CN106330204B - A kind of data processing method and device - Google Patents
A kind of data processing method and device Download PDFInfo
- Publication number
- CN106330204B CN106330204B CN201610796646.4A CN201610796646A CN106330204B CN 106330204 B CN106330204 B CN 106330204B CN 201610796646 A CN201610796646 A CN 201610796646A CN 106330204 B CN106330204 B CN 106330204B
- Authority
- CN
- China
- Prior art keywords
- matrix
- data
- rank
- data processing
- circular
- 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.)
- Active
Links
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
- H03M13/03—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
- H03M13/05—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits
- H03M13/11—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits using multiple parity bits
- H03M13/1102—Codes on graphs and decoding on graphs, e.g. low-density parity check [LDPC] codes
- H03M13/1148—Structural properties of the code parity-check or generator matrix
- H03M13/118—Parity check matrix structured for simplifying encoding, e.g. by having a triangular or an approximate triangular structure
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/007—Transform coding, e.g. discrete cosine transform
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Discrete Mathematics (AREA)
- Multimedia (AREA)
- Probability & Statistics with Applications (AREA)
- Error Detection And Correction (AREA)
Abstract
The present invention provides a kind of data processing method and devices, for decoding to low density parity check code.In data structure, the unit submatrix reduction sparse matrix scale under high code rate specification in left submatrix can use；In data expression, data fixed point scheme can be used；In instruction, check matrix is saved using constant memory, and promote memory efficiency etc. using software pipeline when to global variable access；On platform, frame level degree of parallelism can be set, increase concurrent thread number；It algorithmically, can be using compromise hierarchical decoder processing mode.The quick decoding of low density parity check code is realized in the combination that the application passes through multi-level a variety of optimization means.
Description
Technical field
The present invention relates to coding and decoding technical fields, in particular to a kind of data processing method and device.
Background technique
LDPC code be it is a kind of approach shannon limit, easily realize the outstanding linear error correction code low with system complexity, it is
There is the linear block codes of sparse check matrix by one kind that Robert doctor Gallager proposed in 1963, LDPC code is not
The superperformance of shannon limit is only approached, and decoding complexity is lower, flexible structure is the research of field of channel coding in recent years
Hot spot.It is widely used to the fields such as deep space communication, fiber optic communication, satellite digital video and audio broadcasting at present.LDPC code is
As forth generation communication system (4G) strong competitor, and the encoding scheme based on LDPC code is by next-generation satellite number
Word video broadcast standards DVB-S2 adopts.
Inventors have found that can only be carried out by serial decoded mode, such as in the decoding decoding process of LDPC code matrix
Group/cording quantity in fruit LDPC code matrix is more, and serial decoded mode will take a substantial amount of time, and decoding efficiency is low, occupies
More system resource.
Summary of the invention
In view of this, the present invention provides a kind of data processing method, it can be with Parallel Implementation to the solution of LDPC encoder matrix
Code, decoding efficiency are high.
Technical solution provided by the invention is as follows:
A kind of data processing method, for being decoded to low density parity check code, the verification square of low density parity check code
Battle array includes left submatrix and right submatrix, this method comprises:
Capable transformation is carried out to the left submatrix and right submatrix, rank transformation is carried out to the right submatrix, by the school
Testing matrix conversion is the first matrix comprising multiple n rank square matrixes；
By the first matrix one-dimensional to form linear data structure, all effective members are chosen from first matrix
Number is the circular matrix of the integral multiple of n, records line number, columns and initial shifting of the circular matrix in first matrix
Position position；
Extract multiple circular matrixes parallel simultaneously, according to line number of the circular matrix in first matrix,
Columns and initial displacement position determine first matrix.
Further, it is described by the first matrix one-dimensional to form linear data structure the step of include:
Using compromise layered mode and the number of iterations is minimized, to reduce total operand.
Further, the data of the check matrix indicate to indicate using data fixed point, specially by 32 bit list essences
Degree floating number is converted to 8 bit integers.
Further, the data of the check matrix indicate to indicate using data fixed point, specifically:
Normalized input data is reduced into operation to -127 to 127, in the reduction calculating process using first function
Operated in saturation is used to prevent 8 bit integer from overflowing.
