CN111427713A - Method for estimating service life of storage device by training artificial intelligence - Google Patents
Method for estimating service life of storage device by training artificial intelligence Download PDFInfo
- Publication number
- CN111427713A CN111427713A CN201910022526.2A CN201910022526A CN111427713A CN 111427713 A CN111427713 A CN 111427713A CN 201910022526 A CN201910022526 A CN 201910022526A CN 111427713 A CN111427713 A CN 111427713A
- Authority
- CN
- China
- Prior art keywords
- storage device
- strong
- executing
- artificial intelligence
- correct
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/008—Reliability or availability analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Techniques For Improving Reliability Of Storages (AREA)
Abstract
The invention provides a method for estimating the service life of a storage device by training artificial intelligence, which comprises the following steps: judging whether the operation parameters of the storage device for executing the processing program on the plurality of bit values are smaller than the operation critical parameter value, if so, decoding the bit values stored by the storage device by using a decoder; judging whether the decoder successfully decodes the bit value stored in the storage device, if so, classifying the storage unit of the storage device to belong to a strong correct area, a weak correct area, a strong error area or a weak error area; and judging whether the number of the storage units is within the allowable range of the number, if not, starting an artificial intelligence neural network system, and estimating the service life time of the storage device by using machine learning.
Description
Technical Field
The present invention relates to a storage device, and more particularly, to a method for training artificial intelligence to estimate the lifespan of a storage device.
Background
At present, the memory application is more and more popular, and some factors along with the times of erasing and writing in cause internal damage of the memory in the using process, so that the error rate is increased, and the reliability of a nonvolatile memory (non-volatile memory) is sharply reduced, so that the reliability of the nonvolatile memory can be improved through a reliability design technology, particularly an error correction technology, and the product can be longer in service life and more stable.
Conventionally, BCH (Bose-Chaudhuri-Hochquenghem) Code is adopted as mainstream error correction codes, the calculation speed of the codes is relatively high, and the correction capability is stronger as the redundant bits are more, however, as the manufacturing technology of the nonvolatile memory is higher and higher, the BCH coding technology cannot provide enough correction capability, so that the error correction technology of low Density Parity check codes (L Density Parity Code, L DPC) widely applied in the communication field is turned to be used, and the error correction technology begins to become a new trend in the storage field by virtue of the strong correction capability.
Disclosure of Invention
The present invention is directed to a method for training artificial intelligence to estimate a lifetime of a storage device, which is suitable for a storage device, wherein the storage device includes a plurality of storage units. The method comprises the following steps: (a) performing a data bit processing procedure on one or more bit values using a memory device; (b) calculating the operation parameters of the storage device for executing the data bit processing program to the plurality of bit values; (c) judging whether the operation parameter is smaller than a first operation critical parameter value, if so, executing the step (a), and if not, executing the next step (d); (d) decoding, by a decoder, each of the bit values stored in the storage device; (e) judging whether the decoder successfully decodes the bit values stored in the storage device, if so, executing the step (a), and if not, executing the next step (f); (f) defining a plurality of storage state areas, wherein the storage state areas comprise a strong correct area, a weak correct area, a strong error area and a weak error area; (g) classifying the storage units to belong to a strong correct area, a weak correct area, a strong error area or a weak error area according to the storage states of the storage units; (h) calculating the number of storage units of a plurality of storage units classified in the storage state area; (i) judging whether the number of the storage units is within a first number allowable range, if so, executing the step (a), and if not, executing the next step (j); (j) starting the artificial intelligence neural network system, analyzing whether the operation parameter is smaller than the second operation critical parameter value, whether the decoder successfully decodes the bit values stored in the storage device and whether the number of the storage units is within the second number allowable range by using machine learning, if not, estimating the service life time of the storage device, and if so, executing the step (a).
Preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (k) counting the read-write times of the storage device for reading and writing the bit values; (l) Comparing whether the read-write times are smaller than a read-write times critical value, if so, executing the step (a), and if not, executing the next step (m); (m) counting a number of erases the plurality of bit values by the memory device; (n) comparing whether the erasing times are smaller than an erasing time critical value, if so, executing the step (a), and if not, executing the next step (o); (o) counting busy times for the memory device to read or erase the plurality of bit values; and (p) comparing whether the busy time is smaller than a busy time critical value, if so, executing the step (a), and if not, executing the step (d); wherein the operation parameters include the read-write times, the erase times and the busy time.
Preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (q) calculating a number of strong errors for the memory cells classified in the strong error region; (r) determining whether the number of strong errors is within a first allowable range of the number of strong errors, if so, executing the step (a), and if not, executing the next step(s); starting an artificial intelligence neural network system, analyzing whether the number of the strong errors is within a second allowed range of the number of the strong errors by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device; wherein the number of memory cells includes the number of strong errors, the first allowed range of number includes the first allowed range of number of strong errors, and the second allowed range of number includes the second allowed range of number of strong errors.
Preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (t) calculating a strong error ratio of the number of the plurality of memory cells classified in the strong error region to the number of the plurality of memory cells classified in the sum of the strong error region and the weak error region; (u) determining whether the strong error ratio is less than a first strong error threshold ratio, if so, performing step (a), and if not, performing the next step (v); and (v) starting an artificial intelligence neural network system, analyzing whether the strong error ratio is smaller than a second strong error critical ratio by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device.
Preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (w) calculating a strong correct number of said memory cells classified in said strong correct region; (x) Judging whether the strong correct quantity is larger than a first strong correct quantity threshold value, if so, executing the step (a), and if not, executing the next step (y); and (y) starting an artificial intelligence neural network system, analyzing whether the strong correct quantity is greater than a second strong correct quantity threshold value by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device; wherein the number of memory cells comprises the strong correct number.
Preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (z) calculating a strong-correct proportion of the number of said plurality of memory cells classified in said strong-correct region to the number of said plurality of memory cells classified in the sum of said strong-correct region and said weak-correct region; (aa) determining whether the strong correct ratio is greater than a first strong correct critical ratio, if so, performing step (a), otherwise, performing the next step (bb); and (bb) starting the artificial intelligence neural network system, analyzing whether the strong correct proportion is larger than a second strong correct critical proportion by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device.
Preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (cc) determining whether said decoder successfully decodes said one or more bit values stored in said storage device, if so, performing step (a), and if not, performing the next step (dd); (dd) recording a decoding time for said decoder to decode each of said bit values; and (ee) judging whether the decoding time is less than the first decoding critical time, if so, executing the step (a), and if not, executing the next step (ff); (ff) starting an artificial intelligence neural network system, analyzing whether the decoding time is less than the second decoding critical time by using machine learning, if so, executing the step (a), and if not, estimating the correction capability of the decoder on the bit value stored by the storage device;
preferably, the method for training artificial intelligence to estimate the service life of the storage device further comprises the following steps: (gg) starting an artificial intelligence neural network system, defining a continuous use group and a correction limit group by using machine learning; and (hh) analyzing whether the lifetime of the storage device is greater than a lifetime threshold using machine learning, if so, classifying the storage device into the continuous use group, and if not, classifying the storage device into the correction limit group.
As described above, the method for estimating the lifetime of a memory device by artificial intelligence training provided by the present invention can effectively estimate the lifetime of the memory device according to the operating parameter values such as the read/write times, the erase times, and the busy time of the memory device, the error rate of the data bit value accessed by the memory device (i.e. the number of memory cells classified in a strong error region or other memory states), and the capability of the decoder to correct the data bit value stored by the memory device.
For a better understanding of the features and technical content of the present invention, reference should be made to the following detailed description and accompanying drawings, which are provided for purposes of illustration and description only and are not intended to limit the invention.
Drawings
FIG. 1 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a first embodiment of the present invention.
FIG. 2 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a second embodiment of the present invention.
FIG. 3 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a third embodiment of the present invention.
FIG. 4 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a fourth embodiment of the present invention.
FIG. 5 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a fifth embodiment of the present invention.
FIG. 6 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a sixth embodiment of the present invention.
FIG. 7 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to a seventh embodiment of the present invention.
FIG. 8 is a flowchart illustrating a method for training artificial intelligence to estimate the lifespan of a storage device according to an eighth embodiment of the present invention.
Detailed Description
The following is a description of embodiments of the present invention with reference to specific embodiments, and those skilled in the art will understand the advantages and effects of the present invention from the contents provided in the present specification. The invention is capable of other and different embodiments and its several details are capable of modification and various other changes, which can be made in various details within the specification and without departing from the spirit and scope of the invention. The drawings of the present invention are for illustrative purposes only and are not intended to be drawn to scale. The following embodiments will further explain the related art of the present invention in detail, but the contents are not provided to limit the scope of the present invention.
It will be understood that, although the terms first, second, third, etc. may be used herein to describe various elements or signals, these elements or signals should not be limited by these terms. These terms are used primarily to distinguish one element from another, or from one signal to another. In addition, the term "or" as used herein should be taken to include any one or combination of more of the associated listed items as the case may be.
[ first embodiment ]
Please refer to fig. 1, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a first embodiment of the present invention. As shown in fig. 1, the method for training artificial intelligence to estimate the service life of a storage device according to the present embodiment includes the following steps S101 to S119, which are applicable to a storage device, such as a Solid state drive (Solid state drive), and the storage device includes a plurality of storage units, such as memory cells (memory cells).
In step S101, a data bit processing procedure is performed on one or more bit values using a storage device. For example, the memory device further comprises a memory control unit configured to write one or more data bit values or a bitstream consisting of a plurality of bit values into the memory cells, i.e. to control the memory cells to access the one or more data bit values. Furthermore, the storage control unit of the storage device can read the data bit value stored by the storage unit and control the storage unit to erase the stored data bit value.
In step S103, the operation parameters of the storage device for executing the data bit processing procedure on one or more bit values are calculated, including the number of times the data bit processing procedure is executed, the operation time for executing the data bit processing procedure each time, or a combination thereof.
In step S105, it is determined whether the operating parameter of the memory device is smaller than the first operating threshold parameter value. For example, it is determined whether the number of times the data bit processing program is executed by the memory device is less than a first operation number threshold. Additionally or alternatively, it is determined whether the operation time of the memory device executing the data bit processing program each time is less than a first operation time threshold.
If the operation parameter of the memory device is determined to be less than the first operation threshold parameter value, for example, the operation times of the memory device is less than the first operation time threshold value and the time spent in executing the program each time is less than the first operation time threshold value, the memory device is determined to still have good performance. Therefore, step S101 can be repeatedly executed, and the storage device can execute the data bit processing procedure again, for example, the storage unit can access the new data bit value, or can erase other data bit values stored in the storage unit.
On the contrary, if the operation parameter of the memory device is determined to be equal to or greater than the first operation threshold parameter value, for example, the operation times of the memory device are equal to or greater than the first operation time threshold value and the time spent in executing the program is equal to or greater than the first operation time threshold value, the performance of the memory device may be deteriorated due to the excessive use times or other environmental factors, and the operation time may be prolonged, so as to preliminarily estimate the lifetime of the memory device to be shortened, or even preliminarily estimate that the memory device is about to be damaged or is damaged and may not be used. To estimate the lifetime of the memory device more accurately, the next step S107 can be further performed.
In step S107, the decoder is used to decode one or more bit values stored in the memory cells of the memory device. Specifically, the decoder can decode a plurality of single bit values stored in the same or different memory cells and a plurality of bit values of the same bit stream simultaneously or sequentially. For example, the decoder may sequentially decode the plurality of bit values according to an order in which the plurality of bit values are stored by the storage device. Alternatively, if the plurality of memory cells are arranged in an array in the memory device, the order of sequentially decoding the plurality of bit values may be determined according to the column/row bit values of the plurality of memory cells arranged in the array.
