ReRAM channel decoding method and device based on stealth path distance

By calculating the stealth path distance and combining hard decision and large number logic decoding algorithms, the problem of stealth path interference in ReRAM storage is solved, improving decoding reliability and efficiency. It is applicable to channel coding methods such as LDPC codes and Polar codes.

CN117220827BActive Publication Date: 2026-11-17JINLING INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311010438.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-08-11
Publication Date
2026-11-17
Estimated Expiration
2043-08-11

AI Technical Summary

Technical Problem

In existing ReRAM storage technology, creeping path interference causes the channel decoder to fail to identify storage cells with pseudo-low resistance values, reducing the reliability of the storage channel. Traditional channel error correction codes cannot meet the requirements of high-density storage.

Method used

By calculating the stealth path distance of the memory cell, hard decision and large number logic decoding algorithms are used to reduce stealth path interference and improve decoding reliability.

Benefits of technology

It significantly improves the decoding performance of ReRAM storage channels, reduces the false low resistance value cell error rate, and enhances data reliability and decoding efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117220827B_ABST
    Figure CN117220827B_ABST
Patent Text Reader

Abstract

This invention discloses a ReRAM channel decoding method and device based on stealth path distance. The method includes: Step S1, calculating the stealth path distance of the resistance value sequence to be decoded; Step S2, performing a hard decision based on the stealth path distance, and using the obtained sequence as the initial value of the decoding attempt sequence; Step S3, calculating the reliability of the initial value of the decoding attempt sequence; Step S4, calculating the checksum of the decoding attempt sequence. If the checksum is 0, the current decoding attempt sequence is output as the decoding result; otherwise, proceed to Step S5; Step S5, determining whether the current iteration count has reached the maximum value. If yes, decoding fails; otherwise, proceed to Step S6; Step S6, updating the reliability of the decoding attempt sequence based on the checksum; Step S7, updating the decoding attempt sequence based on the updated reliability, incrementing the iteration count by 1, and proceeding to Step S4. This invention can solve or reduce stealth path interference.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to decoding technology, and more particularly to a ReRAM channel decoding method and apparatus based on stealth path distance. Background Technology

[0002] With the rapid development of emerging technologies such as big data, artificial intelligence (AI), 5G / 6G, cloud computing, and cloud storage, data is growing explosively, driving the demand for storage in terms of large capacity, high performance, and low power consumption. Existing storage technologies can no longer meet the requirements of new scenarios, making novel non-volatile memory (NVM) a current research hotspot. Among them, resistive random access memory (ReRAM) is considered one of the most promising candidates, with its simple device structure, easy miniaturization, high speed, low power consumption, low cost, and easy compatibility with CMOS processes.

[0003] The write and read processing flow of ReRAM is as follows: Figure 1 As shown, the process includes the following steps: (1) Channel coding: Select a channel coding scheme to obtain codeword X. (2) Zero padding: Pad codeword X with zeros to obtain codeword X', with a zero padding length of L0 = M × N - mod(N). c M×N), mod(·) is the remainder function, N c M is the code length of the LDPC code, and M and N are the number of rows and columns of the ReRAM resistor cross array. (3) Block processing: Divide X' into L s individual code blocks ceil(·) is the floor function. (4) ReRAM channel: sequentially... Each storage cell in the memory is assigned a resistance value, using U ij This represents the storage cell in the i-th row and j-th column of the array, if its data value x ij If the value is 0, then its resistance is set to a high resistance value R0; otherwise, it is set to a low resistance value R1.

[0004]

[0005] Taking R0 = 1000Ω and R1 = 100Ω as an example, 1 ≤ l ≤ L s Channel noise n ij It is Gaussian white noise with a noise variance of δ. 2 (5) Resistance measurement: The resistance of L is measured sequentially. s(6) Combining: This process is the reverse of the block-based process, outputting a one-dimensional array. (7) Removing zeros: This process is the reverse of the zero-padding process. (8) Channel decoding: Corresponding to the channel coding scheme, the corresponding decoding algorithm is used for decoding to obtain the estimated information sequence. The reading can be expressed as: y ij =x ij +n ij |1≤i≤M,1≤j≤N. Where (1)~(4) are the writing process, and (5)~(8) are the reading process.

[0006] Crawling paths are a key factor affecting the reliability of ReRAM write and read processes. The essence of the crawling path problem is that during the read process, there is a certain probability that a high-resistance memory cell being read may form a loop with other low-resistance memory cells, resulting in a parallel connection and causing a significant deviation in the read resistance value. For example... Figure 2 As shown, when a memory cell at diagonal cell (1,4) fails, the memory cell (3,2)→(3,4)→(1,4)→(1,2)→(3,2) forms a creeping path (red line) parallel to the target cell (3,2). This creeping path reduces the measured resistance value of the target cell. The memory cell U where the creeping path occurs... ij The other three memory cells that form the stealth path loop must satisfy the following:

[0007] Where c1≠j, l1≠i

[0008] When a stealth path interference occurs, the memory cell U is in a high-resistance state (actual data is "0"). ij It will be incorrectly read as a low resistance state, and the data "1" will be read.

[0009] Interference such as creeping paths leads to uneven reliability in ReRAM storage channels. Traditional error correction codes (ECCs) are limited by their error correction capabilities and cannot meet the data reliability requirements of high-density storage. In ReRAM storage channels, creeping paths prevent the channel decoder from distinguishing between pseudo-low resistance memory cells and original low resistance memory cells. Figure 3 As shown, this increases decoding delay and can even lead to decoding failure. Therefore, it is necessary to develop a channel decoding scheme that can solve or reduce stealth path interference. Summary of the Invention

[0010] Purpose of the invention: This invention addresses the problems existing in the prior art by providing a ReRAM channel decoding method and device based on the stealth path distance that can solve or reduce stealth path interference.

[0011] Technical solution: The ReRAM channel decoding method based on stealth path distance described in this invention includes:

[0012] Step S1: Calculate the stealth path distance of the sequence of resistance values ​​to be decoded based on the measured resistance of the memory cells affected by the stealth path.

[0013] Step S2: Make a hard decision based on the stealth path distance and use the resulting sequence as the initial value of the decoding attempt sequence;

[0014] Step S3: Calculate the reliability of the initial values ​​of the decoding attempt sequence;

[0015] Step S4: Calculate the checksum of the decoding attempt sequence. If the checksum is 0, output the decoding attempt sequence as the decoding result; otherwise, proceed to step S5.

[0016] Step S5: Determine if the current iteration count has reached the maximum value. If yes, the decoding fails; otherwise, proceed to step S6.

[0017] Step S6: Update the reliability of the decoded attempt sequence based on the checksum;

[0018] Step S7: Update the decoding attempt sequence according to the updated reliability, increment the iteration count by 1, and execute step S4.

[0019] Furthermore, the method for calculating the stealth path distance in step S1 is as follows:

[0020]

[0021]

[0022] R s = (1 / R0 + 1 / 3R1) -1

[0023]

[0024] In the formula, The k-th resistor value r in the sequence of resistor values ​​to be decoded k The stealth path distance, W k P sp As an intermediate variable, R s This represents the measured resistance of the memory cell affected by the crawl path, where n is the length of the resistance value sequence to be decoded, R1 and R0 are the resistance values ​​set for data 1 and 0 during encoding, and δ 2 Let be the channel noise variance, and unit(·) be the normalization function.

[0025] Furthermore, the formula for the hard decision in step S2 is:

[0026]

[0027] In the formula, This represents the k-th bit in the initial value of the decoding attempt sequence. To attempt sequence initialization for decoding, This represents the stealth path distance of the k-th resistor value in the sequence of resistor values ​​to be decoded, where n is the length of the sequence of resistor values ​​to be decoded.

[0028] Optionally, the method for calculating the reliability of the initial value of the decoding attempt sequence in step S3 is as follows:

[0029]

[0030] In the formula, To determine the reliability of the k-th bit in the initial value of the decoding attempt sequence, r k For the k-th resistor value to be decoded in the sequence of resistor values, It amp For hard decision amplitude, R1 and R0 are the resistance values ​​set during encoding for data 1 and 0, respectively. s This represents the measured resistance of the memory cell affected by the crawling path, where n is the length of the sequence of resistance values ​​to be decoded.

[0031] Optionally, the method for calculating the reliability of the initial value of the decoding attempt sequence in step S3 is as follows:

[0032]

[0033] In the formula, To determine the reliability of the k-th bit in the initial value of the decoding attempt sequence, r k For the k-th resistor value to be decoded in the sequence of resistor values, It amp For hard decision amplitude, R1 and R0 are the resistance values ​​set during encoding for data 1 and 0, respectively. s This represents the measured resistance of the memory cell affected by the crawling path, where n is the length of the sequence of resistance values ​​to be decoded.

[0034] Furthermore, the calculation method for the checksum in step S4 is as follows:

[0035] s (p) =z (p) H T

[0036] In the formula, s (p) z represents the checksum of the decoding attempt sequence at the p-th iteration. (p) Let H represent the decoding attempt sequence at the p-th iteration, and let H represent the parity check matrix of the LDPC code.

[0037] Furthermore, the method for updating the reliability of the decoding attempt sequence in step S6 is as follows:

[0038]

[0039]

[0040]

[0041] In the formula, This represents the reliability of the k-th bit in the decoding attempt sequence at the (p+1)th and p-th iterations. This indicates an update to the metric. It represents the update metric increment. amp For hard judgment range, H c (:,k) represents the row containing the non-zero element in the k-th column of vector H, length() represents the number of elements in the vector, and s (p) This represents the decoding attempt sequence checksum at the p-th iteration, sum(s) (p) (H c (:,k))) represents the number of elements in column k of H that fail the validation in the row containing the non-zero element. It represents the k-th bit of the decoding attempt sequence at the p-th iteration, and n is the length of the resistor value sequence to be decoded.

[0042] Furthermore, the method for updating the decoding attempt sequence in step S7 is as follows:

[0043] If the reliability of the k-th bit in the decoded sequence after the update is... Then the k-th bit of the decoding attempt sequence will be assigned the value 0. otherwise The updated codewords constitute a new sequence of decoding attempts. n is the length of the resistor value sequence to be decoded.

[0044] The present invention also provides a ReRAM channel decoding device based on stealth path distance, including a processor and an executable program stored in a memory and executable on the processor, wherein the processor implements the above method when executing the executable program.

[0045] The present invention also discloses a storage medium containing a computer-executable program, which, when executed by a computer processor, is used to perform the above-described method.

[0046] Beneficial Effects: Compared with existing technologies, the significant advantages of this invention are: This patent proposes a large-number logic decoding algorithm based on stealth path distance. By calculating the stealth path distance of the storage unit, a preliminary hard decision is made, and then combined with the large-number logic decoding algorithm, the problem of false low-resistance value unit misidentification caused by stealth path is solved or reduced, thus resisting stealth path interference. The scheme proposed in this invention is applicable to various channel coding methods, such as LDPC codes and Polar codes. Attached Figure Description

[0047] Figure 1 This is a flowchart of ReRAM write and read processes;

[0048] Figure 2 This is a schematic diagram of a ReRAM array and its corresponding logic values;

[0049] Figure 3 This is a schematic diagram of the ReRAM resistor state probability distribution;

[0050] Figure 4 This is a flowchart illustrating the ReRAM channel decoding method based on stealth path distance provided by the present invention.

[0051] Figure 5 This is a comparison chart of the experimental results of the present invention and the traditional method;

[0052] Figure 6 This is a schematic diagram of the ReRAM channel decoding device based on stealth path distance provided by the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0055] Example 1

[0056] This embodiment provides a ReRAM channel decoding method based on stealth path distance. This decoding method decodes the sequence of resistance values ​​to be decoded obtained during the ReRAM read process, after resistance measurement, merging, and zero-removal. Figure 4 As shown, the method specifically includes the following steps:

[0057] Step S1: Calculate the stealth path distance of the sequence of resistance values ​​to be decoded based on the measured resistance of the memory cell affected by the stealth path.

[0058] In specific implementation, the method for calculating the stealth path distance is as follows:

[0059]

[0060]

[0061] R s = (1 / R0 + 1 / 3R1) -1

[0062]

[0063] In the formula, The k-th resistor value r in the sequence of resistor values ​​to be decoded k The stealth path distance, W k P sp As an intermediate variable, R s This represents the measured resistance of the memory cell affected by the crawl path, where n is the length of the resistance value sequence to be decoded, R1 and R0 are the resistance values ​​set for data 1 and 0 during encoding, and δ 2 Let be the channel noise variance, and unit(·) be the normalization function.

[0064] Step S2: Make a hard decision based on the stealth path distance and use the resulting sequence as the initial value of the decoding attempt sequence.

[0065] In practice, the formula for a hard decision is:

[0066]

[0067] In the formula, This represents the k-th bit in the initial value of the decoding attempt sequence. This represents the stealth path distance of the k-th resistor value in the sequence of resistor values ​​to be decoded, where n is the length of the sequence of resistor values ​​to be decoded. Initialize the sequence for decoding.

[0068] Step S3: Calculate the reliability of the initial value of the decoding attempt sequence.

[0069] In implementation, the reliability of the initial value of the decoding attempt sequence can be achieved using the following method:

[0070]

[0071] In the formula, To determine the reliability of the k-th bit in the initial value of the decoding attempt sequence, r k For the k-th resistor value to be decoded in the sequence of resistor values, It amp This represents the range of hard judgments.

[0072] In other embodiments, reliability can also be calculated using the following method:

[0073]

[0074] Step S4: Calculate the checksum of the decoding attempt sequence. If the checksum is 0, output the decoding attempt sequence as the decoding result; otherwise, proceed to step S5.

[0075] The calculation method for the checksum is as follows:

[0076] s (p) =z (p) H T

[0077] In the formula, s (p) z represents the checksum of the decoding attempt sequence at the p-th iteration. (p) Let H represent the decoding attempt sequence at the p-th iteration, and let H represent the parity check matrix of the LDPC code.

[0078] Step S5: Determine if the current iteration count has reached the maximum value. If yes, the decoding fails. If no, proceed to step S6.

[0079] Step S6: Update the reliability of the decoded attempt sequence based on the checksum.

[0080] In specific implementation, the method for updating the reliability of the decoding attempt sequence is as follows:

[0081]

[0082]

[0083]

[0084] In the formula, This represents the reliability of the k-th bit in the decoding attempt sequence at the (p+1)th and p-th iterations. This indicates an update to the metric. It represents the update metric increment. amp For hard judgment range, H c(:,k) represents the row containing the non-zero element in the k-th column of vector H, length() represents the number of elements in the vector, and s (p) This represents the decoding attempt sequence checksum at the p-th iteration, sum(s) (p) (H c (:,k))) represents the number of elements in column k of H that fail the validation in the row containing the non-zero element. It represents the k-th bit of the decoding attempt sequence at the p-th iteration, and n is the length of the resistor value sequence to be decoded.

[0085] Step S7: Update the decoding attempt sequence according to the updated reliability, increment the iteration count by 1, and execute step S4.

[0086] In practice, the method for updating the decoding attempt sequence is as follows: if the reliability of the k-th bit in the updated decoding attempt sequence is... Then the k-th bit of the decoding attempt sequence will be assigned the value 0. otherwise The updated codewords constitute a new sequence of decoding attempts. n is the length of the resistor value sequence to be decoded.

[0087] Figure 5 The graph shows the BER performance comparison curves of the proposed stealth path distance-based decoding (Scheme 1 and Scheme 2) with traditional hard decision and HRBI (Hard Reliability-based Iterative) large number logic decoding algorithms under ReRAM storage channels. It is clear from the graph that traditional decoding algorithms perform poorly under ReRAM storage channels, while the decoding algorithm proposed in this invention can achieve a performance gain of up to four orders of magnitude.

[0088] Example 2

[0089] Figure 6 This is a schematic diagram of the structure of a device provided by an embodiment of the present invention. The embodiment of the present invention provides services for the implementation of the method of the above-described embodiment one. For example... Figure 2 As shown, the device may include: a memory 301 storing a computer-executable program; a processor 302 coupled to the memory 301; the processor 302 calls the computer-executable program stored in the memory 301 to perform the steps in the method described in Embodiment 1.

[0090] Memory 301 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, memory 301 may be used to read and write non-removable, non-volatile magnetic media (commonly referred to as a "hard disk drive"). A program / utility having a set (at least one) of program modules may be stored, for example, in memory 301. Such program modules include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. The computer-executable program of the program modules typically performs the functions and / or methods described in the embodiments of the present invention.

[0091] The code for performing the operations of the present invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages.

[0092] The processor 302 executes various functional applications and data processing by running programs stored in the memory 301, such as implementing the method provided in Embodiment 1 of the present invention.

[0093] Example 3

[0094] This invention provides a storage medium containing a computer-executable program, which, when executed by a computer processor, is used to perform the method of Embodiment 1.

[0095] The storage medium of embodiments of the present invention may be any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0096] The code for a computer-executable program that performs the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0097] Of course, the computer-executable program provided in the embodiments of the present invention is not limited to the above-described method operations, but can also perform related operations in the methods provided in any embodiment of the present invention.

Claims

1. A method for ReRAM channel decoding based on a distance of a sneak path, characterized in that, include: Step S1: Calculate the stealth path distance of the sequence of resistance values ​​to be decoded based on the measured resistance of the memory cells affected by the stealth path. Step S2: Make a hard decision based on the stealth path distance and use the resulting sequence as the initial value of the decoding attempt sequence; Step S3: Calculate the reliability of the initial values ​​of the decoding attempt sequence; Step S4: Calculate the checksum of the decoding attempt sequence. If the checksum is 0, output the decoding attempt sequence as the decoding result. Otherwise, proceed to step S5; Step S5: Determine if the current iteration count has reached the maximum value. If yes, the decoding fails. If no, proceed to step S6. Step S6: Update the reliability of the decoded attempt sequence based on the checksum; Step S7: Update the decoding attempt sequence according to the updated reliability, increment the iteration count by 1, and execute step S4; The method for calculating the stealth path distance in step S1 is as follows: , , , , wherein is the distance of the kth resistance value to be decoded in the sequence of resistance values to be decoded is the distance of the kth resistance value to be decoded in the sequence of resistance values to be decoded is an intermediate variable denotes the measured resistance of the storage cell affected by the creeping path n is the length of the sequence of resistance values to be decoded is the resistance value set corresponding to the data 1, 0 during encoding is the channel noise variance, unit is a normalization function The method for calculating the checksum in step S4 is as follows: , wherein denotes the parity check matrix of the LDPC code, denotes the sequence of decoding attempts at the p-th iteration, denotes the parity check matrix of the LDPC code; The method for updating the reliability of the decoding attempt sequence in step S6 is as follows: , , , In the formula, , Indicates the first The reliability of the k-th bit in the decoding attempt sequence during p iterations. This indicates an update to the metric. This indicates an update to the metric increment. For the scope of a hard judgment, The parity check matrix representing the LDPC code The row containing the non-zero elements in the k-th column of the array. This indicates the number of elements in the vector. This represents the decoding attempt sequence checksum during the p-th iteration. express The number of elements in the row containing the non-zero element in the k-th column that failed the validation. This represents the k-th bit of the decoding attempt sequence at the p-th iteration. n The length of the resistor value sequence to be decoded.

2. The method of claim 1, wherein: The formula for the hard decision in step S2 is: , In the formula, represents the kth bit in the initial value of the decoding attempt sequence, is the initial value of the decoding attempt sequence, represents the potential path distance of the kth to-be-decoded resistance value in the to-be-decoded resistance value sequence, n is the length of the to-be-decoded resistance value sequence.

3. The method of claim 1, wherein: The method for calculating the reliability of the initial value of the decoding attempt sequence in step S3 is as follows: , In the formula, is the reliability of the kth bit in the initial value of the decoding attempt sequence, is the kth to-be-decoded resistance value in the to-be-decoded resistance value sequence, is the hard decision amplitude, is the resistance value corresponding to the data 1, 0 set during encoding, represents the measured resistance of the storage unit affected by the sneak path, n is the length of the to-be-decoded resistance value sequence.

4. The method of claim 1, wherein: The method for calculating the reliability of the initial value of the decoding attempt sequence in step S3 is as follows: , In the formula, To assess the reliability of the k-th bit in the initial value of the decoding attempt sequence, This represents the k-th resistor value to be decoded in the sequence of resistor values. For the scope of a hard judgment, These are the resistance values ​​set for data 1 and 0 during encoding. The measured resistance of the memory cell affected by the crawling path. n The length of the resistor value sequence to be decoded.

5. The ReRAM channel decoding method based on stealth path distance according to claim 1, characterized in that: The method for updating the decoding attempt sequence in step S7 is as follows: If the reliability of the kth bit in the updated decoding attempt sequence , the kth bit of the decoding attempt sequence is assigned a value of 0, ; otherwise , the updated codeword forms a new decoding attempt sequence , n is the length of the sequence of resistance values to be decoded.

6. A ReRAM channel decoding device based on a sneak path distance, characterized in that, The invention includes a processor and an executable program stored in a memory and executable on the processor, characterized in that: when the processor executes the executable program, it implements the method as described in any one of claims 1-5.

7. A storage medium containing a computer executable program, characterized in that, The computer executable program, when executed by a computer processor, is used to perform the method as described in any one of claims 1-5.