Further, it in the step of determining first matrix, is determined using multiple concurrent threads, specifically:
Thread block in first matrix is mapped as the form that a check matrix includes multiple circular matrixes, by line
The circular matrix that journey is mapped as a single non-superimposed includes m effectively first forms, m 360.
Further, the low density parity check code is the LDPC code of DVB-S2 Internal Code or the preposition operation that interweaves.
Further, the data processing method is realized using graphics processor GPU.
Present invention also provides a kind of data processing equipments, for being decoded to low density parity check code, low-density parity
The check matrix of check code includes left submatrix and right submatrix, which includes:
Conversion module arranges the right submatrix for carrying out capable transformation to the left submatrix and right submatrix
Transformation, is converted to the first matrix comprising multiple n rank square matrixes for the check matrix；
One-dimensional module, for by the first matrix one-dimensional to form linear data structure, from first matrix
The middle circular matrix for choosing the integral multiple that all effective first numbers are n, records the circular matrix in first matrix
Line number, columns and initial displacement position；
Processing module, for extracting multiple circular matrixes parallel simultaneously, according to the circular matrix described first
Line number, columns and initial displacement position in matrix, determine first matrix.
Further, the one-dimensional module is specifically used for using compromise layered mode and minimizes the number of iterations, to subtract
Few total operand.
Further, the data of the check matrix indicate to indicate using data fixed point, specially by 32 bit list essences
Degree floating number is converted to 8 bit integers.
Using data processing method provided by the invention, during being decoded to low-density odd/even check code matrix,
By way of row transformation is converted to low-density odd/even check code matrix comprising multiple n rank square matrixes, and pass through matrix one-dimensional
Record saves line number, columns and the respective data content of the n rank square matrix of data content non-zero.Thus in decoding, Ke Yitong
The n rank square matrix of the multiple data content non-zeros of extraction of Shi Binghang, determines low-density parity by its line number, columns and data content
The data content of check code matrix realizes decoding.In decoding process, without handling data all in matrix, only need
The data of data content non-zero are handled, data processing amount is less, and parallel processing may be implemented after one-dimensional, will not occur
The loss of interference and encoded information.Faster, decoding efficiency is higher for decoding speed, can make full use of graphics processor and line number
According to processing capacity, the utilization rate of equipment is improved.
To enable the above objects, features and advantages of the present invention to be clearer and more comprehensible, preferred embodiment is cited below particularly, and cooperate
Appended attached drawing, is described in detail below.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not construed as pair
The restriction of range for those of ordinary skill in the art without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is a kind of flow diagram of data processing method provided in an embodiment of the present invention.
Fig. 2 is a kind of the functional block diagram of data processing equipment provided in an embodiment of the present invention.
Specific embodiment
Below in conjunction with attached drawing in the embodiment of the present invention, technical solution in the embodiment of the present invention carries out clear, complete
Ground description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Usually exist
The component of the embodiment of the present invention described and illustrated in attached drawing can be arranged and be designed with a variety of different configurations herein.Cause
This, is not intended to limit claimed invention to the detailed description of the embodiment of the present invention provided in the accompanying drawings below
Range, but it is merely representative of selected embodiment of the invention.Based on the embodiment of the present invention, those skilled in the art are not doing
Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
It should also be noted that similar label and letter indicate similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined in a attached drawing, does not then need that it is further defined and explained in subsequent attached drawing.Meanwhile of the invention
In description, term " first ", " second " etc. are only used for distinguishing description, are not understood to indicate or imply relative importance.
Low density parity check code (Low Density Parity Check Code, LDPC) is a kind of very common
Coding mode, LDPC code have huge application potential, deep space communication, fiber optic communication, satellite digital video, digital watermarking,
It is obtained in magnetic optical/Hologram Storage, movement and fixed radio communication, cable modulator/demodulator and Digital Subscriber Line (DSL) extensively
Using.
In data processing, the transmitting and decoding of coding are realized by low-density odd/even check code matrix.Existing skill
In art, serially executed when being decoded to low-density odd/even check code matrix.Needing will be in low-density odd/even check code matrix
All data elements are saved and are handled.