In step S109, it is determined whether the decoder successfully decodes the bit value. In this embodiment, it is assumed that the decoder does not have the capability of inverting the bit value stored in the memory cell of the memory device, i.e., the decoder does not invert the bit value stored in the memory cell from logic 1 to 0 or invert the bit value from logic 0 to logic 1. Under this condition, if the decoder fails to decode the bit value stored in the storage device, it is determined that the storage device may erroneously determine the bit value when accessing data, for example, erroneously determine the bit value as logic 1 when the bit value is logic 0 or erroneously determine the logic 1 as logic 0, and store the erroneous bit value. If so, step S101 can be repeated to re-access the data bit value by the storage device.
On the contrary, if the decoder successfully decodes the bit value stored in the storage device, the decoder can correctly interpret the bit value and store the correct bit value when judging that the storage device is accessing the data bit value, thereby preliminarily deducing that the storage device still has good performance. Therefore, step S111 is then performed.
In step S111, a plurality of memory state areas are defined. For example, the plurality of memory state regions may include more regions such as a strong correct region, a weak correct region, a strong error region, and a weak error region.
In step S113, the memory cells are classified as belonging to a strong-correct region, a weak-correct region, a strong-error region or a weak-error region according to the storage states of the memory cells. For example, when the accuracy of the bit value read by the memory cell is higher than an accuracy threshold, the memory cell is classified in the strong accuracy region. On the contrary, when the accuracy of the memory cell reading bit value is lower than the accuracy threshold, the memory cell is weakened to the correct area. In addition, when the error rate of the memory cell reading bit value is judged to be higher than the error rate threshold value, the memory cell is classified in a strong error area. On the contrary, when the error rate of the memory cell reading bit value is lower than the error rate threshold value, the memory cell is classified in the weak error region.
In step S115, the number of memory cells classified in any one or each of the memory status regions, for example, the strong correct region, the weak correct region, the strong error region, and the weak error region, is calculated.
In step S117, it is determined whether the number of memory cells classified into any or each of the memory state regions, for example, the strong correct region, the weak correct region, the strong error region, and the weak error region, is within a first number allowable range. If the number of the memory cells classified in any or each memory state area according to the memory states of the memory cells is within the default first number allowable range, the memory device is determined to be still usable. On the contrary, when the number of the storage units classified in the storage status area is determined to be beyond the first allowable number range, the storage device is preliminarily estimated to be possibly unusable.
In step S119, the artificial intelligence neural network system is started, and the service life of the storage device is analyzed more precisely by using the relevant values of the storage device through machine learning. The operation of the artificial intelligence-like neural network system is described in more detail below with respect to the second embodiment.
[ second embodiment ]
Please refer to fig. 2, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a second embodiment of the present invention. As shown in FIG. 2, the method for training artificial intelligence to estimate the lifespan of a storage device according to the present embodiment includes the following steps S201-S215, which are applicable to a storage device including a plurality of storage units. It should be understood that the steps of the various embodiments described herein can be combined as desired.
After the artificial intelligence neural network system is started up by executing step S119 or step S203, the following steps S205 to S215 may be executed. That is, the first test of the memory device in steps S103 to S117 and the second test of the memory device in steps S205 to S215 can accurately estimate the lifetime of the memory device by performing two tests on the memory device.
Alternatively, as shown in fig. 2, after the storage device performs the data bit processing procedure on the bit value in step S101 or S201, steps S103 to S117 may be omitted, and the artificial intelligence neural network system is directly activated to perform steps S205 to S215 to estimate the service life of the storage device. It should be understood that the steps described herein, such as steps S207, S209, S213, may be executed in different execution sequences or simultaneously according to actual requirements.
In step S201, the storage device executes a data bit processing procedure on the bit value, such as writing the bit value into the memory cell, reading the bit value stored in the memory cell, and erasing the bit value stored in the memory cell.
In step S203, the artificial intelligence neural network system is started.
In step S205, the artificial intelligence neural network system is used to analyze the second operation threshold parameter value for different types of storage devices with different service lives by using machine learning.
In step S207, machine learning is used to determine whether the operating parameter of the storage device is smaller than the second operating threshold parameter value. In the embodiment, the second operation threshold parameter value is smaller than the first operation threshold parameter value, but the invention is not limited thereto.
For example, the operation threshold parameter values include the operation time for accessing the data bit values, for example, if the memory device is to access data with a bit size within a specific bit size range, the first operation threshold parameter value is set as the access time 0.5ms, and the second operation threshold parameter value is set as the access time 0.2 ms. First, in step S105, the time for accessing the data bit value by the artificial intelligence neural network system using the machine learning analysis storage device is, for example, 0.3ms less than the first operation threshold parameter value, for example, 0.5 ms. Next, in step S203, the time for accessing the data bit value by the artificial intelligence neural network system using the machine learning analysis storage device is less than the second operation threshold parameter value, such as 0.2 ms.
Additionally or alternatively, the corresponding operating threshold parameter values may be set based on an upper threshold amount accessible by memory cells of the memory device, such as 1 bit value accessible by a Single-level Cell (S L C), 2 bit values accessible by a Multi-level Cell (M L C), 4 bit values accessible by a four-level Cell (Quad-L) C, and the currently actually accessed data bit amount.
If the operation parameter of the storage device is analyzed by using the machine learning through the artificial intelligence neural network system to be smaller than the second operation threshold parameter value, for example, the operation times of the storage device is smaller than the second operation time threshold value and the time spent in executing the program each time is smaller than the second operation time threshold value, the storage device is judged to have good performance and can be continuously used. Thus, step S201 may be performed again.
On the contrary, if the operation parameter of the storage device is analyzed by the artificial intelligence neural network system using machine learning and is greater than or equal to the second operation threshold parameter value, for example, the number of operations of the storage device is greater than or equal to the second operation threshold value and the time spent in executing the program each time is greater than or equal to the second operation time threshold value, it is preliminarily inferred that the storage device may not be used any more or that a small number of operations may remain. Further, to more accurately estimate the lifetime of the storage device, step S209 is performed.
In step S209, the artificial intelligence neural network system determines whether the decoder successfully decodes the one or more bit values stored in the memory cells of the memory device using machine learning.
If the decoding of the bit value stored in the decoder storage unit by the artificial intelligence neural network system using machine learning is failed, step S201 is repeatedly executed. On the contrary, if the artificial intelligence neural network system uses the machine learning judgment decoder, the bit value stored in the storage unit can be successfully decoded by one or more decoding procedures, and step S211 is executed.
In step S211, a second allowable number of storage devices for different types and having different lifetimes are analyzed using machine learning. The second allowable range is smaller than the first allowable range, but the invention is not limited thereto.
In step S213, it is determined whether the number of memory cells classified in any or each of the memory status regions, such as the strong correct region, the weak correct region, the strong error region, and the weak error region, is within the second number allowable range using machine learning.
In step S215, the artificial intelligence neural network system is activated, and the machine learning is utilized to determine whether the storage device is currently usable or not according to the determination results of steps S201 to S213, and to determine whether the storage device is unusable due to old or damaged, and whether the storage device needs to be repaired or directly replaced with a new one. Even more, the artificial intelligence neural network system is started, and the steps S201 to S213 are executed again by using machine learning to more accurately estimate whether the lifetime of the storage device is about to expire, for example, whether the lifetime is less than the lifetime threshold, so as to approximately estimate how long the storage device can still be used, for example, more than half a year.
[ third embodiment ]
Please refer to fig. 3, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a third embodiment of the present invention. As shown in fig. 3, the method for training artificial intelligence to estimate the lifespan of a storage device according to the present embodiment includes the following steps S301 to S317, which are applied to a storage device including a plurality of storage units. Steps S301 to S317 of the present embodiment are further examples of steps S103 and S105 of the first embodiment and steps S205 and S207 of the second embodiment. It should be understood that the steps of the various embodiments described herein can be combined as desired.
In step S301, a memory control module of the memory device writes a data bit value into a memory cell. For example, the same memory cell of the memory device can write only a single data bit value, or sequentially read and write a plurality of data bit values, or write a data bit stream composed of a plurality of data bit values. The data bit value is, for example, a logical 0 or a logical 1. It should be understood that the present invention is not limited by the number of bit values read by the memory cell.
In step S303, the number of times of reading and writing of the storage device is counted.
In step S305, it is compared whether the currently accumulated read/write times of the storage device is smaller than a preset read/write times threshold. If the current accumulated read-write times of the storage device is smaller than the preset read-write times threshold value, it is determined that the storage device can be reused for executing the operation of step S301. On the contrary, if the current accumulated read-write times of the storage device is greater than or equal to the preset read-write times critical value, the subsequent steps are further executed, so as to more accurately estimate whether the storage device is still usable or not by various operation parameters of the storage device.
In step S307, the memory control module of the memory device is used to erase one or more data bit values stored in one or more memory cells.
In step S309, the number of times that the memory control module of the memory device erases the data bit value stored in the memory cell is counted.
In step S311, it is compared whether the currently accumulated erase count of the memory device is smaller than a preset erase count threshold. If the current accumulated erase count of the storage device is smaller than the preset erase count threshold value, it is determined that the storage device can be reused to perform the operation of step S301. On the contrary, if the currently accumulated erase count of the memory device is not less than the preset erase count threshold value, the subsequent step S313 is executed to further determine the state of the memory device.
In step S313, the busy time of the storage device is counted. For example, the time for the data bit value to be processed by the counting storage device includes the time for the storage control module to receive and transmit the data bit value to the storage unit, the time for the storage control module to control the storage unit to access the data bit value, and the time for erasing the bit value stored in the storage unit.
In step S315, it is compared whether the busy time of the storage device is less than the busy time threshold. If the busy time of the storage device is smaller than the busy time threshold, it is determined that the accumulated operation time of the storage device is not long, and the step S301 should be performed repeatedly with good operation performance. On the contrary, if the busy time of the storage device is not less than the busy time threshold, it is determined that the storage device is still continuously used to perform step S317.
In step S317, it is estimated whether the life time of the storage device will expire. In detail, if the currently accumulated read/write times of the memory device are not less than the predetermined threshold value in step S305, the currently accumulated erase times of the memory device are not less than the predetermined threshold value in step S311, and the latest busy time of the memory device is not less than the busy time threshold value in step S315, it can be preliminarily estimated that the usable time of the memory device may not be left long, i.e. the usable life may be soon expired.
[ fourth embodiment ]
Please refer to fig. 4, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a fourth embodiment of the present invention. As shown in fig. 4, the method for training artificial intelligence to estimate the lifespan of a storage device according to the present embodiment includes the following steps S401 to S409, which are applied to a storage device including a plurality of storage units. Steps S401 to S403 in the present embodiment are steps S115 to S117 and step S407 in the first embodiment, which are further examples of step S213 in the first embodiment. It should be understood that the steps of the various embodiments described herein can be combined as desired.
In step S401, a plurality of memory cells of the memory device are classified as belonging to the same or different memory state regions, such as more regions including a strong correct region, a weak correct region, a strong error region, and a weak error region, according to the memory states of the data bit values of the memory cells of the memory device, for example, according to the correct rate and the error rate of the bit values of the memory cells. Then, the number of memory cells classified in the strong error region is calculated, i.e. the number of memory cells with the error rate of the misjudgment bit value higher than the error rate threshold value is calculated.
In step S403, it is determined whether the number of memory cells classified in the strong error area is within the allowable range of the first strong error number, i.e. whether the number is smaller than the threshold of the first strong error number. If the number of the memory cells classified in the strong error region is within the allowable range of the first strong error number, i.e. smaller than the threshold of the first strong error number, the memory device is determined to have a certain capability of correctly reading the bit value, and the memory device can be continuously used.
After the memory device is used, the operation performance of the memory device may be affected due to the increase of the number of times of using the memory device, the increase of the usage time, environmental factors, and the like, so that the memory states of the memory cells, such as the interpretation capability, are reduced, and the plurality of memory cells are reclassified into different memory state regions. For example, a memory cell originally classified as a weak error region is reclassified as a strong error region after a period of use. Therefore, if the number of the memory cells classified in the strong error area is determined to be within the first allowable range of the number of strong errors, step S401 can be executed again. If the number of the memory cells classified in the strong error area is determined to be greater than the allowable range of the first strong error number, i.e. determined to be greater than the threshold of the first strong error number, the next step S405 is executed.
In step S405, the artificial intelligence neural network system is started.
In step S407, the artificial intelligence neural network system is used to analyze whether the number of the storage units classified in the strong error region exceeds the allowable range of the second strong error number, i.e. to determine whether the number is smaller than the threshold of the second strong error number. The second allowable range of the number of strong errors may be smaller than the first allowable range of the number of strong errors, i.e., the threshold value of the number of strong errors may be smaller than the threshold value of the number of strong errors.
If the number of the storage units classified in the strong error region by using the machine learning analysis of the artificial intelligence neural network system is within the allowable range of the second strong error number, that is, if it is determined that the number is smaller than the threshold of the second strong error number, step S401 may be executed again after a period of time elapses. On the contrary, when the number of the storage units classified into the strong error region exceeds the second strong error number allowable range, i.e. is judged to be greater than the second strong error number threshold value, step S409 is executed.
In step S409, the lifetime of the storage device is estimated by using the machine learning of the artificial intelligence neural network system according to the number of the storage units in the strong error region and the difference between the number of the storage units and the threshold of the second strong error.
[ fifth embodiment ]
Please refer to fig. 5, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a fifth embodiment of the present invention. As shown in fig. 5, the method for training artificial intelligence to estimate the lifetime of a storage device according to the present embodiment includes the following steps S501 to S513, which are applied to a storage device including a plurality of storage units. It should be understood that the steps of the various embodiments described herein can be combined as desired.
Before the artificial intelligence neural network system is started, the steps S501 to S511 of the present embodiment are preliminarily executed using a system built in a storage control module of another system, for example, a storage device. Then, after the artificial intelligence neural network system is started, the steps S501 to S511 of this embodiment are executed again, wherein different second allowable ranges of the number of strong errors and second critical ratios of the number of strong errors can be set to replace the first allowable ranges of the number of strong errors and the first critical ratios of the strong errors, respectively. In another embodiment, after the artificial intelligence neural network system is started, the steps S501 to S511 of this embodiment are executed.
In step S501, a data bit processing procedure is performed on a plurality of bit values by using a plurality of memory cells of the memory device.
In step S503, the number of strong errors of the memory cells classified in the strong error area is calculated.
In step S505, it is determined whether the number of strong errors is within the first allowable range of the number of strong errors. If the number of strong errors is determined to be within the allowable range of the first number of strong errors, step S501 may be executed again. On the contrary, if the number of strong errors is determined to be beyond the allowable range of the first number of strong errors, step S507 is executed.
In step S507, the number of weak error storage units of the storage units classified in the weak error area is calculated.
In step S509, a strong error ratio is calculated in which the number of the plurality of memory cells classified in the strong error region accounts for the number of the plurality of memory cells classified in the sum of the strong error region and the weak error region. The Strong Error Ratio (SER) is expressed by the following calculation formula:
where SER represents the ratio of strong errors, SE represents the number of memory cells in the strong error area, and WE represents the number of memory cells in the weak error area.
In step S511, it is determined whether the strong error ratio is smaller than the first strong error threshold ratio. If the strong error ratio is smaller than the first strong error threshold ratio, step S501 may be executed again. Conversely, if the strong error ratio is equal to or greater than the first strong error threshold ratio, step S513 is executed.
In step S513, the artificial intelligence neural network system is activated to estimate the service life of the storage device, and more precisely estimate the service life of each storage unit of the storage device.
It should be understood that, in practice, the comparison of the number of memory cells classified in the strong error region and the first allowable range of the number of strong errors in step S505 may be omitted, and only the comparison of the ratio of the strong errors and the ratio smaller than the first critical ratio of the strong errors in step S511 is performed.
[ sixth embodiment ]
Please refer to fig. 6, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a sixth embodiment of the present invention. As shown in fig. 6, the method for training artificial intelligence to estimate the lifespan of a storage device according to the present embodiment includes the following steps S601 to S609, which are applied to a storage device including a plurality of storage units. It should be understood that the steps of the various embodiments described herein can be combined as desired.
In step S601, a plurality of memory cells of the memory device are classified to belong to the same or different memory state regions, for example, more regions including a strong correct region, a weak correct region, a strong error region, and a weak error region, according to the memory states of the data bit values of the memory cells of the memory device, for example, according to the correct rate and the error rate of the bit values of the memory cells. Next, the strong correct number of memory cells classified in the strong correct region is calculated.
In step S603, it is determined whether the strong correct number of memory cells classified in the strong correct area is greater than a first strong correct number threshold, e.g., 8 memory cells. If the strong correct number of the memory cells classified in the strong correct area, for example, 10 memory cells, is greater than the first strong correct number threshold, for example, 8 memory cells, it is determined that the memory device can continue to operate, and step S601 can be executed again during the subsequent operation of the memory device. On the contrary, if it is determined that the strong correct number of the storage units classified in the strong correct area, for example, 6 storage units, is less than or equal to the first strong correct number threshold, for example, 8 storage units, step S605 is executed.
In step S605, the artificial intelligence neural network system is started.
In step S607, a machine learning analysis is used to determine whether the strong correctness number of the memory cells classified in the strong correctness region is greater than a second strong correctness number threshold. For example, the second strong correct number threshold, e.g., 5 memory cells, is less than the first strong correct number threshold, e.g., 8 memory cells.
If the strong correct number of the memory cells classified in the strong correct area, for example, 6 memory cells, is greater than the second strong correct number threshold, for example, 5 memory cells, it is determined that the memory device can continue to operate, and step S601 can be executed again during the subsequent operation of the memory device. Conversely, if the strong correct number of the memory cells classified in the strong correct area, for example, 4 memory cells, is smaller than or equal to the second strong correct number threshold, for example, 5 memory cells, step S609 is executed.
In step S609, the lifetime of the storage device is estimated to expire soon.
[ seventh embodiment ]
Please refer to fig. 7, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a storage device according to a seventh embodiment of the present invention. As shown in FIG. 7, the method for training artificial intelligence to estimate the lifespan of a storage device according to the present embodiment includes the following steps S701-S713, which are applied to a storage device including a plurality of storage units. It should be understood that the steps of the various embodiments described herein can be combined as desired.
In step S701, a data bit processing procedure is performed on a plurality of bit values by using a plurality of memory cells of the memory device.
In step S703, the strong correct number of memory cells classified in the strong correct area is calculated.
In step S705, it is determined whether the strong correct number of memory cells classified in the strong correct area is greater than a first strong correct number threshold. If the strong correct number of the memory cells classified in the strong correct area is greater than the first strong correct number threshold, step S701 may be repeated. Conversely, if the strong correct number of the memory cells classified in the strong correct area is determined to be less than or equal to the first strong correct number threshold, step S707 is executed.
In step S707, the number of weakly correct memory cells of the memory cells classified in the weakly correct region is calculated.
In step S709, a strong correct ratio of the number of memory cells classified in the strong correct region to the number of memory cells classified in the sum of the strong correct region and the weak correct region is calculated. The Strong Correct Ratio (SCR) is expressed as the following calculation:
where SCR represents the strong correct ratio, SC represents the number of memory cells in the strong correct region, and WC represents the number of memory cells in the weak correct region.
In step S711, it is determined whether the strong correct ratio is greater than the first strong correct critical ratio. If the strong correct ratio is greater than the first strong correct threshold ratio, step S701 may be repeated. If the strong correct ratio is not greater than the first strong correct threshold ratio, step S713 is executed.
In step S713, the artificial intelligence neural network system is started to estimate the service life of the storage device according to the determination result. Even more, the artificial intelligence neural network system may be started to perform the above steps S701 to S711 again, and the first strong correct quantity threshold and the first strong correct critical ratio may be replaced by a second strong correct quantity threshold and a second strong correct critical ratio having different values.
As described above, the sixth embodiment calculates a strong error ratio, and the seventh embodiment calculates a strong correct ratio. Alternatively, in yet another embodiment, a strong error ratio of the number of memory cells classified in the strong error region to the number of memory cells classified in the strong error region is calculated. Then, it is determined whether the ratio of the strong errors is smaller than a critical ratio of the strong errors, or whether the number of the memory cells classified in the strong error area is smaller than the number of the memory cells classified in the strong error area. If the strong correctness ratio is less than the critical strong correctness ratio, or the number of the memory cells classified in the strong error region is less than the number of the memory cells classified in the strong error region, step S701 may be repeatedly performed. Conversely, step S713 is executed.
[ eighth embodiment ]
Please refer to fig. 8, which is a flowchart illustrating a method for training artificial intelligence to estimate a lifetime of a memory device according to an eighth embodiment of the present invention. As shown in fig. 8, the method for training artificial intelligence to estimate the lifespan of a storage device according to the present embodiment includes the following steps S801 to S817, which are applied to a storage device including a plurality of storage units. It should be understood that the steps of the various embodiments described herein can be combined as desired.
In step S801, a decoder decodes a bit value stored in a memory cell of a memory device.
In step S803, it is determined whether the decoder successfully decodes the bit value stored in the memory cell of the memory device. In order to accurately estimate the lifetime of the memory device, the decoder may decode all or most of the bit values stored in the plurality of memory cells of the memory device and record the time consumed by each decoding, but the invention is not limited thereto. In practice, to save the estimation time, only a part of the bit values stored in the memory cell may be decoded.
If the decoder fails to decode the memory cells of the memory device, step S805 is executed: recording the decoding time of the decoder for decoding the bit value. Then, step S803 is repeated, and the decoder can flip the storage device to store the erroneous bit value with a certain probability, or the decoder can decode the bit value re-accessed by the storage device. Conversely, if the decoder successfully decodes the memory cells of the memory device, step S807 is executed.
In step S807, it is determined whether the decoding time of the current decoded bit value of the decoder is less than the default first decoding threshold time.
In step S809, the artificial intelligence type neural network system is started.
In step S811, the artificial intelligence neural network system is used to determine whether the decoding time is less than the second decoding threshold time by using machine learning analysis. The second decoding critical time may be less than the first decoding critical time.
In step S813, the machine learning is used by the artificial intelligence neural network system to estimate the correction capability of the decoder for the storage device according to the determination result.
In particular, in an ideal situation, when decoding the bit value stored in the memory cell of the storage device, the decoder may execute a bit value correction procedure to flip the erroneous bit value stored in the memory cell with a specific probability to be the correct bit value, for example, the bit value of the storage device to be stored with logic 0 is erroneously determined and erroneously stored as logic 1 when accessing, and the decoder executes the bit value correction procedure to flip the logic 1 back to the bit value of the logic 0 to realize the correction of the bit value and store the correct bit value in the memory cell of the storage device.
However, if the performance of the storage device is too poor, which may cause the storage device to misjudge a large number of data bit values or store invalid data, i.e., the error rate of misjudged bit values is higher than the error rate threshold, or the performance of the decoder becomes too poor, which may cause a load level exceeding the correction capability of the decoder or the decoder not having a good correction capability, which may cause the decoder to be unable to effectively correct the erroneous data stored in the storage device, it may be estimated that the lifetime of the decoder or the storage device will expire, and the available decoder, the storage device, or both may need to be replaced with a new one.
In step S815, the artificial intelligence neural network system is used to define the continuous use group and the correction limit group by using machine learning.
In step S817, the storage devices are classified into continuous use groups when the storage devices are analyzed for continuous use according to the estimation result by using machine learning through the artificial intelligence neural network system. Conversely, if the storage device is determined to be incapable of further use, the storage device is classified into the correction limit group.
[ advantageous effects of the embodiments ]
In summary, the method for training artificial intelligence to estimate the lifetime of a memory device provided by the present invention can effectively estimate the lifetime of the memory device according to the operation parameter values such as the read/write times, the erase times, and the busy time of the memory device, the error rate of the data bit value accessed by the memory device (i.e. the number of memory cells classified in the strong error area or other memory states), and the capability of the decoder to correct the data bit value stored by the memory device.
It should be finally noted that while in the foregoing specification, the present inventive concept has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the present inventive concept as defined by the appended claims.
Claims (8)
1. A method for training artificial intelligence to estimate a lifetime of a storage device, the method being adapted for use with a storage device, the storage device comprising a plurality of storage cells, the method comprising:
(a) performing a data bit processing procedure on one or more bit values using the memory device;
(b) calculating the operation parameters of the storage device for executing the data bit processing program on the bit values;
(c) judging whether the operation parameter is smaller than a first operation critical parameter value, if so, executing the step (a), and if not, executing the next step (d);
(d) decoding each of the bit values stored in the storage device using a decoder;
(e) judging whether the decoder successfully decodes the bit values stored in the storage device, if so, executing the step (a), and if not, executing the next step (f);
(f) defining a plurality of storage state areas, wherein the storage state areas comprise a strong correct area, a weak correct area, a strong error area and a weak error area;
(g) classifying the memory cells belonging to the strong error region, the weak error region, the strong error region or the weak error region according to the memory states of the memory cells;
(h) calculating the number of memory cells of the plurality of memory cells classified in the memory state region;
(i) judging whether the number of the storage units is within a first number allowable range, if so, executing the step (a), and if not, executing the next step (j);
(j) starting an artificial intelligence neural network system, analyzing whether the operation parameter is smaller than a second operation critical parameter value, whether the decoder successfully decodes the bit values stored in the storage device and whether the number of the storage units is within a second number allowable range by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device.
2. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(k) counting the read-write times of the storage device for reading and writing the bit values;
(l) Comparing whether the read-write times are smaller than a read-write times critical value, if so, executing the step (a), and if not, executing the next step (m);
(m) counting a number of erases the plurality of bit values by the memory device;
(n) comparing whether the erasing times are smaller than an erasing time critical value, if so, executing the step (a), and if not, executing the next step (o);
(o) counting busy times for the memory device to read or erase the plurality of bit values; and
(p) comparing whether the busy time is smaller than a busy time critical value, if so, executing the step (a), and if not, executing the step (d);
wherein the operation parameters include the read-write times, the erase times and the busy time.
3. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(q) calculating a number of strong errors for the memory cells classified in the strong error region;
(r) determining whether the number of strong errors is within a first allowable range of the number of strong errors, if so, executing the step (a), and if not, executing the next step(s);
(s) starting an artificial intelligence neural network system, analyzing whether the number of the strong errors is within a second allowed range of the number of the strong errors by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device; and
wherein the number of memory cells includes the number of strong errors, the first allowed range of number includes the first allowed range of number of strong errors, and the second allowed range of number includes the second allowed range of number of strong errors.
4. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(t) calculating a strong error ratio of the number of the plurality of memory cells classified in the strong error region to the number of the plurality of memory cells classified in the sum of the strong error region and the weak error region;
(u) determining whether the strong error ratio is less than a first strong error threshold ratio, if so, performing step (a), and if not, performing the next step (v); and
(v) starting an artificial intelligence neural network system, analyzing whether the strong error ratio is smaller than a second strong error critical ratio by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device.
5. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(w) calculating a strong correct number of said memory cells classified in said strong correct region;
(x) Judging whether the strong correct quantity is larger than a first strong correct quantity threshold value, if so, executing the step (a), and if not, executing the next step (y); and
(y) starting an artificial intelligence neural network system, analyzing whether the strong correct quantity is greater than a second strong correct quantity threshold value or not by using machine learning, if so, executing the step (a), and if not, estimating the service life time of the storage device;
wherein the number of memory cells comprises the strong correct number.
6. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(z) calculating a strong-correct proportion of the number of said plurality of memory cells classified in said strong-correct region to the number of said plurality of memory cells classified in the sum of said strong-correct region and said weak-correct region;
(aa) determining whether the strong correct ratio is greater than a first strong correct critical ratio, if so, performing step (a), otherwise, performing the next step (bb); and
(bb) starting the artificial intelligence neural network system, analyzing whether the strong correct proportion is larger than a second strong correct critical proportion by using machine learning, if so, executing the step (a), and if not, estimating the service life of the storage device.
7. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(cc) determining whether said decoder successfully decodes said one or more bit values stored in said storage device, if so, performing step (a), and if not, performing the next step (dd);
(dd) recording a decoding time for said decoder to decode each of said bit values; and
(ee) judging whether the decoding time is less than a first decoding critical time, if so, executing the step (a), and if not, executing the next step (ff);
(ff) starting an artificial intelligence neural network system, analyzing whether the decoding time is less than the second decoding critical time by using machine learning, if so, executing the step (a), and if not, estimating the correction capability of the decoder on the bit value stored by the storage device.
8. The method for estimating the lifespan of a storage device according to claim 1, wherein the method for estimating the lifespan of a storage device by artificial intelligence further comprises the steps of:
(gg) starting an artificial intelligence neural network system, defining a continuous use group and a correction limit group by using machine learning; and
(hh) using machine learning to analyze whether the lifetime of the storage device is greater than a lifetime threshold, if so, classifying the storage device into the continuous use group, and if not, classifying the storage device into the correction limit group.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910022526.2A CN111427713B (en) | 2019-01-10 | 2019-01-10 | Method for training artificial intelligence to estimate service life of storage device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910022526.2A CN111427713B (en) | 2019-01-10 | 2019-01-10 | Method for training artificial intelligence to estimate service life of storage device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111427713A true CN111427713A (en) | 2020-07-17 |
| CN111427713B CN111427713B (en) | 2023-08-22 |
Family
ID=71545848
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910022526.2A Active CN111427713B (en) | 2019-01-10 | 2019-01-10 | Method for training artificial intelligence to estimate service life of storage device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111427713B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101211306A (en) * | 2006-12-27 | 2008-07-02 | 宇瞻科技股份有限公司 | Solid-state memory life-span warning device and its method |
| CN104733044A (en) * | 2013-12-24 | 2015-06-24 | 群联电子股份有限公司 | Decoding method, memory storage device and memory control circuit unit |
| US20160011930A1 (en) * | 2014-07-11 | 2016-01-14 | Phison Electronics Corp. | Method for preventing read-disturb errors, memory control circuit unit and memory storage apparatus |
| CN105988910A (en) * | 2015-02-05 | 2016-10-05 | 纬创资通股份有限公司 | Storage device life monitoring system and storage device life monitoring method thereof |
| CN108228371A (en) * | 2016-12-15 | 2018-06-29 | 发那科株式会社 | Machine learning device and method, life predication apparatus, numerical control device |
| CN108369818A (en) * | 2016-03-09 | 2018-08-03 | 华为技术有限公司 | A kind of method for refreshing and device of flash memory device |
-
2019
- 2019-01-10 CN CN201910022526.2A patent/CN111427713B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101211306A (en) * | 2006-12-27 | 2008-07-02 | 宇瞻科技股份有限公司 | Solid-state memory life-span warning device and its method |
| CN104733044A (en) * | 2013-12-24 | 2015-06-24 | 群联电子股份有限公司 | Decoding method, memory storage device and memory control circuit unit |
| US20160011930A1 (en) * | 2014-07-11 | 2016-01-14 | Phison Electronics Corp. | Method for preventing read-disturb errors, memory control circuit unit and memory storage apparatus |
| CN105988910A (en) * | 2015-02-05 | 2016-10-05 | 纬创资通股份有限公司 | Storage device life monitoring system and storage device life monitoring method thereof |
| CN108369818A (en) * | 2016-03-09 | 2018-08-03 | 华为技术有限公司 | A kind of method for refreshing and device of flash memory device |
| CN108228371A (en) * | 2016-12-15 | 2018-06-29 | 发那科株式会社 | Machine learning device and method, life predication apparatus, numerical control device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111427713B (en) | 2023-08-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10283215B2 (en) | Nonvolatile memory system with background reference positioning and local reference positioning | |
| EP2530595B1 (en) | System and method of tracking error data within a storage device | |
| KR101951988B1 (en) | Word line defect detection and handling for a data storage device | |
| US8711619B2 (en) | Categorizing bit errors of solid-state, non-volatile memory | |
| KR100321302B1 (en) | A method of controlling flash memory and a flash memory chip | |
| US8467249B2 (en) | Systems and methods for storing, retrieving, and adjusting read thresholds in flash memory storage system | |
| US9042169B2 (en) | Shifting cell voltage based on grouping of solid-state, non-volatile memory cells | |
| CN1578988B (en) | Error management for writable tracking storage units storing reference values | |
| US8539311B2 (en) | System and method for data recovery in multi-level cell memories | |
| US9342399B2 (en) | Threshold voltage calibration using reference pattern detection | |
| US8996936B2 (en) | Enhanced error correction in memory devices | |
| US20220165348A1 (en) | Method and apparatus for determining when actual wear of a flash memory device differs from reliability states for the flash memory device | |
| US8730724B2 (en) | Common line current for program level determination in flash memory | |
| US9021340B2 (en) | Error recovery using erasures for NAND Flash | |
| CN111427713B (en) | Method for training artificial intelligence to estimate service life of storage device | |
| TWI690860B (en) | Method of training artificial intelligence to estimate lifetime of storage device | |
| CN109215716B (en) | Method and device for improving reliability of NAND type floating gate memory | |
| US11361221B2 (en) | Method of training artificial intelligence to estimate lifetime of storage device | |
| CN110970080B (en) | Method for training artificial intelligence to estimate sensing voltage of storage device | |
| CN113380309A (en) | Method for searching optimal reading voltage and flash memory system | |
| CN110969256A (en) | Method for training artificial intelligence to correct logarithm probability ratio of storage device | |
| CN115273942A (en) | Solid state disk and configuration method, system and device thereof | |
| CN116665745A (en) | Flash memory reading method and device | |
| CN114023372A (en) | Memory management method, memory storage device and memory controller | |
| CN111785314A (en) | Method, system, storage medium and terminal for reducing over-erasure phenomenon of non-flash memory |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |