSICS
    • G11INFORMATION STORAGE
    • G11CSTATIC STORES
    • G11C29/00Checking stores for correct operation ; Subsequent repair; Testing stores during standby or offline operation
    • G11C29/04Detection or location of defective memory elements, e.g. cell constructio details, timing of test signals
    • G11C29/08Functional testing, e.g. testing during refresh, power-on self testing [POST] or distributed testing

Definitions

  • the present invention relates to the field of information technology, and in particular, to a memory detecting method and a memory detecting device. Background technique
  • the existing memory detection method performs write and readback detection on the memory in the order of increasing or decreasing logical addresses, however, the logical addresses are sequentially incremented or sequentially decremented during the process of writing and reading back, in the physical structure.
  • the adjacent cells may write the same data, and when the adjacent cell data of the physical structure are the same (that is, the charges are the same), the data changes are usually not affected, so the failed cells cannot be detected, resulting in The test results are inaccurate, which in turn affects the reliability of the memory and equipment. Summary of the invention
  • the embodiment of the invention provides a memory detection method and a memory detection device, which are used to solve the problem that the prior art cannot find the failed storage unit in time, thereby affecting the reliability of the memory and the device.
  • the embodiment of the invention provides a memory detection method, including:
  • a readback detection is performed on the basic unit that performs the write operation to determine the basic unit in which the written value is different from the readback value.
  • An embodiment of the present invention provides a memory detecting apparatus, including:
  • the memory detecting method and the memory detecting device provided by the embodiment of the present invention can perform a gap writing operation on the basic unit on the physical unit of the memory, and can perform a readback of the basic unit after performing the writing operation, and can timely find that there is a fault. Adjacent base units that affect each other, thereby improving the accuracy of memory detection and improving the reliability of memory and equipment.
  • FIG. 1 is a flowchart of a first embodiment of a memory detecting method provided by the present invention
  • FIG. 2a is a schematic diagram showing the execution sequence of the row-by-row write operation in the interval 2 columns provided by the present invention
  • FIG. 2b is a schematic diagram showing the data change of the physical unit of the memory in the interval 2 column of the row-by-row write operation provided by the present invention
  • FIG. 3a is a schematic diagram showing the execution sequence of the row-by-row write operation in the interval 4 columns provided by the present invention
  • FIG. 3b is a schematic diagram showing the data change of the physical unit of the memory in the row-by-row write operation of the interval 4 columns provided by the present invention
  • FIG. 4a is a schematic diagram showing the execution sequence of the row-by-row write operation in the interval 3 columns provided by the present invention
  • FIG. 4b is a schematic diagram showing the data change of the physical unit of the memory in the row-by-row write operation of the interval 3 columns provided by the present invention
  • FIG. 5 is a structural diagram of a first embodiment of a memory detecting apparatus according to the present invention.
  • FIG. 6 is a structural diagram of a first embodiment of a memory detecting apparatus provided by the present invention. detailed description
  • FIG. 1 is a flowchart of a first embodiment of a memory detecting method according to the present invention. As shown in FIG. 1, the method includes:
  • the memory involved in the memory detection method provided by the embodiment of the present invention may be a memory on a communication device such as a base station, a wireless network controller, a server, a core network device, or a memory on a terminal device such as a mobile phone or a computer.
  • a communication device such as a base station, a wireless network controller, a server, a core network device, or a memory on a terminal device such as a mobile phone or a computer.
  • the physical unit of the memory involved in the embodiment of the present invention may be the entire memory in the above various devices, or may be a physical storage body in the memory of the above various devices (the number of physical storage bodies is usually represented by BANK).
  • the basic unit is used to represent the smallest storage unit in the physical unit of memory, that is, the most basic storage structure in the physical unit of memory.
  • Each data unit stores a data message of 0 or 1 (ie, uncharged or fully charged), and the basic units are arranged in a matrix.
  • the memory detection method provided by the present invention performs a space write operation on the basic unit in the physical unit of the memory so that the data in the adjacent basic unit is different, since the failed basic unit is stored in the adjacent basic unit.
  • the basics of failure due to the failure of the basic unit The cell and the adjacent base cell store data are changed. Therefore, when the readback detection operation is performed after the write operation, the failed basic unit can be determined by whether the readback data and the write data are inconsistent.
  • the method for detecting the memory may perform a write 1 operation on the basic unit if the initial value of the basic unit is 0; if the initial value of the basic unit is 1, the write operation of the basic unit may be performed.
  • the interval write operation to the basic unit can be performed row by row, and the write operation is performed at a certain number of columns in each row.
  • the so-called row-by-row interval write operation can mean: writing to one basic unit in any row After the operation, skip the set number of columns and then perform the write operation until the number of basic units that have not been written at the end of the line is less than the set interval column number, and then perform the interval write operation on the next line;
  • the interval write operation to the basic unit can also be performed column by column. Specifically, after performing a write operation on one of the basic units in any column, the set number of rows is skipped and then the write operation is performed until the end of the column is not The number of basic units for the write operation is less than the number of set interval lines, and the next column is subjected to the interval write operation.
  • the number of columns in the interval may be n, n is greater than 0, and is less than or equal to the number of basic units in each row.
  • the interval column The number can be selected, 2 m is an integer greater than 0, and 2 is less than or equal to the number of basic units in each row.
  • the number of lines in the interval can also be n or 2 m .
  • one traversal means: the write operation starts from the first line until completion In the last row of write operations, at least half of the base cells in each row are not written. Therefore, a second traversal can be performed, and a basic write operation in which no write operation is performed in each row is performed until the write operation is completed for the basic cells in all the memory physical cells.
  • the number of traversal depends on the number of columns in the interval. For example: When the number of columns in the interval is 2, the writing operation to all the basic units in the physical unit of the memory can be completed after only two traversals. The larger the number of interval columns, the more traversal operations are required.
  • the physical unit of any size when the physical unit of any size is detected, if the write operation is performed row by row, the physical unit of the memory may be implemented by at least two traversals. All basic units are tested, and the physical addresses of the basic units detected during each traversal are continuous. Therefore, when performing memory detection, the entire memory can be divided into arbitrary block memory physical units according to the spatial structure (where each physical unit of memory can be a BANK or several BANKs), and the physical unit of the memory can be detected block by block (for Each memory physical unit performs at least two traversal to detect all the basic units in the physical unit of the block memory, thereby detecting all the basic units in the entire memory.
  • a readback operation can be performed immediately to detect whether the value in the basic unit is consistent with the written value, and after writing to at least one row or one column of the basic unit, Then perform a readback detection operation. If the value read back by a basic unit does not match the written value, the basic unit may be a failed basic unit, or the adjacent basic unit of the basic unit is a failed basic unit, and the basic unit is subjected to The effect of the adjacent failed basic unit results in a different readback value than the written value. That is, if the adjacent two basic units have different read value and write value, it is not determined which one of the basic units is invalid. Therefore, both basic units need to be recorded and subsequently Dimensional processing.
  • the memory usually includes several BANKs. If each BANK corresponds to 1 bit of data, it can be detected synchronously by 8 bits when detecting.
  • the data used for the detection can be: 0x00 ⁇ 0xff, 0x1 1 ⁇ 0xee, 0x22 ⁇ 0xdd 0x44 ⁇ Oxbb and 0x88 ⁇ 0x77; Synchronous detection can also be performed in groups of 16 bits.
  • the data used for detection can be: 0x0000 ⁇ 0xffff, 0x111 1 ⁇ 0xeeee, 0x2222 ⁇ 0xdddd, 0x4444 ⁇ Oxbbbb and 0x8888 ⁇ 0x7777.
  • synchronous detection can be performed in groups of 32 bits/64 bits, and the data used for the detection can be: 0x00 ⁇ 0xff, 0x11 ⁇ 0xee, 0x22 ⁇ 0xdd, 0x44 ⁇ Oxbb, and 0x88 ⁇ 0x77.
  • the memory detection method provided by the embodiment of the present invention performs interval write operation on the basic unit on the physical unit of the memory, so that the data in the adjacent basic unit is different, because the failed basic unit is stored in the adjacent basic unit.
  • the data is different, the failed basic unit and the adjacent basic unit may change data due to the failure of the basic unit. Therefore, when the basic unit after performing the write operation is read back, the data can be passed. Whether the readback data is consistent with the written data to determine the basic unit of failure, and timely discover the adjacent basic units that have faults and affect each other, thereby improving the accuracy of the memory detection and improving the reliability of the memory and the device.
  • the interval write operation may perform the write operation on the basic unit line by line by the interval setting column number until the traversal of the physical unit of the memory All basic units.
  • the start position of the first row in the physical unit of the memory is used as a starting point, and the number of columns is set to write the basic unit row by row.
  • the last row in the physical unit of the memory is written, the first row is returned.
  • the row performs a write operation, that is, after the last row is reached, the first row and the last row are connected to perform a write operation. Or, after writing to the last line, you can also use the next basic unit of the starting position as the starting point, and set the number of columns to write the basic unit line by line.
  • the specific offset may be determined according to the number of columns in the interval, for example: If the number of spaced columns If it is 2, it means that the next basic unit of the starting position has not been written and readback detection operation, then it can be detected by shifting a basic unit from the starting position; if the number of spacing columns is 3, the starting position is indicated. If the next two basic units are not subjected to the write operation and the readback detection operation, the detection may be started by shifting one or two basic units from the starting position, and so on, and the various cases are not enumerated.
  • the interval write operation may also perform a write operation on the basic unit column by column by the interval setting row number until all the basic units in the physical unit of the memory are traversed.
  • the column-by-column write can also start from the start position of the first column in the physical unit of the memory, and set the number of rows to write the base unit column by column; After the last column in the physical unit is written, the first column can be returned for writing.
  • the next basic unit at the starting position can be used as the starting point, and the number of rows can be set to write the basic unit column by column.
  • the number of set columns of the interval when performing row-by-line detection, may be n, n is greater than 0 and less than or equal to the number of basic units in each row.
  • Another implementation manner is as follows: Considering that when the detection operation is actually performed, it is usually required to perform addressing according to a logical address to know the actual physical address of the basic unit in the memory, and the logical address is usually binary, so that a better one is preferable.
  • the implementation may be: writing at intervals of 2 m columns, m being an integer greater than 0, and 2 m being less than or equal to the number of basic units in each row.
  • the number of set lines may be n, n is greater than 0 and less than or equal to the number of basic units in each column, or the number of set lines may be 2 m , and m is greater than or equal to 0. An integer, and 2 m is less than or equal to the number of basic units in each column.
  • a feasible manner can be performed immediately after performing a write operation on each basic unit, and another feasible manner can be performed after performing a write operation on at least one row or a column of basic units of the memory unit.
  • all the basic units in the physical unit of memory to be detected may be initialized before the write operation.
  • the write operation performed may be an inverse operation of the initialization operation, that is, if the basic unit is initialized to 0, the value written by the write operation may be 1, and if the basic unit is valid, the value read at the time of detection is read back. Also 1 , if the basic unit fails, the value read during the readback detection may become 0; if the basic unit is initialized to 1, the value written by the write operation may be 0, if the basic unit is valid, then The value read during the readback detection is also 0. If the basic unit fails, the value read during the readback detection may become 1.
  • the basic unit can be read to determine the read value and the initial value.
  • Different basic units Still taking a write operation line by line, before all the basic units in the physical unit of memory have been initialized but have not yet started the write operation, since all the basic units in the physical unit of the memory are initialized to 1 or 0, In this case, even if there is a failed basic unit in the physical unit of the memory, the charge (data) stored in each adjacent basic unit is generally the same, and the charge flow does not normally occur, and thus the basic unit does not usually appear. The situation in which data changes.
  • the first traversal operation can be directly performed (one traversal means: the writing operation is started from the first line until the writing operation of the last line is completed).
  • the next basic unit of the starting position of the first traversal operation can be used as the starting point, in this case, if the starting point of the first traversal operation is the basic unit and the second operation.
  • the value stored in the starting point of the second traversal should be the initialized value, and the starting point of the first traversal adjacent to the starting point of the second traversal The value stored in it should be the value of the write operation.
  • the values of the starting base unit of the first traversal operation and the starting base unit of the second operation will be A change has occurred. It can be understood that, besides the starting point of the second traversal, for the other basic units traversed the second time, there are also basic units adjacent thereto that have completed the writing operation in the first traversal. Accordingly, before the second traversal, the basic unit may be first read. If the read value is different from the initialized value, the read basic unit and its adjacent ones have been performed before the first traversal. There is a failed basic unit in the basic unit of the write operation.
  • FIG. 2a is a schematic diagram showing the execution sequence of the row-by-row write operation in intervals of 2 columns.
  • the memory physical unit for detecting in FIG. 2a is composed of 8 rows by 8 columns of basic units. Among them, 1, 2, 3, ..., 64 respectively indicate the execution order of the write operation.
  • r0 indicates the read operation before the write operation, and reads 1 bit of data stored in the base unit (if the base unit is valid, the value read by the read operation is 0, otherwise, the value read by the read operation is 0)
  • w1 indicates that 1 bit of data is written
  • r1 indicates Read 1 bit data (if the basic unit is valid, the data to be read back is 1, if the basic unit fails, the data to be read back is 0).
  • Figure 2b is a diagram showing the data changes of the physical unit of the memory in the row-by-row write operation of the interval 2 columns.
  • Fig. 3a is a diagram showing the execution sequence of the row-by-row write operation in intervals of 4 columns, wherein 1, 2, 3, ... 64 represent the execution order of the write operation.
  • the default is to initialize all the basic units in the physical unit of memory to 1.
  • r1-w0-r0 shown in Figure 3a r1 indicates the read operation before the write operation, and reads the 1 bit stored in the base unit. Data (if the basic unit is valid, the value read by the read operation is 1, otherwise, the value read by the read operation is 0), wO means write 1 bit data 0, r0 means read back 1 bit data (if basic If the unit is valid, the data to be read back is 0. If the basic unit fails, the data to be read back is 1).
  • Figure 3b is a diagram showing the data changes of the physical unit of the memory in the row-by-row write operation of the interval 4 columns.
  • FIG. 4a is a diagram showing the execution sequence of a row-by-row write operation in intervals of 3 columns, wherein 1, 2, 3, ... 64 represent the execution order of the write operation.
  • the default is to initialize all the basic units in the physical unit of memory to 1.
  • r1-w0-r0 shown in Figure 4a r1 indicates the read operation before the write operation, and reads the 1 bit stored in the base unit. Data (if the basic unit is valid, the read operation reads the value as 1, otherwise, the read operation reads the value as 0), wO means write 1 bit data 0, r0 means read 1 bit data (if basic If the unit is valid, the data read back is 0. If the basic unit fails, the data read back is 1).
  • FIG. 4b is a schematic diagram showing data changes of a physical unit of a memory in a row-by-row write operation in intervals of three columns.
  • the failed basic unit Since the failed basic unit is different from the adjacent basic unit storage data, the failed basic unit and the adjacent basic unit store data may be changed due to the failure of the basic unit, and therefore, the execution is performed on the pair.
  • the basic unit after the operation it is possible to determine the basic unit of the failure by the data read back and the written data, and to find the adjacent basic unit that has the fault and affect each other in time, thereby improving the accuracy of the memory detection.
  • Sexuality improves memory and device reliability.
  • the memory detection method provided by the embodiment of the present invention combines the row and column structure of the memory by performing a space write operation on the basic unit on the physical unit of the memory, so that the data in the adjacent basic unit is different, due to the failure of the basic unit.
  • the data of the failed basic unit and the adjacent basic unit may be changed due to the influence of the failed basic unit, and therefore, the basic after performing the write operation
  • the unit performs readback it can determine the basic unit of failure by whether the data read back is consistent with the written data.
  • the basic unit of the failure may be determined by the read operation before the write operation.
  • the physical address of the basic unit for performing the writing operation is continuous, and therefore, the storage unit can be relatively completed and detected, thereby improving the comprehensiveness and accuracy of the memory detection.
  • Improve storage particle failure screening rate and discover memory in time In the basic unit of effectiveness, the products with faulty memory are intercepted before shipment, which reduces maintenance costs and improves the reliability of memory and equipment.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).
  • FIG. 5 is a structural diagram of a first embodiment of a memory detecting apparatus according to the present invention. As shown in FIG. 5, the apparatus includes: a writing module 11 and a readback detecting module 12;
  • the write module 11 is configured to perform a write operation on the basic unit in the physical unit of the memory; and the read-back detection module 12 is configured to perform readback detection on the basic unit that performs the write operation to determine the write value and read back.
  • the memory detecting apparatus can check the memory on the communication device such as the base station, the wireless network controller, the server, the core network device, and the like, and can also detect the memory on the terminal device such as the mobile phone and the computer.
  • the physical unit of the memory involved in the embodiment of the present invention may be the entire memory in the above various devices, or may be a physical storage body in the memory of the above various devices (the number of physical storage bodies is usually represented by BANK).
  • the basic unit is used to represent the smallest storage unit in the physical unit of memory, that is, the most basic storage structure in the physical unit of memory.
  • Each data unit stores a data message of 0 or 1 (ie, uncharged or fully charged), and the basic units are arranged in a matrix.
  • the basic unit in the physical unit of the memory can be intermittently written by the writing module 11 so that the data in the adjacent basic unit is different, because the failed basic unit is in phase with If the neighboring basic unit stores different data, the failed basic unit and the adjacent basic unit may change data due to the failure of the basic unit. Therefore, after the write operation, the detection module may be read back. 12 when performing a readback detection operation, The basic unit of failure is determined by whether the data read back is inconsistent with the written data.
  • the writing module 11 can perform a write 1 operation on the basic unit; if the initial value of the basic unit is 1, the writing module 11 can perform a write 0 operation on the basic unit.
  • the write module 11 can perform interval write operations on the basic unit, and can perform row-by-row operations, and perform a write operation at a certain number of columns in each row.
  • the so-called row-by-row interval write operation can mean: one in any row After the basic unit performs the write operation, skip the set number of columns and then perform the write operation until the number of basic units that have not been written at the end of the line is less than the set interval column number, and then write the interval to the next line. Operation
  • the write module 1 1 can also perform a column-by-column write operation on the basic unit. Specifically, after performing a write operation on one of the basic units in any column, the set line number is skipped and then the write operation is performed until the At the end of the column, the number of basic units that have not been written is less than the number of set interval lines, and the next column is written to the interval.
  • the number of columns in the interval may be n, n is greater than 0, and is less than or equal to the number of basic units in each row.
  • the interval column The number can be selected, 2 m is an integer greater than 0, and 2 is less than or equal to the number of basic units in each row.
  • the number of lines in the interval can also be n or 2 m .
  • the readback detection module 12 can immediately perform a readback operation to detect whether the value in the basic unit is consistent with the written value, and the readback detection module 12 further The readback detection operation may be performed after the write operation is performed by the write module 11 on at least one row or one column of the basic cells. If the value read back by a basic unit does not match the written value, the basic unit may be a failed basic unit, or the adjacent basic unit of the basic unit is a failed basic unit, and the basic unit is subjected to The effect of the adjacent failed basic unit results in a different readback value than the written value. That is, if both adjacent basic units are read back In the case where the value is not the same as the written value, it is not possible to determine which of the basic units has failed. Therefore, both basic units need to be recorded and subjected to subsequent maintenance processing.
  • the memory detection device provided in this embodiment corresponds to the memory detection method provided by the embodiment of the present invention.
  • the memory detecting apparatus performs interval writing operation on the basic unit on the physical unit of the memory, so that the data in the adjacent basic unit is different, because the failed basic unit is stored in the adjacent basic unit.
  • the failed basic unit and the adjacent basic unit may change data due to the failure of the basic unit. Therefore, when the basic unit after performing the write operation is read back, the data can be passed. Whether the readback data is consistent with the written data to determine the basic unit of failure, and timely discover the adjacent basic units that have faults and affect each other, thereby improving the accuracy of the memory detection and improving the reliability of the memory and the device.
  • 6 is a structural diagram of a second embodiment of the memory detecting apparatus provided by the present invention. As shown in FIG. 6, the apparatus includes: a writing module 1 1 and a readback detecting module 12;
  • the writing module 11 may be specifically configured to: perform a write operation on the basic unit row by row until the basic unit is traversed in the physical unit of the memory; or Interval setting the number of rows to perform a write operation on the basic unit column by column until all the basic units in the physical unit of the memory are traversed.
  • the number of the set columns is n, n is greater than 0 and less than or equal to the number of basic units in each row, or the set number of columns is 2 m , m is an integer greater than 0, and 2 is less than or equal to each The number of basic units in the line;
  • the set number of rows is n, n is greater than 0 and less than or equal to the number of basic units in each column, or the set number of rows is 2 m , m is an integer greater than 0, and 2 m is less than or equal to each column The number of basic units in the middle.
  • the writing module 11 is further configured to: use the first row in the physical unit of the memory
  • the starting position is a starting point, and the interval setting column number performs a writing operation on the basic unit line by line; after writing to the last line in the memory physical unit, returning the first line to write Operation, or, starting from the next basic unit of the starting position, and setting the number of columns to perform the writing operation on the basic unit line by line;
  • the memory detecting apparatus may further include: an initializing module 13 configured to perform basic operations on the physical unit of the memory before the writing module 11 performs a space write operation on the basic unit in the physical unit of the memory The unit performs an initialization operation.
  • the write operation performed by the write module 11 is an inverse operation of the initialization operation performed by the initialization module 13.
  • the apparatus may further include: a reading module 14 configured to perform a reading operation on the basic unit after the initializing operation to determine a basic unit whose reading value is different from the initializing value.
  • a reading module 14 configured to perform a reading operation on the basic unit after the initializing operation to determine a basic unit whose reading value is different from the initializing value.
  • the readback detection module 12 may be specifically configured to: perform a readback detection operation immediately after performing a write operation on each of the basic units; or perform a write operation on the at least one row or a column of basic units. Thereafter, the basic unit that performs the write operation on the at least one row or column performs a readback detection operation.
  • the writing module 11 can perform a writing operation on the basic unit row by row by using the interval setting column number until all the basic units in the physical unit of the memory are traversed.
  • the writing module 11 may start from the starting position of the first row in the physical unit of the memory, and set the number of columns to write the basic unit row by row; when writing to the last row in the physical unit of the memory, Return to the first line for write operation, that is, after the last line is reached, the first line and the last line are connected to perform the write operation. Or, after writing to the last line, The base unit can be written to the base unit line by line with the next base unit of the start position as the starting point.
  • the specific offset may be determined according to the number of columns in the interval, for example: If the number of spaced columns If it is 2, it means that the next basic unit of the starting position has not been written and readback detection operation, then it can be detected by shifting a basic unit from the starting position; if the number of spacing columns is 3, the starting position is indicated. If the next two basic units are not subjected to the write operation and the readback detection operation, the detection may be started by shifting one or two basic units from the starting position, and so on, and the various cases are not enumerated.
  • the write module 11 can also write the base unit column by column by the number of rows, until all the basic units in the physical unit of the memory are traversed.
  • the start position of the first column in the physical unit of the memory may be used as a starting point, and the number of rows is set to write the basic unit column by column. Operation; After writing to the last column in the physical unit of memory, you can return to the first column for writing, or you can start with the next basic unit at the starting position, and set the number of rows by column. Write to the base unit.
  • the number of set columns of the interval may be n, n is greater than 0 and less than or equal to the number of basic cells in each row.
  • a preferred embodiment may be: The write module 11 performs a write operation with 2 m columns, m is an integer greater than 0, and 2 is less than or equal to the number of basic units in each row.
  • the number of set lines may be n, n is greater than 0 and less than or equal to the number of basic units in each column, or the number of set lines may be 2 m , and m is greater than or equal to 0. An integer, and 2 m is less than or equal to the number of basic units in each column.
  • the readback detection module 12 can be immediately read back after the write operation of the write operation by the write module 11 or after the write operation is performed by the write module 11 on at least one row or a column of the basic cells of the memory cell.
  • the memory physics to be detected by the initialization module 13 before the write operation is performed All basic units in the unit are initialized.
  • the write operation performed by the write module 11 may be an inverse operation of the initialization operation performed by the initialization module 13.
  • the reading operation of the basic unit by the reading module 14 may be performed to determine the read value and Initialize the basic unit with different values. For example: If the write module 11 performs a write-by-row operation, the base unit that needs to perform a write operation during these traversal operations before the write module 11 performs the second traversal, the third traversal, or the subsequent traversal , there are basic units of the completed write operation adjacent thereto, and therefore, the basic unit can be read by the reading module 14 before the writing module 11 performs the third traversal, the fourth traversal or more traversal A read operation is performed, and if the value read by the reading module 14 is different from the initialized value, there is a failed basic unit in the read basic unit and the adjacent basic unit in which the write operation has been performed.
  • the memory detection device provided in this embodiment corresponds to the memory detection method provided by the embodiment of the present invention.
  • DSP Digital Signal Processing
  • FPGA Field-Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • the memory detecting apparatus combines the row and column structure of the memory by performing a space write operation on the basic unit on the physical unit of the memory, so that the data in the adjacent basic unit is different, due to the failure of the basic unit.
  • the data is stored differently from the adjacent basic unit
  • the data of the failed basic unit and the adjacent basic unit may be changed due to the influence of the failed basic unit, and therefore, the basic after performing the write operation
  • the unit can determine the basic unit of failure by whether the data read back is consistent with the written data.
  • the third traversal or the subsequent traversal before the write operation, you can also read The fetch operation determines the basic unit of failure.
  • the physical address of the basic unit for performing the writing operation is continuous, and therefore, the storage unit can be relatively completed and detected, thereby improving the comprehensiveness and accuracy of the memory detection. It can improve the storage particle failure screening rate, and timely discover the effective basic unit in the memory, so as to intercept the products with fault hidden memory before shipment, reduce maintenance costs, and thus improve the reliability of memory and equipment.

Abstract

一种内存检测方法和内存检测装置。方法包括以下步骤:对内存物理单元中的基本单元进行间隔写入操作(S101);对进行写入操作的基本单元进行回读检测,以确定写入数值与回读数值不同的基本单元(S102)。本内存检测方法和内存检测装置,通过对内存物理单元上的基本单元进行间隔写入操作,在对执行写入操作后的基本单元进行回读时,能够及时发现存在故障而互相影响的相邻基本单元,从而提高了内存检测的准确性,提高了内存和设备的可靠性。

Description

内存检测方法和内存检测装置 技术领域
本发明涉及信息技术领域,特别涉及一种内存检测方法和内存检测装置。 背景技术
随着存储芯片集成度的不断提高, 芯片内部存储单元间相互干扰问题变 得越来越突出, 而内存是否可靠对设备的可靠性起到至关重要的作用。
现有的内存检测方法, 按照逻辑地址依次递增或依次递减的顺序, 对内 存进行写入和回读检测, 然而, 逻辑地址依次递增或依次递减进行写入和回 读的过程中, 在物理结构上相邻的单元可能写入相同的数据, 而当物理结构 上相邻的单元数据相同 (即电荷相同) 时, 通常不会相互影响进而产生数据 变化, 因此无法检测到其中的失效单元, 导致检测结果不准确, 进而影响到 内存和设备的可靠性。 发明内容
本发明实施例提供了一种内存检测方法和内存检测装置, 用以解决现有 技术无法及时发现失效的存储单元, 进而影响到内存和设备可靠性的问题。
本发明实施例提供一种内存检测方法, 包括:
对内存物理单元中的基本单元进行间隔写入操作;
对进行写入操作的基本单元进行回读检测, 以确定写入数值与回读数值 不同的基本单元。
本发明实施例提供一种内存检测装置, 包括:
写入模块, 用于对内存物理单元中的基本单元进行间隔写入操作; 回读检测模块, 用于对进行写入操作的基本单元进行回读检测, 以确定 写入数值与回读数值不同的基本单元。 本发明实施例提供的内存检测方法和内存检测装置, 通过对内存物理单 元上的基本单元进行间隔写入操作, 在对执行写入操作后的基本单元进行回 读时, 能够及时发现存在故障而互相影响的相邻基本单元, 从而提高了内存 检测的准确性, 提高了内存和设备的可靠性。 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案, 下面将对实 施例或现有技术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面 描述中的附图仅仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付出创造性劳动的前提下, 还可以根据这些附图获得其他的附图。
图 1为本发明提供的内存检测方法第一实施例的流程图;
图 2a为本发明提供的间隔 2列进行逐行写入操作的执行顺序示意图; 图 2b为本发明提供的间隔 2列进行逐行写入操作内存物理单元的数据变 化示意图;
图 3a为本发明提供的间隔 4列进行逐行写入操作的执行顺序示意图; 图 3b为本发明提供的间隔 4列进行逐行写入操作内存物理单元的数据变 化示意图;
图 4a为本发明提供的间隔 3列进行逐行写入操作的执行顺序示意图; 图 4b为本发明提供的间隔 3列进行逐行写入操作内存物理单元的数据变 化示意图;
图 5为本发明提供的内存检测装置第一实施例的结构图;
图 6为本发明提供的内存检测装置第一实施例的结构图。 具体实施方式
下面将结合本发明实施例中的附图, 对本发明实施例中的技术方案进行 清楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明一部分实施例, 而 不是全部的实施例。 基于本发明中的实施例, 本领域普通技术人员在没有做 出创造性劳动前提下所获得的所有其他实施例, 都属于本发明保护的范围。
图 1为本发明提供的内存检测方法第一实施例的流程图, 如图 1所示, 该 方法包括:
S101、 对内存物理单元中的基本单元进行间隔写入操作;
S102、 对进行写入操作的基本单元进行回读检测, 以确定写入数值与回 读数值不同的基本单元。
本发明实施例提供的内存检测方法中涉及的内存, 可以是基站、 无线网 络控制器、 服务器、 核心网设备等通信设备上的内存, 还可以是手机、 计算 机等终端设备上的内存。
本发明实施例中涉及的内存物理单元可以是上述各种设备中的整个内 存, 也可以是上述各种设备的内存中的某一块物理存储体(物理存储体的数 量通常以 BANK表示) 。 基本单元用来表示内存物理单元中的最小存储单元, 即内存物理单元中最基本的存储结构。 每个基本单元中存储一个数据信息 0 或 1 (即未充电荷或充满电荷) , 基本单元呈矩阵状排列。
如果内存物理单元存在设计缺陷, 则会导致某些基本单元失效, 而失效 的基本单元会影响或干扰相邻的基本单元, 失效的基本单元及其相邻的受到 影响的基本单元内的电荷会发生变化, 即失效的基本单元及其相邻的受到影 响的基本单元内存储的数据可能从 1变成 0, 或者从 0变成 1 , 由此引发内存数 据错误。 然而, 同行或同列的相邻基本单元, 如果其数值均为 0或均为 1 (即 都未充电荷或充满电荷)时, 通常不会相互影响而导致存储的数据发生变化。 而如果同行或同列相邻的基本单元一个数据为 1 , 另一个数据为 0, 则可能会 出现数据为 1的基本单元上的电荷流向数据为 0的基本单元的情况。 据此, 本 发明提供的内存检测方法,对内存物理单元中的基本单元进行间隔写入操作, 以使相邻的基本单元中的数据不同, 由于失效的基本单元在与相邻的基本单 元存储数据不同的情况下, 会由于失效的基本单元的影响而导致失效的基本 单元和相邻的基本单元存储数据发生变化, 因此, 在写入操作之后再进行回 读检测操作时, 可以通过回读的数据与写入数据是否不一致来确定失效的基 本单元。
本发明实施例提供的内存的检测方法, 如果基本单元的初始值为 0, 则可 以对基本单元进行写 1操作; 如果基本单元的初始值为 1 , 则可以对基本单元 进行写 0操作。
对基本单元进行间隔写入操作可以逐行进行, 在每行中间隔一定列数进 行写入操作, 所谓逐行进行间隔写入操作可以是指: 在对任一行中的一个基 本单元进行写入操作后, 跳过设定列数再进行写入操作, 直到该行末尾未进 行写入操作的基本单元个数小于设定间隔列数为止, 再对下一行进行间隔写 入操作;
对基本单元进行间隔写入操作还可以逐列进行, 具体可以是: 在对任一 列中的一个基本单元进行写入操作后, 跳过设定行数再进行写入操作, 直到 该列末尾未进行写入操作的基本单元个数小于设定间隔行数为止, 再对下一 列进行间隔写入操作。
以逐行进行写入操作为例, 间隔的列数可以为 n , n大于 0, 且小于等于每 行中基本单元的个数。 还有另外一种实现方式: 考虑到在实际进行检测操作 时, 通常需要根据逻辑地址进行寻址获知基本单元在内存中的实际的物理位 置, 而逻辑地址通常是二进制的, 因此, 间隔的列数可以选取, 2m为大于 0 的整数, 且2 小于等于每行中基本单元的个数。 逐列进行写入操作时, 间隔 的行数也可以为 n或 2m
仍以逐行进行写入操作为例, 由于在每行中进行的是间隔写入操作, 因 此当第一次遍历所有行之后 (一次遍历是指: 从第一行开始进行写入操作直 至完成最后一行的写入操作) , 每行中仍有至少一半的基本单元没有进行写 入操作。 因此, 可以进行第二次遍历, 对每行中没有进行写入操作的基本单 元进行间隔写入操作,直至对所有内存物理单元中的基本单元完成写入操作。 需要说明的是, 遍历的次数取决于间隔的列数, 例如: 当间隔的列数为 2时, 只需经过两次遍历便可以完成对内存物理单元中的所有基本单元的写入操 作。 间隔列数越大, 所需的遍历操作次数越多。
可以看出, 本发明实施例提供的内存检测方法中, 在对任意大小的内存 物理单元进行检测时, 如果逐行进行写入操作, 则可以通过至少两次遍历实 现对该内存物理单元中的所有基本单元进行检测, 而在每一次遍历过程中检 测的基本单元的物理地址都是连续的。 因此, 在进行内存检测时, 可以将整 个内存按照空间结构划分为任意块内存物理单元(其中每一块内存物理单元 可以是一个 BANK或几个 BANK ) , 可以逐块对内存物理单元进行检测(对于 每块内存物理单元进行至少两次遍历以实现对该块内存物理单元中所有基本 单元进行检测) , 从而实现对整个内存中的所有基本单元进行检测。
在对每个基本单元进行写入操作之后, 可以立即进行回读操作, 以检测 基本单元中的数值是否与写入的值一致, 还可以在对至少一行或一列基本单 元进行写入操作之后, 再进行回读检测操作。 如果某一基本单元回读出的数 值与写入的数值不一致, 则该基本单元可能为失效的基本单元, 或者是, 该 基本单元的相邻基本单元为失效的基本单元, 该基本单元由于受到相邻的失 效基本单元的影响而导致回读数值与写入数值不相同。 即, 如果相邻两个基 本单元均出现回读数值与写入数值不相同的情况, 并不能确定是其中哪一个 基本单元失效, 因此, 需要将这两个基本单元都进行记录并且进行后续的维 爹处理。
内存中通常包括数个 BANK, 如果每个 BANK对应 1 bit数据, 则在进行检 测时,可以 8bit为一组同步检测,检测采用的数据可以为: 0x00 ~ 0xff、 0x1 1 ~ 0xee、 0x22 ~ 0xdd、 0x44 ~ Oxbb以及 0x88 ~ 0x77; 还可以 16bit为一组进行 同步检测, 检测采用的数据可以为: 0x0000 ~ 0xffff、 0x111 1 ~ 0xeeee、 0x2222 ~ 0xdddd、 0x4444 ~ Oxbbbb以及 0x8888 ~ 0x7777。 或者, 可以以 32bit/64bit为一组进行同步检测,检测采用的数据可以为: 0x00 ~ 0xff、 0x11 ~ 0xee、 0x22 ~ 0xdd、 0x44 ~ Oxbb以及 0x88 ~ 0x77。
本发明实施例提供的内存检测方法, 通过对内存物理单元上的基本单元 进行间隔写入操作, 以使相邻的基本单元中的数据不同, 由于失效的基本单 元在与相邻的基本单元存储数据不同的情况下, 会由于失效的基本单元的影 响而导致失效的基本单元和相邻的基本单元存储数据发生变化, 因此, 在对 执行写入操作后的基本单元进行回读时, 能够通过回读的数据与写入数据是 否一致来确定失效的基本单元, 及时发现存在故障而互相影响的相邻基本单 元, 从而提高了内存检测的准确性, 提高了内存和设备的可靠性。 本发明提供的内存检测方法第二实施例, 作为一种可选的实施方式, 进 行间隔写入操作可以采用间隔设定列数逐行对基本单元进行写入操作, 直至 遍历内存物理单元中的全部基本单元。
具体可以以内存物理单元中第一行的起始位置为起点, 间隔设定列数逐 行对基本单元进行写入操作; 当对内存物理单元中的最后一行进行写入操作 后, 返回第一行进行写入操作, 即, 当到达最后一行后, 将第一行与最后一 行接续起来进行写入操作。 或者, 在对最后一行进行写入操作后, 还可以以 起始位置的下一个基本单元为起点, 间隔设定列数逐行对基本单元进行写入 操作。 即, 当到达最后一行后, 重新回到第一行, 从原来的起始位置偏移至 少一个基本单元开始检测, 具体的偏移量可以根据间隔的列数而定, 例如: 如果间隔列数为 2, 则说明起始位置的下一个基本单元尚未进行写入操作和 回读检测操作, 则可以从起始位置偏移一个基本单元开始检测; 如果间隔列 数为 3, 则说明起始位置的下两个基本单元均未进行写入操作和回读检测操 作, 则可以从起始位置偏移一个或 2个基本单元开始检测, 依次类推, 各种 情况不一一列举。
作为另一种可选的实施方式, 进行间隔写入操作还可以采用间隔设定行 数逐列对基本单元进行写入操作,直至遍历内存物理单元中的全部基本单元。 与逐行进行写入操作类似的, 逐列写入时也可以以内存物理单元中第一 列的起始位置为起点, 间隔设定行数逐列对基本单元进行写入操作; 当对内 存物理单元中的最后一列进行写入操作后, 可以返回第一列进行写入操作, 或者, 还可以以起始位置的下一个基本单元为起点, 间隔设定行数逐列对基 本单元进行写入操作。
本实施例提供的内存检测方法中, 在进行逐行检测时, 间隔的设定列数 可以为 n , n大于 0且小于等于每行中基本单元的个数。 另一种实现方式为: 考虑到在实际进行检测操作时, 通常需要根据逻辑地址进行寻址获知基本单 元在内存中的实际的物理地址, 而逻辑地址通常是二进制的, 因此, 一种较 佳的实施方式可以是: 间隔 2m列进行写入操作, m为大于 0的整数, 且 2m 小于等于每行中基本单元的个数。
类似的, 在进行逐列检测时, 设定行数可以为 n , n大于 0且小于等于每 列中基本单元的个数, 或者, 设定行数可以为 2m, m为大于等于 0的整数, 且 2m小于等于每列中基本单元的个数。
进行回读检测操作, 一种可行的方式可以在对每个基本单元进行写入操 作后立刻进行, 另一种可行的方式可以在对存储单元的至少一行或一列基本 单元进行写入操作后进行。
另外, 在进行写入操作前, 可以首先对待检测的内存物理单元中的全部 基本单元进行初始化操作。 进行的写入操作可为初始化操作的取反操作, 即, 如果将基本单元初始化为 0, 则写入操作写入的数值可以为 1 ,如果基本单元 有效, 则回读检测时读取的数值也为 1 , 如果基本单元失效, 则回读检测时 读取的数值可能会变为 0; 如果将基本单元初始化为 1 , 则写入操作写入的数 值可以为 0, 如果基本单元有效, 则回读检测时读取的数值也为 0, 如果基本 单元失效, 则回读检测时读取的数值可能会变为 1。
其中, 在对基本单元进行初始化操作之后, 并且在对基本单元进行写入 操作之前, 还可以对基本单元进行读取操作, 以确定读取数值与初始化数值 不同的基本单元。 仍以逐行进行写入操作为例, 在内存物理单元中的所有基 本单元都完成初始化但尚未开始写入操作之前, 由于内存物理单元中的所有 基本单元都被初始化为 1 或 0, 在这种情况下, 即使内存物理单元中存在失 效的基本单元, 也会由于各相邻的基本单元中存储的电荷(数据)都相同而 通常不会出现电荷流动的情况, 进而通常不会出现基本单元中数据变化的情 况。 因此, 在完成基本单元的初始化操作后, 可以直接进行第一次遍历操作 (一次遍历是指: 从第一行开始进行写入操作直至完成最后一行的写入操 作)。 而在进行第二次遍历时, 由于可以以第一次遍历操作的起始位置的下一 个基本单元为起点, 在这种情况下, 如果第一次遍历操作的起点基本单元和 第二次操作的起点基本单元均有效的情况下, 进行第二次遍历之前, 第二次 遍历的起点中存储的数值应为初始化的数值, 而与第二次遍历的起点相邻的 第一次遍历的起点中存储的数值应为写入操作的数值。 但如果第一次遍历操 作的起点基本单元和第二次操作的起点基本单元中存在失效的基本单元, 则 第一次遍历操作的起点基本单元和第二次操作的起点基本单元存储的数值都 会发生变化。 可以理解的是, 除了第二次遍历的起点之外, 对于第二次遍历 的其他基本单元, 也存在与之相邻的已在第一次遍历中完成写入操作的基本 单元。 据此, 在进行第二次遍历之前, 可以首先对基本单元进行读取操作, 如果读取的数值与初始化的数值不同, 则读取的基本单元与其相邻的已在第 一次遍历之前进行写入操作的基本单元中存在失效的基本单元。
对于逐行进行写入操作, 当间隔的列数大于 2时, 需要经过两次以上遍 历来完成对内存物理单元中的所有基本单元的写入操作。 间隔列数越大, 所 需的遍历操作次数越多。 可以理解的是, 当需要进行第三次遍历、 第四次遍 历或者更多次遍历时, 在进行遍历之前, 对于需要在这些遍历过程中进行写 入操作的基本单元, 均存在与之相邻的已完成写入操作的基本单元, 因此, 在需要进行第三次遍历、 第四次遍历或者更多次遍历之前, 可以首先对基本 单元进行读取操作, 如果读取的数值与初始化的数值不同, 则读取的基本单 元与其相邻的已进行写入操作的基本单元中存在失效的基本单元。 图 2a所示为间隔 2列进行逐行写入操作的执行顺序示意图, 图 2a中进 行检测的内存物理单元由 8行乘以 8列个基本单元组成。 其中, 1、 2、 3... ... 64分别表示进行写入操作的执行顺序。
其中, 默认为将该 8行乘以 8列的内存物理单元中的所有基本单元初始 化为 0, 图 2a所示的 r0-w1 -r1 中, r0表示写入操作之前的读取操作, 读取 基本单元中存储的 1 bit数据(如果基本单元有效, 则读取操作读取的数值为 0, 否则, 读取操作读取的数值为 0 ), w1表示写入 1 bit数据 1 , r1表示回读 1 bit数据(如果基本单元有效, 则回读的数据为 1 , 如果基本单元失效, 则回 读的数据为 0 )。 图 2b为间隔 2列进行逐行写入操作内存物理单元的数据变 化示意图。
图 3a所示为间隔 4列进行逐行写入操作的执行顺序示意图, 其中, 1、 2、 3... ...64表示进行写入操作的执行顺序。 图 3a中默认为将内存物理单元 中的所有基本单元初始化为 1 , 图 3a所示的 r1-w0-r0中, r1表示写入操作 之前的读取操作, 读取基本单元中存储的 1 bit数据(如果基本单元有效, 则 读取操作读取的数值为 1 , 否则, 读取操作读取的数值为 0 ), wO表示写入 1 bit数据 0, r0表示回读 1 bit数据 (如果基本单元有效, 则回读的数据为 0, 如果基本单元失效, 则回读的数据为 1 )。 图 3b为间隔 4列进行逐行写入操 作内存物理单元的数据变化示意图。
图 4a所示为间隔 3列进行逐行写入操作的执行顺序示意图, 其中, 1、 2、 3... ...64表示进行写入操作的执行顺序。 图 4a中默认为将内存物理单元 中的所有基本单元初始化为 1 , 图 4a所示的 r1-w0-r0中, r1表示写入操作 之前的读取操作, 读取基本单元中存储的 1 bit数据(如果基本单元有效, 则 读取操作读取的数值为 1 , 否则, 读取操作读取的数值为 0 ), wO表示写入 1 bit数据 0, r0表示读取 1 bit数据(如果基本单元有效, 则回读的数据为 0, 如果基本单元失效, 则回读的数据为 1 )。 图 4b为间隔 3列进行逐行写入操 作内存物理单元的数据变化示意图。
需要说明的是, 从图 2b、 图 3b和图 4b中可以看出, 在以任意一种间隔 进行逐行或逐列写入操作时, 在进行最后一次遍历操作时, 在每行或每列中, 尚未进行写入操作的基本单元,其相邻的两个基本单元都已进行了写入操作。 而在这种情况下, 可以直接通过写入操作之前的读取操作便能够获知该尚未 进行写入操作的基本单元是否受到相邻基本单元的干扰, 或者是否影响相邻 的基本单元而出现数据错误。 如果读取操作中读取的数值正确, 则可进一步 进行写入操作及回读检测操作。
由于失效的基本单元在与相邻的基本单元存储数据不同的情况下, 会由 于失效的基本单元的影响而导致失效的基本单元和相邻的基本单元存储数据 发生变化, 因此, 在对执行写入操作后的基本单元进行回读时, 能够通过回 读的数据与写入数据是否一致来确定失效的基本单元, 及时发现存在故障而 互相影响的相邻基本单元, 从而提高了内存检测的准确性, 提高了内存和设 备的可靠性。
本发明实施例提供的内存检测方法, 结合内存的行、 列结构, 通过对内 存物理单元上的基本单元进行间隔写入操作, 以使相邻的基本单元中的数据 不同, 由于失效的基本单元在与相邻的基本单元存储数据不同的情况下, 会 由于失效的基本单元的影响而导致失效的基本单元和相邻的基本单元存储数 据发生变化, 因此, 在对执行写入操作后的基本单元进行回读时, 能够通过 回读的数据与写入数据是否一致来确定失效的基本单元。 另外, 还第二次遍 历、 第三次遍历或后续的遍历过程中, 在进行写入操作之前, 还可以通过读 取操作确定失效的基本单元。
本发明实施例提供的内存检测方法, 进行写入操作的基本单元的物理地 址是连续的, 因此, 能过对存储单元进行一次较为完成的检测, 因此, 提高 了内存检测的全面和准确性, 能够提升存储颗粒故障筛选率, 及时发现内存 中的实效基本单元, 从而把含故障隐患内存类的产品拦截在发货前, 降低维 护成本, 进而提高了内存和设备的可靠性。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流 程, 是可以通过计算机程序来指令相关的硬件来完成, 所述的程序可存储于 一计算机可读取存储介质中, 该程序在执行时, 可包括如上述各方法的实施 例的流程。 其中, 所述的存储介质可为磁碟、 光盘、 只读存储记忆体 ( Read-Only Memory, ROM )或随机存者 i己忆体( Random Access Memory, RAM )等。
图 5为本发明提供的内存检测装置第一实施例的结构图, 如图 5所示, 该装置包括: 写入模块 11和回读检测模块 12;
写入模块 11 , 用于对内存物理单元中的基本单元进行间隔写入操作; 回读检测模块 12, 用于对进行写入操作的基本单元进行回读检测, 以确 定写入数值与回读数值不同的基本单元。
本发明实施例提供的内存检测装置, 可以对基站、 无线网络控制器、 服 务器、 核心网设备等通信设备上的内存进行检查, 还可以对手机、 计算机等 终端设备上的内存进行检测。
本发明实施例中涉及的内存物理单元可以是上述各种设备中的整个内 存, 也可以是上述各种设备的内存中的某一块物理存储体(物理存储体的数 量通常以 BANK表示) 。 基本单元用来表示内存物理单元中的最小存储单元, 即内存物理单元中最基本的存储结构。 每个基本单元中存储一个数据信息 0 或 1 (即未充电荷或充满电荷) , 基本单元呈矩阵状排列。
本实施例提供的内存检测装置中, 可以通过写入模块 11对内存物理单元 中的基本单元进行间隔写入操作, 以使相邻的基本单元中的数据不同, 由于 失效的基本单元在与相邻的基本单元存储数据不同的情况下, 会由于失效的 基本单元的影响而导致失效的基本单元和相邻的基本单元存储数据发生变 化, 因此, 在写入操作之后, 可以通过回读检测模块 12进行回读检测操作时, 通过回读的数据与写入数据是否不一致来确定失效的基本单元。
其中,如果基本单元的初始值为 0,则写入模块 11可以对基本单元进行写 1操作; 如果基本单元的初始值为 1 , 则写入模块 11可以对基本单元进行写 0 操作。
写入模块 11可以对基本单元进行间隔写入操作可以逐行进行, 在每行中 间隔一定列数进行写入操作, 所谓逐行进行间隔写入操作可以是指: 在对任 一行中的一个基本单元进行写入操作后, 跳过设定列数再进行写入操作, 直 到该行末尾未进行写入操作的基本单元个数小于设定间隔列数为止, 再对下 一行进行间隔写入操作;
写入模块 1 1还可以对基本单元进行逐列写入操作, 具体可以是: 在对任 一列中的一个基本单元进行写入操作后, 跳过设定行数再进行写入操作, 直 到该列末尾未进行写入操作的基本单元个数小于设定间隔行数为止, 再对下 一列进行间隔写入操作。
以逐行进行写入操作为例, 间隔的列数可以为 n , n大于 0, 且小于等于每 行中基本单元的个数。 还有另外一种实现方式: 考虑到在实际进行检测操作 时, 通常需要根据逻辑地址进行寻址获知基本单元在内存中的实际的物理位 置, 而逻辑地址通常是二进制的, 因此, 间隔的列数可以选取, 2m为大于 0 的整数, 且2 小于等于每行中基本单元的个数。 逐列进行写入操作时, 间隔 的行数也可以为 n或 2m
在写入模块 1 1对每个基本单元进行写入操作之后, 回读检测模块 12可以 立即进行回读操作, 以检测基本单元中的数值是否与写入的值一致, 回读检 测模块 12还可以在写入模块 11对至少一行或一列基本单元进行写入操作之 后, 再进行回读检测操作。 如果某一基本单元回读出的数值与写入的数值不 一致, 则该基本单元可能为失效的基本单元, 或者是, 该基本单元的相邻基 本单元为失效的基本单元, 该基本单元由于受到相邻的失效基本单元的影响 而导致回读数值与写入数值不相同。 即, 如果相邻两个基本单元均出现回读 数值与写入数值不相同的情况, 并不能确定是其中哪一个基本单元失效, 因 此, 需要将这两个基本单元都进行记录并且进行后续的维修处理。
本实施例提供的内存检测装置与本发明实施例提供的内存检测方法相对 应, 为执行内存检测方法的功能设备, 其具体的执行过程可参见本发明提供 的内存检测方法实施例。
本发明实施例提供的内存检测装置, 通过对内存物理单元上的基本单元 进行间隔写入操作, 以使相邻的基本单元中的数据不同, 由于失效的基本单 元在与相邻的基本单元存储数据不同的情况下, 会由于失效的基本单元的影 响而导致失效的基本单元和相邻的基本单元存储数据发生变化, 因此, 在对 执行写入操作后的基本单元进行回读时, 能够通过回读的数据与写入数据是 否一致来确定失效的基本单元, 及时发现存在故障而互相影响的相邻基本单 元, 从而提高了内存检测的准确性, 提高了内存和设备的可靠性。 图 6为本发明提供的内存检测装置第二实施例的结构图, 如图 6所示, 该装置包括: 写入模块 1 1和回读检测模块 12;
作为一个较佳的实施例, 写入模块 1 1可以具体用于: 间隔设定列数逐行 对所述基本单元进行写入操作, 直至遍历所述内存物理单元中的全部基本单 元; 或者, 间隔设定行数逐列对所述基本单元进行写入操作, 直至遍历所述 内存物理单元中的全部基本单元。
其中, 所述设定列数为 n , n大于 0且小于等于每行中基本单元的个数, 或者, 所述设定列数为 2m, m为大于 0的整数, 且2 小于等于每行中基本 单元的个数;
所述设定行数为 n , n大于 0且小于等于每列中基本单元的个数, 或者, 所述设定行数为 2m, m为大于 0的整数, 且 2m小于等于每列中基本单元的 个数。
进一步的, 写入模块 11还可以具体用于: 以所述内存物理单元中第一行 的起始位置为起点, 间隔设定列数逐行对所述基本单元进行写入操作; 当对所述内存物理单元中的最后一行进行写入操作后, 返回所述第一行 进行写入操作, 或者, 以所述起始位置的下一个基本单元为起点, 间隔设定 列数逐行对所述基本单元进行写入操作;
或者, 以所述内存物理单元中第一列的起始位置为起点, 间隔设定行数 逐列对所述基本单元进行写入操作;
当对所述内存物理单元中的最后一列进行写入操作后, 返回所述第一列 进行写入操作, 或者, 以所述起始位置的下一个基本单元为起点, 间隔设定 行数逐列对所述基本单元进行写入操作。
本实施例提供的内存检测装置, 还可以包括: 初始化模块 13, 用于在所 述写入模块 11对内存物理单元中的基本单元进行间隔写入操作之前,对所述 内存物理单元中的基本单元进行初始化操作。
其中, 写入模块 11执行的写入操作为所述初始化模块 13执行的初始化 操作的取反操作。
进一步的, 该装置还可以包括: 读取模块 14, 用于对初始化操作后的基 本单元进行读取操作, 以确定读取数值与初始化数值不同的基本单元。
进一步的, 所述回读检测模块 12可以具体用于: 对每个所述基本单元进 行写入操作后, 立刻进行回读检测操作; 或者, 对所述至少一行或一列基本 单元进行写入操作后, 对所述至少一行或一列进行写入操作的基本单元进行 回读检测操作。
具体的,写入模块 11可以采用间隔设定列数逐行对基本单元进行写入操 作, 直至遍历内存物理单元中的全部基本单元。
写入模块 11可以以内存物理单元中第一行的起始位置为起点,间隔设定 列数逐行对基本单元进行写入操作; 当对内存物理单元中的最后一行进行写 入操作后, 返回第一行进行写入操作, 即, 当到达最后一行后, 将第一行与 最后一行接续起来进行写入操作。 或者, 在对最后一行进行写入操作后, 还 可以以起始位置的下一个基本单元为起点, 间隔设定列数逐行对基本单元进 行写入操作。 即, 当到达最后一行后, 重新回到第一行, 从原来的起始位置 偏移至少一个基本单元开始检测, 具体的偏移量可以根据间隔的列数而定, 例如: 如果间隔列数为 2, 则说明起始位置的下一个基本单元尚未进行写入 操作和回读检测操作, 则可以从起始位置偏移一个基本单元开始检测; 如果 间隔列数为 3, 则说明起始位置的下两个基本单元均未进行写入操作和回读 检测操作, 则可以从起始位置偏移一个或 2个基本单元开始检测,依次类推, 各种情况不一一列举。
写入模块 11还可以采用间隔设定行数逐列对基本单元进行写入操作,直 至遍历内存物理单元中的全部基本单元。
与逐行进行写入操作类似的,写入模块 11进行逐列写入时也可以以内存 物理单元中第一列的起始位置为起点, 间隔设定行数逐列对基本单元进行写 入操作; 当对内存物理单元中的最后一列进行写入操作后, 可以返回第一列 进行写入操作, 或者, 还可以以起始位置的下一个基本单元为起点, 间隔设 定行数逐列对基本单元进行写入操作。
写入模块 11逐行进行写入操作时, 间隔的设定列数可以为 n, n大于 0 且小于等于每行中基本单元的个数。 而一种较佳的实施方式可以是: 写入模 块 11间隔 2m列进行写入操作, m为大于 0的整数,且2 小于等于每行中基 本单元的个数。
类似的, 在进行逐列检测时, 设定行数可以为 n, n大于 0且小于等于每 列中基本单元的个数, 或者, 设定行数可以为 2m, m为大于等于 0的整数, 且 2m小于等于每列中基本单元的个数。
回读检测模块 12可以在写入模块 11对每个基本单元进行写入操作后立 刻进行回读,或者是在写入模块 11对存储单元的至少一行或一列基本单元进 行写入操作后进行。
另外,在进行写入操作前, 可以通过初始化模块 13对待检测的内存物理 单元中的全部基本单元进行初始化操作。写入模块 11进行的写入操作可为初 始化模块 13进行初始化操作的取反操作。
其中,在初始化模块 13对基本单元进行初始化操作之后, 并且在写入模 块 11对基本单元进行写入操作之前, 还可以通过读取模块 14对基本单元进 行读取操作, 以确定读取数值与初始化数值不同的基本单元。 例如: 如果写 入模块 1 1进行逐行写入操作, 则在写入模块 11进行第二次遍历、 第三次遍 历或后续遍历之前, 对于需要在这些遍历过程中进行写入操作的基本单元, 均存在与之相邻的已完成写入操作的基本单元, 因此, 在写入模块 11进行第 三次遍历、第四次遍历或者更多次遍历之前, 可以通过读取模块 14对基本单 元进行读取操作,如果读取模块 14读取的数值与初始化的数值不同, 则读取 的基本单元与其相邻的已进行写入操作的基本单元中存在失效的基本单元。
本实施例提供的内存检测装置与本发明实施例提供的内存检测方法相对 应, 为执行内存检测方法的功能设备, 其具体的执行过程可参见本发明提供 的内存检测方法实施例。
本发明实施例提供的内存检测装置, 其功能可以集成到数字信号处理
( Digital Signal Processing; 简称: DSP ) 芯片或现场可编程门阵列 ( Field-Programmable Gate Array; 简称: FPGA ) 芯片上进行软件开发来 实现, 也可以通过专用集成电路 ( Application Specific Integrated Circuit; 简 称: ASIC )或其他硬件电路的方式固化实现。
本发明实施例提供的内存检测装置, 结合内存的行、 列结构, 通过对内 存物理单元上的基本单元进行间隔写入操作, 以使相邻的基本单元中的数据 不同, 由于失效的基本单元在与相邻的基本单元存储数据不同的情况下, 会 由于失效的基本单元的影响而导致失效的基本单元和相邻的基本单元存储数 据发生变化, 因此, 在对执行写入操作后的基本单元进行回读时, 能够通过 回读的数据与写入数据是否一致来确定失效的基本单元。 另外, 还第二次遍 历、 第三次遍历或后续的遍历过程中, 在进行写入操作之前, 还可以通过读 取操作确定失效的基本单元。
本发明实施例提供的内存检测装置, 进行写入操作的基本单元的物理地 址是连续的, 因此, 能过对存储单元进行一次较为完成的检测, 因此, 提高 了内存检测的全面和准确性, 能够提升存储颗粒故障筛选率, 及时发现内存 中的实效基本单元, 从而把含故障隐患内存类的产品拦截在发货前, 降低维 护成本, 进而提高了内存和设备的可靠性。
最后应说明的是: 以上实施例仅用以说明本发明的技术方案, 而非对其 限制; 尽管参照前述实施例对本发明进行了详细的说明, 本领域的普通技术 人员应当理解: 其依然可以对前述各实施例所记载的技术方案进行修改, 或 者对其中部分技术特征进行等同替换; 而这些修改或者替换, 并不使相应技 术方案的本质脱离本发明各实施例技术方案的精神和范围。

Claims

权 利 要求
1、 一种内存检测方法, 其特征在于, 包括:
对内存物理单元中的基本单元进行间隔写入操作;
对进行写入操作的基本单元进行回读检测, 以确定写入数值与回读数值 不同的基本单元。
2、 根据权利要求 1所述的方法, 其特征在于, 所述对内存物理单元中的 基本单元进行间隔写入操作, 包括:
间隔设定列数逐行对所述基本单元进行写入操作, 直至遍历所述内存物 理单元中的全部基本单元; 或者,
间隔设定行数逐列对所述基本单元进行写入操作, 直至遍历所述内存物 理单元中的全部基本单元。
3、 根据权利要求 2所述的方法, 其特征在于, 所述设定列数为 n, n大 于 0且小于等于每行中基本单元的个数, 或者, 所述设定列数为 2 , m为大 于 0的整数, 且 2 小于等于每行中基本单元的个数;
所述设定行数为 n , n大于 0且小于等于每列中基本单元的个数, 或者, 所述设定行数为 2m, m为大于等于 0的整数, 且 2 小于等于每列中基本单 元的个数。
4、根据权利要求 2或 3所述的方法, 其特征在于, 所述间隔设定列数逐 行对所述基本单元进行写入操作, 直至遍历所述内存物理单元中的全部基本 单元, 包括:
以所述内存物理单元中第一行的起始位置为起点, 间隔设定列数逐行对 所述基本单元进行写入操作;
当对所述内存物理单元中的最后一行进行写入操作后, 返回所述第一行 进行写入操作, 或者, 以所述起始位置的下一个基本单元为起点, 间隔设定 列数逐行对所述基本单元进行写入操作。
5、根据权利要求 2或 3所述的方法, 其特征在于, 所述间隔设定行数逐 列对所述基本单元进行写入操作, 直至遍历所述内存物理单元中的全部基本 单元, 包括:
以所述内存物理单元中第一列的起始位置为起点, 间隔设定行数逐列对 所述基本单元进行写入操作;
当对所述内存物理单元中的最后一列进行写入操作后, 返回所述第一列 进行写入操作, 或者, 以所述起始位置的下一个基本单元为起点, 间隔设定 行数逐列对所述基本单元进行写入操作。
6、 根据权利要求 1-3任一项所述的方法, 其特征在于, 所述对内存物理 单元中的基本单元进行间隔写入操作之前, 还包括:
对所述内存物理单元中的基本单元进行初始化操作。
7、 根据权利要求 6所述的方法, 其特征在于, 所述写入操作为所述初始 化操作的取反操作。
8、 根据权利要求 6所述的方法, 其特征在于, 所述对内存物理单元中的 基本单元进行间隔写入操作之前, 还包括:
对初始化操作后的基本单元进行读取操作, 以确定读取数值与初始化数 值不同的基本单元。
9、根据权利要求 2或 3所述的方法, 其特征在于, 所述对进行写入操作 的基本单元进行回读检测, 包括:
对每个所述基本单元进行写入操作后, 立刻进行回读检测操作; 或者, 对至少一行或一列基本单元进行写入操作后, 对所述至少一行或 一列进行写入操作的基本单元进行回读检测操作。
10、 一种内存检测装置, 其特征在于, 包括:
写入模块, 用于对内存物理单元中的基本单元进行间隔写入操作; 回读检测模块, 用于对进行写入操作的基本单元进行回读检测, 以确定 写入数值与回读数值不同的基本单元。
11、 根据权利要求 10 所述的内存检测装置, 其特征在于, 所述写入模 块具体用于: 间隔设定列数逐行对所述基本单元进行写入操作, 直至遍历所 述内存物理单元中的全部基本单元; 或者, 间隔设定行数逐列对所述基本单 元进行写入操作, 直至遍历所述内存物理单元中的全部基本单元。
12、 根据权利要求 11 所述内存检测装置, 其特征在于, 所述设定列数 为 n, n 大于 0且小于等于每行中基本单元的个数, 或者, 所述设定列数为 2m, m为大于 0的整数, 且2 小于等于每行中基本单元的个数;
所述设定行数为 n, n大于 0且小于等于每列中基本单元的个数, 或者, 所述设定行数为 2m, m为大于等于 0的整数, 且 2m小于等于每列中基本单 元的个数。
13、 根据权利要求 11或 12所述的内存检测装置, 其特征在于, 所述写 入模块具体用于: 以所述内存物理单元中第一行的起始位置为起点, 间隔设 定列数逐行对所述基本单元进行写入操作;
当对所述内存物理单元中的最后一行进行写入操作后, 返回所述第一行 进行写入操作, 或者, 以所述起始位置的下一个基本单元为起点, 间隔设定 列数逐行对所述基本单元进行写入操作;
或者, 以所述内存物理单元中第一列的起始位置为起点, 间隔设定行数 逐列对所述基本单元进行写入操作;
当对所述内存物理单元中的最后一列进行写入操作后, 返回所述第一列 进行写入操作, 或者, 以所述起始位置的下一个基本单元为起点, 间隔设定 行数逐列对所述基本单元进行写入操作。
14、 根据权利要求 10-12任一项所述的内存检测装置, 其特征在于, 还 包括: 初始化模块, 用于在所述写入模块对内存物理单元中的基本单元进行 间隔写入操作之前, 对所述内存物理单元中的基本单元进行初始化操作。
15、 根据权利要求 14 所述的内存检测装置, 其特征在于, 所述写入模 块执行的写入操作为所述初始化模块执行的初始化操作的取反操作。
16、 根据权利要求 14所述的内存检测装置, 其特征在于, 还包括: 读取模块, 用于对初始化操作后的基本单元进行读取操作, 以确定读取 数值与初始化数值不同的基本单元。
17、 根据权利要求 1 1或 12所述的内存检测装置, 其特征在于, 所述回 读检测模块具体用于: 对每个所述基本单元进行写入操作后, 立刻进行回读 检测操作; 或者, 对所述至少一行或一列基本单元进行写入操作后, 对所述 至少一行或一列进行写入操作的基本单元进行回读检测操作。
PCT/CN2011/074099 2011-05-16 2011-05-16 内存检测方法和内存检测装置 WO2011124185A2 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201180001054.4A CN102893263B (zh) 2011-05-16 2011-05-16 内存检测方法和内存检测装置
PCT/CN2011/074099 WO2011124185A2 (zh) 2011-05-16 2011-05-16 内存检测方法和内存检测装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2011/074099 WO2011124185A2 (zh) 2011-05-16 2011-05-16 内存检测方法和内存检测装置

Publications (2)

Publication Number Publication Date
WO2011124185A2 true WO2011124185A2 (zh) 2011-10-13
WO2011124185A3 WO2011124185A3 (zh) 2012-04-19

Family

ID=44763329

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/074099 WO2011124185A2 (zh) 2011-05-16 2011-05-16 内存检测方法和内存检测装置

Country Status (2)

Country Link
CN (1) CN102893263B (zh)
WO (1) WO2011124185A2 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109961824A (zh) * 2019-03-27 2019-07-02 苏州浪潮智能科技有限公司 一种内存测试方法

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106776173B (zh) * 2016-12-15 2019-09-17 郑州云海信息技术有限公司 一种内存检测方法及装置
CN107516546B (zh) * 2017-07-07 2020-09-22 中国航空工业集团公司西安飞行自动控制研究所 一种随机存储器的在线检测装置及方法
CN110647436A (zh) * 2018-06-26 2020-01-03 北京自动化控制设备研究所 一种ddr2/ddr3存储器的快速检测方法
CN112053731A (zh) * 2020-07-28 2020-12-08 深圳市宏旺微电子有限公司 Ddr测试方法和装置
CN113407372B (zh) * 2021-06-01 2023-10-20 中国科学院计算技术研究所 一种独立于操作系统的计算机系统内存检测方法及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5410687A (en) * 1990-03-19 1995-04-25 Advantest Corporation Analyzing device for saving semiconductor memory failures
CN1479207A (zh) * 2002-08-29 2004-03-03 深圳市中兴通讯股份有限公司 内存检测方法
CN101576838A (zh) * 2009-05-13 2009-11-11 北京中星微电子有限公司 一种检测内存的方法和装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5410687A (en) * 1990-03-19 1995-04-25 Advantest Corporation Analyzing device for saving semiconductor memory failures
CN1479207A (zh) * 2002-08-29 2004-03-03 深圳市中兴通讯股份有限公司 内存检测方法
CN101576838A (zh) * 2009-05-13 2009-11-11 北京中星微电子有限公司 一种检测内存的方法和装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109961824A (zh) * 2019-03-27 2019-07-02 苏州浪潮智能科技有限公司 一种内存测试方法

Also Published As

Publication number Publication date
CN102893263B (zh) 2016-12-07
CN102893263A (zh) 2013-01-23
WO2011124185A3 (zh) 2012-04-19

Similar Documents

Publication Publication Date Title
WO2011124185A2 (zh) 内存检测方法和内存检测装置
US9812222B2 (en) Method and apparatus for in-system management and repair of semi-conductor memory failure
CN104051024B (zh) 用于内建错误更正的储存装置及其操作方法
US10204698B2 (en) Method to dynamically inject errors in a repairable memory on silicon and a method to validate built-in-self-repair logic
US9818492B2 (en) Method for testing a memory and memory system
RU2682387C1 (ru) Полупроводниковое запоминающее устройство
US20050289423A1 (en) Built-in self test systems and methods for multiple memories
JPS5936358B2 (ja) 半導体記憶装置に於ける予防的保守を計画的に行なう方法
JP2010537314A5 (zh)
US10249380B2 (en) Embedded memory testing with storage borrowing
CN101477480A (zh) 内存控制方法、装置及内存读写系统
Hsiao et al. Built-in self-repair schemes for flash memories
CN113035259A (zh) Dram测试方法、装置、可读存储介质及电子设备
WO2017012460A1 (zh) 检测随机存储器故障的方法及装置、处理器
CN110942798A (zh) 半导体存储器件、存储系统及操作半导体存储器件的方法
CN113160876A (zh) Dram测试方法、装置、计算机可读存储介质及电子设备
JP5737753B2 (ja) ネットワーク装置、ネットワーク装置の管理方法、管理プログラム、パケットネットワークシステム
JP2004086996A (ja) メモリテスト回路
JP4962277B2 (ja) 半導体メモリ試験装置
CN112466386A (zh) 一种面向故障分类的存储器测试系统及方法
CN113488100A (zh) Dram测试方法、装置、计算机可读存储介质及电子设备
JP2012038168A (ja) 記録ユニット及び故障チップ特定方法
CN103198864B (zh) 双分离栅快闪存储器的访问方法
US11327867B2 (en) On-die logic analyzer
WO2023137855A1 (zh) 存储芯片的测试方法及设备

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 201180001054.4

Country of ref document: CN

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11765103

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11765103

Country of ref document: EP

Kind code of ref document: A2