It include a large amount of neutral element in sparse matrix since low-density odd/even check code matrix can be sparse matrix,
When being decoded, the treatment process to neutral element is not no practical significance, but due to the substantial amounts of neutral element, so that low
The decoding process efficiency of density parity check code matrix is lower.
When use graphics processor (Graphics Processing Unit, GPU) is to low-density odd/even check code matrix
When being decoded, due to can only serially execute decoding process, the parallel processing capability of GPU cannot be made full use of, reduces GPU's
Service efficiency causes the waste of system resource.
In view of this, the embodiment of the present application provides a kind of data processing method, for low density parity check code square
Battle array decoding, as shown in Figure 1, this method comprises:
Step 101, capable transformation is carried out to the left submatrix and right submatrix, rank transformation is carried out to the right submatrix,
The check matrix is converted into the first matrix comprising multiple n rank square matrixes.
The left submatrix of low-density odd/even check code matrix and right submatrix in the embodiment of the present application, and left submatrix and
Right submatrix is all sparse matrix, and the quantity for the element that numerical value is 0 in matrix will be far more than the quantity of nonzero element.The application
In capable transformation carried out to left submatrix and right submatrix, and rank transformation is carried out to right submatrix, by variable node information code element square
Battle array is converted to the form comprising multiple n rank square matrixes.Check matrix after row-column transform is still sparse matrix, but if from
The dimensional analysis of the n rank square matrix check matrix, complexity have obtained a degree of simplification.
In the embodiment of the present application, the low-density odd/even check code matrix after converting is the first matrix.In the first matrix
In, both included the n rank square matrix that data content is zero, and had also included the n rank square matrix of data content non-zero.
Step 102, by the first matrix one-dimensional to form linear data structure, institute is chosen from first matrix
Effectively member number be n integral multiple circular matrix, record line number of the circular matrix in first matrix, columns
With initial displacement position.
Specifically, during carrying out one-dimensional to the first matrix using compromise layered mode and iteration can be minimized
Number, to reduce total operand.
For 2/3 yard, it can be that unit is divided by the size (360) of circular matrix, make check matrix natively
The macro matrix of a 15*45 is formed, a line is one layer corresponding.
Wherein, hierarchical decoder Layered decoding calculating process are as follows:
qm,n[k]=qn[k*G+g-1]-rm,n[k-1]
qn[k*G+g]=qn[k*G+g-1]+rm,n[k]-rm,n[k-1]
Enabling k is iteration index, and g is layer index, and G is the number of plies, then layer operation is as shown in above formula.qnIt is initialized to
Channel info, and exported eventually as the foundation of symbol judgement.The results of property of existing algorithm is compared it is found that connecing in performance
In the case where close, the number of iterations of hierarchical decoder layered decoding algorithm halves.
After converting to low-density odd/even check code matrix, there is no variations for conceptual data amount, if directly
The first matrix obtained after conversion is decoded, decoding operand is compared with directly decoding low-density odd/even check code matrix
It is not significantly reduced.In the embodiment of the present application, the operation of one-dimensional further has been carried out to the first matrix.
Only the n rank square matrix of data content non-zero is saved, by the n rank square matrix of data content non-zero in the first square
Line number, columns and specific data content in battle array are saved.Since low-density odd/even check code matrix is sparse matrix,
The first matrix obtained after conversion is also sparse matrix, and the quantity of the n rank square matrix of data content non-zero will be much in the first matrix
Less than the quantity for the n rank square matrix that data content is zero.Therefore, line number, the column of the n rank square matrix of data content non-zero are only saved
The operand of the parameters such as number, data content will far smaller than save the operand of entire low-density odd/even check code matrix, also want
Far smaller than save the operand of entire first matrix.Realizing two-dimensional first matrix conversion is one-dimensional data.
Step 103, multiple circular matrixes while are parallel extracted, according to the circular matrix in first matrix
Line number, columns and initial displacement position, determine first matrix.
In the embodiment of the present application, the data of the check matrix indicate to include: by 32 ratios using the expression of data fixed point
Special single precision floating datum is converted to 8 bit integers.And by normalized input data using first function reduce operation to-
127 to 127, operated in saturation is used in the reduction calculating process to prevent 8 bit integer from overflowing.
During the one-dimensional of the first matrix, the note of line number and columns has been carried out to the n rank square matrix of data content non-zero
Record, the line number columns saved according to a n rank square matrix record, so that it may know specific position of the n rank square matrix in the first matrix
It sets.In the first matrix, total line number is the integral multiple of n, and total columns is also the integral multiple of n.For example, low density parity check code square
Battle array includes 16200 column, 5400 row data, n 360, then just including 45 column, 15 row n ranks in the first matrix being converted to
Square matrix.Wherein, the columns of n rank square matrix be 1 to 15 wherein any one, line number be 1 to 45 wherein any one.According to the rank side n
The specific columns and line number of battle array are assured that the specific location of the n rank square matrix in the first matrix.
Determining n rank square matrix behind the position in the first matrix by line number, columns, so that it may according to one-dimensional process
The data content for the n rank square matrix that middle record saves determines the data content of the first matrix.
After the n rank square matrix to the data content non-zero in the first matrix carries out record preservation, so that it may according to data
The n rank square matrix of content non-zero restores the coding of the first matrix.Specifically, can be according to the n rank of the data content non-zero of preservation
The line number and columns of square matrix determine the specific location of the n rank square matrix in the first matrix, and the number of the n rank square matrix according to preservation
The data content of the specific location is determined according to content.
In the embodiment of the present application, the n rank square matrix of multiple data content non-zeros can be handled simultaneously, due to
The n rank square matrix of data content non-zero is all independent from each other, while will not there is a situation where interfere with each other when being decoded.
For example, if every a line includes m n rank square matrix in the first matrix, can from the n rank square matrix of all data content non-zeros,
The identical n rank square matrix of line number is extracted simultaneously, i.e., extracts m n rank square matrix simultaneously, is distinguished further according to the line number of n rank square matrix
It determines the specific location of each n rank square matrix, and determines the specific data content of the position further according to the data content of n rank square matrix.
Characteristic based on low-density odd/even check code matrix, in the first matrix converted in every a line data content non-zero n rank
The quantity of square matrix is identical.The other positions in the processing for the n rank square matrix for having carried out all data content non-zeros, the first matrix
N rank square matrix each element for including all be just zero, to complete the decoding of entire low-density odd/even check code matrix.
In actual treatment, can be with parallel data processing using GPU the characteristics of, using identical instruction place parallel simultaneously
Manage the n rank square matrix of data content non-zero.It is serially decoded compared to the prior art, faster, decoding efficiency is higher for processing speed.
In the embodiment of the present application, during being decoded to low-density odd/even check code matrix, by low close
Degree odd/even check code matrix is converted, so that the matrix after conversion includes multiple n rank square matrixes, in decoding process, only for
The n rank square matrix of data content non-zero is handled, since low-density odd/even check code matrix is sparse matrix, data content non-zero
The quantity of n rank square matrix to be far smaller than the quantity of neutral element so that the operand of decoding process reduces, decoding efficiency is mentioned
It is high.Also, the characteristics of using GPU parallel data processing, can carry out with n rank square matrix of the time-division multiple threads to data content non-zero
Processing, improves the decoding speed of low-density odd/even check code matrix, improves the utilization rate of GPU.
Left submatrix and right submatrix are all sparse matrix, so that the first matrix after converting is also into sparse matrix.Existing
Have in technology, since low-density odd/even check code matrix is being decoded for sparse matrix wherein including a large amount of neutral element
In the process, it is still desirable to the operation that these neutral elements are saved, are judged, will duplicate multiple carry out data reading
It takes, judge, and these neutral elements are substantially without coding meaning, reading and judgement for neutral element will be occupied largely
Decoding time and hardware resource.
In the embodiment of the present application, in data structure, it can use unit under high code rate specification in left submatrix
Matrix condensation sparse matrix scale；In data expression, data fixed point scheme can be used；In instruction, use
Constant memory saves check matrix, and promotes memory efficiency using software pipeline when to global variable access
Deng；On platform, frame level degree of parallelism can be set, increase concurrent thread number；Algorithmically, at can be using compromise hierarchical decoder
Reason mode.The quick decoding of low density parity check code is realized in the combination that the application passes through multi-level a variety of optimization means.
Present invention also provides a kind of data processing equipments 200, as shown in Fig. 2, including conversion module 201, one-dimensional mould
Block 202 and processing module 203.Concrete methods of realizing is identical as above method embodiment, and which is not described herein again.
In conclusion being solved using data processing method provided by the invention to low-density odd/even check code matrix
During code, by way of row transformation is converted to low-density odd/even check code matrix comprising multiple n rank square matrixes, and pass through square
Battle array one-dimensional record saves line number, columns and the respective data content of the n rank square matrix of data content non-zero.To decode
When, the n rank square matrix of the multiple data content non-zeros of extraction that can be parallel simultaneously is determined by its line number, columns and data content
The data content of low-density odd/even check code matrix realizes decoding.In decoding process, without being carried out to data all in matrix
Processing need to only handle the data of data content non-zero, and data processing amount is less, and parallel processing may be implemented after one-dimensional,
Interference and the loss of encoded information will not occur.Faster, decoding efficiency is higher for decoding speed, can make full use of graphics process
The parallel data processing capacity of device, improves the utilization rate of equipment.
In several embodiments provided herein, it should be understood that disclosed device and method can also pass through
Other modes are realized.The apparatus embodiments described above are merely exemplary, for example, flow chart and block diagram in attached drawing
Show the device of multiple embodiments according to the present invention, the architectural framework in the cards of method and computer program product,
Function and operation.In this regard, each box in flowchart or block diagram can represent the one of a module, section or code
Part, a part of the module, section or code, which includes that one or more is for implementing the specified logical function, to be held
Row instruction.It should also be noted that function marked in the box can also be to be different from some implementations as replacement
The sequence marked in attached drawing occurs.For example, two continuous boxes can actually be basically executed in parallel, they are sometimes
It can execute in the opposite order, this depends on the function involved.It is also noted that every in block diagram and or flow chart
The combination of box in a box and block diagram and or flow chart can use the dedicated base for executing defined function or movement
It realizes, or can realize using a combination of dedicated hardware and computer instructions in the system of hardware.
In addition, each functional module in each embodiment of the present invention can integrate one independent portion of formation together
Point, it is also possible to modules individualism, an independent part can also be integrated to form with two or more modules.
It, can be with if the function is realized and when sold or used as an independent product in the form of software function module
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.It needs
Illustrate, herein, relational terms such as first and second and the like be used merely to by an entity or operation with
Another entity or operation distinguish, and without necessarily requiring or implying between these entities or operation, there are any this realities
The relationship or sequence on border.Moreover, the terms "include", "comprise" or its any other variant are intended to the packet of nonexcludability
Contain, so that the process, method, article or equipment for including a series of elements not only includes those elements, but also including
Other elements that are not explicitly listed, or further include for elements inherent to such a process, method, article, or device.
In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including the element
Process, method, article or equipment in there is also other identical elements.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any to repair
Change, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.It should also be noted that similar label and letter exist
Similar terms are indicated in following attached drawing, therefore, once being defined in a certain Xiang Yi attached drawing, are then not required in subsequent attached drawing
It is further defined and explained.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. a kind of data processing method, which is characterized in that for being decoded to low density parity check code, low density parity check code
Check matrix include left submatrix and right submatrix, this method comprises:
Capable transformation is carried out to the left submatrix and right submatrix, rank transformation is carried out to the right submatrix, by the verification square
Battle array is converted to the first matrix comprising multiple n rank square matrixes, has both included the n rank square matrix that data content is zero in first matrix,
It also include the n rank square matrix of data content non-zero；
By the first matrix one-dimensional to form linear data structure, all effective first numbers are chosen from first matrix
For the circular matrix of the integral multiple of n, line number, columns and initial displacement position of the circular matrix in first matrix are recorded
It sets；
Multiple circular matrixes are extracted parallel simultaneously, according to line number of the circular matrix in first matrix, columns
With initial displacement position, first matrix is determined；
Wherein, line number, columns and the initial displacement position according to the circular matrix in first matrix determines described
The step of one matrix includes:
The coding that the first matrix is restored according to the n rank square matrix of data content non-zero, according to the n rank of the data content non-zero of preservation
The line number and columns of square matrix determine the specific location of the n rank square matrix in the first matrix, and the number of the n rank square matrix according to preservation
The data content of the specific location is determined according to content.
2. data processing method according to claim 1, which is characterized in that it is described by the first matrix one-dimensional with shape
The step of linear data structure includes:
Using compromise layered mode and the number of iterations is minimized, to reduce total operand.
3. data processing method according to claim 2, which is characterized in that the data of the check matrix indicate to use number
It is indicated according to fixed point, 32 bit single precision floating datums is specially converted into 8 bit integers.
4. data processing method according to claim 3, which is characterized in that the data of the check matrix indicate to use number
It is indicated according to fixed point, specifically:
Normalized input data is reduced into operation to -127 to 127 using first function, is used in the reduction calculating process
Operated in saturation is to prevent 8 bit integer from overflowing.
5. data processing method according to claim 1, which is characterized in that in the step of determining first matrix,
It is determined using multiple concurrent threads, specifically:
Thread block in first matrix is mapped as the form that a check matrix includes multiple circular matrixes, thread is reflected
Penetrating as the circular matrix of a single non-superimposed includes m effectively first forms, m 360.
6. data processing method according to claim 1, which is characterized in that the low density parity check code is DVB-S2
Internal Code or the LDPC code of the preposition operation that interweaves.
7. data processing method according to claim 1, which is characterized in that the data processing method uses graphics process
Device GPU is realized.
8. a kind of data processing equipment, which is characterized in that for being decoded to low density parity check code, low density parity check code
Check matrix include left submatrix and right submatrix, which includes:
Conversion module carries out rank transformation to the right submatrix for carrying out capable transformation to the left submatrix and right submatrix,
The check matrix is converted into the first matrix comprising multiple n rank square matrixes, was both including data content in first matrix
Zero n rank square matrix also includes the n rank square matrix of data content non-zero；
One-dimensional module, for form linear data structure, selecting the first matrix one-dimensional from first matrix
The circular matrix for taking the integral multiple that all effective first numbers are n, record line number of the circular matrix in first matrix,
Columns and initial displacement position；
Processing module, for extracting multiple circular matrixes parallel simultaneously, according to the circular matrix in first matrix
In line number, columns and initial displacement position, determine first matrix；
Wherein, line number, columns and initial displacement position of the processing module according to the circular matrix in first matrix
It sets, the method for determining first matrix includes:
The coding that the first matrix is restored according to the n rank square matrix of data content non-zero, according to the n rank of the data content non-zero of preservation
The line number and columns of square matrix determine the specific location of the n rank square matrix in the first matrix, and the number of the n rank square matrix according to preservation
The data content of the specific location is determined according to content.
9. data processing equipment according to claim 8, which is characterized in that the one-dimensional module is specifically used for using folding
Middle layered mode and minimum the number of iterations, to reduce total operand.
10. data processing equipment according to claim 8, which is characterized in that the data of the check matrix indicate to use
Data fixed point indicates, 32 bit single precision floating datums are specially converted to 8 bit integers.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610796646.4A CN106330204B (en) | 2016-08-31 | 2016-08-31 | A kind of data processing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610796646.4A CN106330204B (en) | 2016-08-31 | 2016-08-31 | A kind of data processing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106330204A CN106330204A (en) | 2017-01-11 |
| CN106330204B true CN106330204B (en) | 2019-08-09 |
Family
ID=57787390
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610796646.4A Active CN106330204B (en) | 2016-08-31 | 2016-08-31 | A kind of data processing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106330204B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006020205A1 (en) * | 2004-07-19 | 2006-02-23 | Texas Instruments Incorporated | Improvements to a layered decoding approach for low density parity check (ldpc) codes |
| CN102315899A (en) * | 2010-06-30 | 2012-01-11 | 中国科学院微电子研究所 | Quasi-cyclic decoding method for low-density parity check code |
| US8347200B1 (en) * | 2011-07-29 | 2013-01-01 | Stec, Inc. | Reduced memory multi-channel parallel encoder system |
| CN102932003A (en) * | 2012-09-07 | 2013-02-13 | 上海交通大学 | Accelerated QC-LDPC (Quasi-Cyclic Low-Density Parity-Check Code) decoding method based on GPU (Graphics Processing Unit) framework |
| CN103427848A (en) * | 2012-08-03 | 2013-12-04 | 上海数字电视国家工程研究中心有限公司 | Method for constructing LDPC codes and codes |
-
2016
- 2016-08-31 CN CN201610796646.4A patent/CN106330204B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006020205A1 (en) * | 2004-07-19 | 2006-02-23 | Texas Instruments Incorporated | Improvements to a layered decoding approach for low density parity check (ldpc) codes |
| CN102315899A (en) * | 2010-06-30 | 2012-01-11 | 中国科学院微电子研究所 | Quasi-cyclic decoding method for low-density parity check code |
| US8347200B1 (en) * | 2011-07-29 | 2013-01-01 | Stec, Inc. | Reduced memory multi-channel parallel encoder system |
| CN103427848A (en) * | 2012-08-03 | 2013-12-04 | 上海数字电视国家工程研究中心有限公司 | Method for constructing LDPC codes and codes |
| CN102932003A (en) * | 2012-09-07 | 2013-02-13 | 上海交通大学 | Accelerated QC-LDPC (Quasi-Cyclic Low-Density Parity-Check Code) decoding method based on GPU (Graphics Processing Unit) framework |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106330204A (en) | 2017-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111699696B (en) | Method and apparatus for encoding and decoding byte stream | |
| US11632560B2 (en) | Methods and apparatuses for encoding and decoding a bytestream | |
| CN100596029C (en) | Method of constructing check matrix for LDPC code, and encoding and decoding device of using the method | |
| CN105684316B (en) | Polar code encoding method and device | |
| CN104124980B (en) | It is adapted to the high speed secret negotiation method of continuous variable quantum key distribution | |
| Yerva et al. | Lossless image compression based on data folding | |
| WO2013156084A1 (en) | Method for processing an image | |
| KR20140096298A (en) | Position coding based on spatial tree with duplicate points | |
| US9244935B2 (en) | Data encoding and processing columnar data | |
| KR100484137B1 (en) | Improved huffman decoding method and apparatus thereof | |
| CN100583649C (en) | Method and apparatus for encoding/decoding vector as well as flow medium player | |
| Safavi et al. | Sparsity‐aware adaptive block‐based compressive sensing | |
| CN101594152B (en) | LDPC code decoding method for realizing simultaneous operation of horizontal operation and vertical operation | |
| CN109075805A (en) | Realize the device and method of polarization code | |
| KR20180077060A (en) | Method and apparatus for encoding and decoding lists of pixels | |
| CN108200439B (en) | Method for improving digital signal conversion performance and digital signal conversion method and device | |
| KR101627000B1 (en) | Method for treating digital data | |
| CN106330204B (en) | A kind of data processing method and device | |
| CN102547260A (en) | Decoding method of adaptive variable length coding based on context and system thereof | |
| Golin et al. | Encoding 2D range maximum queries | |
| Wu et al. | Application of quantisation‐based deep‐learning model compression in JPEG image steganalysis | |
| CN102315899A (en) | Quasi-cyclic decoding method for low-density parity check code | |
| KR101577848B1 (en) | Method for counting vectors in regular point networks | |
| CN107992555B (en) | Method for storing and reading raster data | |
| CN107018287A (en) | The method and apparatus for carrying out noise reduction to image using video epitome |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20170428 Address after: 610000 Chengdu, Wuhou District Wuhou New Town Management Committee No. Hing Road, No. two, No. 17 Applicant after: Chengdu Fourier Electronic Technology Co., Ltd. Applicant after: Shenzhen SDG Information Co., Ltd. Address before: 610000 Chengdu, Wuhou District Wuhou New Town Management Committee No. Hing Road, No. two, No. 17 Applicant before: Chengdu Fourier Electronic Technology Co., Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |