CN106354581A - Cyclic redundancy check method and multi-core processor - Google Patents
Cyclic redundancy check method and multi-core processor Download PDFInfo
- Publication number
- CN106354581A CN106354581A CN201610698751.4A CN201610698751A CN106354581A CN 106354581 A CN106354581 A CN 106354581A CN 201610698751 A CN201610698751 A CN 201610698751A CN 106354581 A CN106354581 A CN 106354581A
- Authority
- CN
- China
- Prior art keywords
- message segment
- sub
- kernel
- crc
- code
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/08—Error detection or correction by redundancy in data representation, e.g. by using checking codes
- G06F11/10—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's
- G06F11/1004—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's to protect a block of data words, e.g. CRC or checksum
Abstract
The invention discloses a cyclic redundancy check method and a multi-core processor. According to the embodiment, the CRC method can improve the CRC treatment efficiency of the multi-core processor. The CRC method comprises the steps that an original message segment is cut into P sub- message segments by a first kernel; the cutting information of the P sub-message segments is transmitted to Q corresponding kernels by the first kernel; the corresponding sub-message segments are obtained by the Q kernels according to the corresponding cutting information of the sub-message segments; the check information of the corresponding sub-message segments is determined and transmitted to the first kernel by the Q kernels; according to the check information of the P sub-message segments, the CRC code of the original message segment is determined by the first kernel. The invention further provides a related multi-core processor.
Description
Technical field
The application is related to the communications field, more particularly, to a kind of cyclic redundancy check method and polycaryon processor.
Background technology
Cyclic redundancy check (CRC) (English: cyclic redundancy check, abbreviation: crc) is former using division and remainder
The method to carry out error detection (English: error detecting) for the reason.The network equipment, such as base station controller, equipment of the core network
Deng in communication process, what a crc constant transmitting terminal and receiving terminal arrange.Transmitting terminal is treated using this crc constant as divisor
The former message segment sending carries out complementation, and the remainder obtaining is the crc code of former message segment.Transmitting terminal is by calculated former report
The crc code of civilian section is attached to former message segment end and is sent to receiving terminal.Receiving terminal recalculates former report after receiving former message segment
The crc code of civilian section, and the crc code that calculated crc code is sent with transmitting terminal compares.If two crc codes different then it is assumed that
Communication occurs in that mistake.
Crc verification to be realized by the processor of the network equipment typically in upper-layer protocol.But processor at this stage enters
It is less efficient that row crc verifies, and especially cannot meet application in the scene of the mass data such as video communication in actual applications
Require.
Content of the invention
This application provides a kind of cyclic redundancy check method, for lifting the crc treatment effeciency of polycaryon processor.This Shen
Please in additionally provide correlation polycaryon processor.
The first aspect of the application provides a kind of crc method it is adaptable to polycaryon processor.Wherein, this polycaryon processor
Multiple kernels including the first kernel.The former message segment that pending crc verifies is cut into p sub- message by the first kernel
Section, this p sub- message segment is followed successively by the first sub- message segment to the sub- message segment of pth according to order from front to back, p be more than or equal to
2 positive integer.First kernel determines the q kernel for processing this p sub- message segment, the sub- message that each kernel is processed with it
Section is corresponding.One of kernel can correspond to one or more sub- message segments, but a sub- message segment can only correspond in one
Core, q≤p.First kernel after determining the corresponding kernel of the sub- message segment of p, by the cutting information of this p sub- message segment respectively
It is sent to corresponding kernel, cutting information is used for each sub- message segment of unique determination, for example, can be the address letter of each sub- message segment
Breath or positional information in former message segment for each sub- message segment or other types of information.This q kernel obtains
After the cutting information of corresponding sub- message segment, according to this cutting information, and the former message segment in internal memory, obtain corresponding sub- report
Civilian section.Then this q kernel determines the check information of itself corresponding sub- message segment respectively, and by itself corresponding sub- message segment
Check information be sent to the first kernel.The verification letter of corresponding sub- message segment is all have sent in this q kernel to the first kernel
After breath, the first kernel, according to the check information of this p sub- message segment, is calculated the crc code of former message segment.By such side
Method, the first kernel does not need to calculate the crc code of entirely former message segment, but is assigned with q kernel to calculate the school of sub- message segment
Test information.First kernel only needs to carry out the fortune of the check information of simply former message segment cutting operation and p sub- message segment
Calculation operates.Thus the calculating of the crc code of former message segment operation is shared to undertake to q kernel, greatly reduce the
The amount of calculation of one kernel, has equalized the load of each kernel in processor, is conducive to lifting the efficiency that processor carries out crc verification.
Optionally, in this p sub- message segment, the cutting information of the i-th sub- message segment includes the i-th sub- message segment with respect to former report
Side-play amount offset_i of civilian section and length length_i of the i-th sub- message segment, 1≤i≤p.I-th sub- message segment is corresponding
According to offset_i and length_i, kernel can determine that the i-th sub- message segment is the offset_i byte from former message segment
Start, the sequence of message of a length of length_i byte.
Optionally, this q kernel determines the check information of corresponding sub- message segment, is specifically as follows this q kernel and calculates
The check code of equal value of corresponding sub- message segment.Check code wherein of equal value is used for representing fills j 0 at the end of the i-th sub- message segment
The crc code of the message segment obtaining afterwards, the length length_sum-offset_i length_i of the former message segment of j=.First kernel
Determine the crc code of former message segment according to the check information of p sub- message segment, be specifically as follows the verification of this p sub- message segment
Information carries out XOR, obtains the crc code of former message segment.So only need to carry out simply former message segment cutting operation and p
The xor operation of the check code of equal value of sub- message segment, further reduces the amount of calculation of the first kernel.
Optionally, this q kernel calculates the check code of equal value of corresponding sub- message segment, is specifically as follows in this q and assesses
Calculate the crc code of corresponding sub- message segment, then the crc code according to corresponding sub- message segment calculates the equivalence of corresponding sub- message segment
Check code, to reduce the number of times of tabling look-up of the direct check code of equal value calculating corresponding sub- message segment.
Optionally, this q kernel can determine the crc code of corresponding sub- message segment by searching a crc table.Wherein
One crc table lists the crc code of the sequence of single byte.
Optionally, this q kernel can search the 2nd crc table according to the crc code of corresponding sub- message segment, obtains corresponding
The check code of equal value of sub- message segment.The crc code of multiple single-bit ordered sequences, the effective sequence of single-bit is listed in 2nd crc table
Arrange for representing that the first bit is the sequence that 1 other bits are 0.
Optionally, this q kernel searches the 2nd crc form, can be specifically that this q kernel determines corresponding sub- report respectively
In the crc code of civilian section, each bit corresponding single-bit ordered sequence, then searches each bit pair in the 2nd crc table
The crc code of corresponding for each bit single-bit ordered sequence is finally also carried out by the crc code of the single-bit ordered sequence answered
Or, just having obtained the check code of equal value of corresponding sub- message segment.So obtain corresponding son compared to a direct crc form of searching
The check code of equal value of message segment, can reduce number of times of tabling look-up.
Optionally, also include internal memory in polycaryon processor, this q kernel obtains according to the cutting information of corresponding sub- message segment
Take corresponding sub- message segment, can be specifically former message segment to be obtained from internal memory by shared drive mode, then according to correspondence
The cutting information of sub- message segment and former message segment obtain corresponding sub- message segment.
The application second aspect provides a kind of polycaryon processor, and this polycaryon processor is multiple including the first kernel
Kernel.Wherein, the first kernel is used for: the former message segment of pending crc verification is cut into p sub- message segment, this p sub- message
Section is followed successively by the first sub- message segment to the sub- message segment of pth according to order from front to back, and p is the integer more than or equal to 2.Determine and use
In q kernel processing this p sub- message segment, wherein each kernel is corresponding with the sub- message segment that it is processed, q≤p.By this p
The cutting information of sub- message segment is sent to corresponding kernel.This q kernel is used for: in the cutting obtaining corresponding sub- message segment
After information, according to cutting information, obtain corresponding sub- message segment.After obtaining corresponding sub- message segment, calculate corresponding sub- report
Then the check information of corresponding sub- message segment is sent to the first kernel by the check information of civilian section.First kernel is additionally operable to: receives
The check information of the sub- message segment of correspondence that each kernel sends in this q kernel, obtains the check information of this p sub- message segment.
Then the check information according to this p sub- message segment, calculates the crc code of former message segment.In the polycaryon processor that the application provides,
The arithmetic operation that first kernel only needs to carry out the check information of simply former message segment cutting operation and p sub- message segment is
Can.Thus the calculating operation of the crc code of former message segment is shared to undertake to q kernel, greatly reduce the first kernel
Amount of calculation, in processor, the load of each kernel more equalizes, and the efficiency carrying out crc verification is higher.
Optionally, in this p sub- message segment, the cutting information of the i-th sub- message segment includes the i-th sub- message segment with respect to former report
Side-play amount offset_i of civilian section and length length_i of the i-th sub- message segment, 1≤i≤p.I-th sub- message segment is corresponding
Kernel specifically for: according to offset_i and length_i, determine that the i-th sub- message segment is the offset_i from former message segment
Byte starts, the sequence of message of a length of length_i byte.
Optionally, this q kernel is specifically with calculating the check code of equal value of corresponding sub- message segment.Check code wherein of equal value is used
In the crc code representing the message segment obtaining after the end of the i-th sub- message segment fills j 0, the length of the former message segment of j=
length_sum-offset_i–length_i.First kernel is different specifically for carrying out the check information of this p sub- message segment
Or, obtaining the crc code of former message segment.Such first kernel only needs to carry out simply former message segment cutting operation and p son
The xor operation of the check code of equal value of message segment, can reduce the amount of calculation of the first kernel further.
Optionally, this q kernel specifically for calculate corresponding sub- message segment crc code, then according to corresponding sub- report
The crc code of civilian section calculates the check code of equal value of corresponding sub- message segment, to reduce the direct equivalence calculating corresponding sub- message segment
The number of times of tabling look-up of check code.
Optionally, this q kernel is additionally operable to, by searching a crc table, determine the crc code of corresponding sub- message segment.Wherein
First crc table lists the crc code of the sequence of single byte.
Optionally, the crc code that this q kernel is additionally operable to according to corresponding sub- message segment searches the 2nd crc table, is corresponded to
Sub- message segment check code of equal value.The crc code of multiple single-bit ordered sequences is listed, single-bit is effective in 2nd crc table
Sequence is used for representing that the first bit is the sequence that 1 other bits are 0.
Optionally, to be additionally operable to each bit in the crc code determine respectively corresponding sub- message segment corresponding for this q kernel
Single-bit ordered sequence, then searches the crc code of each bit corresponding single-bit ordered sequence, in the 2nd crc table
Afterwards the crc code of corresponding for each bit single-bit ordered sequence is also carried out or, just obtained corresponding sub- message segment etc.
Valency check code.So search, compared to direct, the check code of equal value that a crc form obtains corresponding sub- message segment, can reduce and look into
Tabular order number.
Optionally, also include internal memory in polycaryon processor, this q kernel is additionally operable to by shared drive mode from internal memory
Obtain former message segment, then the cutting information according to corresponding sub- message segment and former message segment obtain corresponding sub- message segment.
Brief description
One example structure figure of polycaryon processor that Fig. 1 provides for the application;
One embodiment flow chart of crc method that Fig. 2 provides for the application.
Specific embodiment
This application provides a kind of cyclic redundancy check (CRC) crc method, the crc treatment effeciency of polycaryon processor can be lifted.
Present invention also provides the polycaryon processor of correlation, it is described respectively below.
Processor kernel (English: core, also known as: die) be also called the core of processor, be most important group of processor
Become part.One or more kernels can be deployed with single processor.If kernel number m of deployment is not less than 2 in processor,
Then processor is referred to as polycaryon processor.Fig. 1 is only described taking m=8 as a example.As shown in figure 1, the eight of polycaryon processor 100
Kernel is divided into in the first kernel 101, the second kernel 102, the 3rd kernel 103, the 4th kernel 104, the 5th kernel the 105, the 6th
Core 106, the 7th kernel 107 and the 8th kernel 108.Kernel based on wherein first kernel, is responsible for coordinating verification message in other
Processed.The internal memory 109 for data storage is also included, such as Double Data Rate synchronous DRAM in polycaryon processor
(English: double data rate synchronous dynamic random access memory, abbreviation: ddr
Sdram, referred to as: ddr).Wherein, each kernel and described internal memory are attached in the way of bus, and each kernel passes through altogether
The mode enjoying internal memory accesses the data in described internal memory.
The embodiment one of the crc method of the application offer, on the basis of the polycaryon processor shown in below in Fig. 1, is provided,
Its basic procedure refers to Fig. 2, comprising:
201st, former message segment is cut into p sub- message segment by the first kernel.
If there being the pending crc of former message segment to verify in internal memory, this former message segment is cut into p sub- message by the first kernel
Section, p is the positive integer more than or equal to 2.This p sub- message segment is followed successively by the first sub- message segment x according to order from front to back1、
Second sub- message segment x2... pth sub- message segment xp.
For example, if former message segment is 9 byte message segment message segment a1a2a3b1b2b3c1c2c3, the first kernel is by former report
Civilian section cuts into 3 sub- message segment x1、x2、x3, then x1=a1a2a3, x2=b1b2b3, x3=c1c2c3.
Optionally, the first kernel can averagely cut this former message segment as far as possible.For example: in a kind of average cutting
In method, former message segment can be cut into isometric p sub- message segment by the first kernel.If former message segment unnecessary go out k byte
Can not be divided exactly by p, then the first kernel shares this k byte in k sub- message segment.If former segment length is 500 words
Section, p=8, then the sub- message segment that cutting obtains sub- message segment that 4 length is 62 and 4 length is 63.Another kind of average
In cutting method, if former message segment unnecessary go out k byte can not be divided exactly by p, this k byte is assigned to wherein by the first kernel
On one sub- message segment.If former segment length is 500 bytes, p=8, then cutting obtains 7 length is that the sub of 62 bytes is reported
Civilian section and the sub- message segment that 1 length is 66.
First kernel can obtain the cutting information of each sub- message segment after cutting former message, cutting information is used for uniquely true
Fixed each sub- message segment, is specifically as follows the address information or each sub- message segment of each sub- message segment position letter in former message segment
Breath or other types of information.For example: cutting information can include each sub- message segment with respect to former message segment
Side-play amount, and the length of each sub- message segment.As long as each kernel in processor obtains the cutting information of sub- message segment, with regard to energy
Enough the content in sub- message segment is gone out according to former message segment and this cutting information reverting.For the ease of description, use in the application
Offset_i represents xiSide-play amount (i.e. x with respect to former message segmentiOriginal position apart from the original position of former message segment away from
From), represent x with length_iiLength, represent the length of former message segment, 1≤i≤p with length_sum.
It is worthy of note that, the sub- message segment after cutting also can be reduced into former message segment.For example, can be in sub- report
The end zero padding of civilian section is so that alignd with the end of former message segment in the end of sub- message segment, that is, in xiEnd fill one complete
Zero sequences yi, wherein yiByte number be equal to: length_sum-offset_i length_i.For the ease of description, the application
In use xiyiRepresent yiImmediately xiSequence of message (i.e. the after zero padding i-th sub- message segment) afterwards, by xiyiCrc code be referred to as i-th
The check code of equal value of sub- message segment.It should be understood that each sub- message segment after zero padding is carried out XOR can get former message segment,
I.e. former message segment=x1y1^x2y2^……^xiyi……^xpyp.Wherein need to particularly point out, ypDigit be 0, i.e. xpyp=
xp.
For example, if former message segment is a1a2a3b1b2b3c1c2c3, x1=a1a2a3, x2=b1b2b3, x3=c1c2c3, then
x1y1=a1a2a3000000, x2y2=b1b2b3000, x3y3=c1c2c3.Former message segment a1a2a3b1b2b3c1c2c3=
x1y1^x2y2^x3y3=a1a2a30 0 0 0 0 0^b1b2b30 0 0^c1c2c3.
202nd, the cutting information of described p sub- message segment is sent to corresponding q kernel by the first kernel.
First kernel determines the q kernel for processing this p sub- message segment.For the ease of statement, will use in the application
It is referred to as the corresponding kernel of this sub- message segment in the kernel processing certain sub- message segment, the sub- message segment of certain kernel processes is referred to as
The corresponding sub- message segment of this kernel.Wherein, a sub- message segment can only correspond to a kernel.Many individual sub- message segments can correspond to same
One kernel is it is also possible to correspond to different kernels.This p sub- message segment corresponds to q kernel altogether.
Optionally, this p sub- message segment can be one-to-one relation with this q kernel, i.e. p=q.
Optionally, the first kernel can be one of this q kernel it is also possible to be not belonging to this q kernel.Particularly,
If the first kernel is one of this q kernel, the first kernel can not execute and send corresponding sub- message segment to oneself
Step.
Optionally, this q kernel can be the part kernel in polycaryon processor, but in order to make full use of polycaryon processor
Process resource, this q kernel can also be whole kernels of polycaryon processor, i.e. q=m.
After first kernel determines the corresponding kernel of the sub- message segment of this p, by the cutting information of this p sub- message segment respectively
It is sent to corresponding kernel.
203rd, q kernel is according to the corresponding sub- message segment of cutting acquisition of information of corresponding sub- message segment.
Each kernel in processor can access the former message segment in internal memory by way of shared drive.This q kernel
After obtaining the cutting information of corresponding sub- message segment, according to this cutting information, and the former message segment in internal memory, obtain and correspond to
Sub- message segment.
For example: assume that the second kernel corresponds to the second sub- message segment x2, then the second kernel receive x at the first kernel2's
Cutting information offset_2 and length_2, and obtain former message segment by way of shared drive, it is then determined that x2From former
The offset_2 byte of message segment starts, a length of length_2 byte.Such second kernel just obtains x2.
204th, q kernel determines the check information of corresponding sub- message segment and is sent to the first kernel.
By step 201~203, this q kernel has obtained each self-corresponding sub- message segment.Then this q kernel
Determine the check information of itself corresponding sub- message segment respectively, and the check information of itself corresponding sub- message segment is sent to
One kernel.
Wherein, the check information of sub- message segment is according to the calculated information of sub- message segment, for the first kernel according to
The crc code to calculate former message segment for the check information of every sub- message segment.The concrete form of check information and computational methods will be
Describe in detail in latter embodiments, do not limit in the present embodiment.
Particularly, if the first kernel is one of this q kernel, the first kernel can not execute right to oneself transmission
The step of the check information of sub- message segment answered.
205th, the first kernel, according to the check information of p sub- message segment, determines the crc code of former message segment.
After this q kernel all have sent the check information of corresponding sub- message segment to the first kernel, the first kernel is with regard to energy
Enough obtain the check information of this p sub- message segment.Then the first kernel, according to the check information of this p sub- message segment, calculates
Crc code to former message segment.
The concrete grammar of the crc code that the first kernel calculates former message segment according to the check information of p sub- message segment will be rear
Describe in detail in the embodiment in face, do not limit herein.
In the present embodiment, former message segment is cut into p sub- data segment by the first kernel, and cutting this p sub- message segment
The information of cutting is sent to corresponding q kernel.The check information that this q kernel is calculated this p sub- message segment is sent to first
Kernel, the first kernel obtains the crc code of former message segment according to the check information of this p sub- message segment.Permissible by said method
Find out, in the present embodiment, the first kernel does not need to calculate the crc code of entirely former message segment, but is assigned with q kernel to calculate
The check information of sub- message segment.First kernel only needs to carry out simply former message segment cutting operation and p sub- message segment
The arithmetic operation of check information.Thus the calculating operation of the crc code of former message segment is shared to undertake to q kernel,
Greatly reduce the amount of calculation of the first kernel, equalized the load of each kernel in processor, be conducive to lifting processor to carry out crc
The efficiency of verification.
Discuss in above embodiment the application offer crc method basic procedure, below will to step 204~
The check information of the sub- message segment of calculating mentioned in 205, and the crc calculating former message segment according to the check information of sub- message segment
The concrete grammar of code is described:
In embodiments herein two, q kernel described by step 204 determines the verification letter of corresponding sub- message segment
Breath, is specifically as follows the check code of equal value that this q kernel calculates corresponding sub- message segment.Kernel calculates the school of equal value of sub- message segment
The method testing code has a lot, will describe in detail, do not limit herein in embodiment below.This q kernel is by corresponding sub- message
Section check information be sent to and be sent to the first kernel, be specifically as follows by this q kernel by corresponding sub- message segment etc.
Valency check code is sent to the first kernel.: assume that the second kernel corresponds to the second sub- message segment x taking the second kernel as a example2, then in second
Core is according to the x getting in step 2032, calculate x2Check code of equal value (i.e. x2y2Crc code) as x2Check information, and
By x2Check information be sent to the first kernel.The first kernel described by step 205 is according to the check information of p sub- message segment
Determine the crc code of former message segment, be specifically as follows and the check information of this p sub- message segment is carried out XOR, obtain former message segment
Crc code.
In embodiment two, q kernel calculates the check code of equal value of corresponding sub- message segment respectively, and the first kernel is sub by this p
The check code of equal value of message segment carries out XOR, has just obtained the crc code of former message segment.As can be seen that the first kernel does not need to count
Calculate the crc code of entirely former message segment it is only necessary to carry out simply former message segment cutting operation and the school of equal value of p sub- message segment
Test the xor operation of code, further reduce the amount of calculation of the first kernel.
In the crc method that embodiment two provides, each kernel calculates the check code of equal value of corresponding sub- message segment, the first kernel
The check code of equal value of each sub- message segment is carried out XOR, obtains the crc code of former message segment.Wherein, the equivalence of each sub- message segment
The reason XOR of check code is equal to the crc code of former message segment is as follows:
In clear and definite first the application, "/" represents mould 2 division operation, and " ^ " represents XOR, and " * " represents modular two multiplication computing.If
Two sequence of message are respectively d1, d2, and crc constant is p.And:
R1 more than d1/p=q1, that is, the crc code of d1 is r1; (1.1)
R2 more than d2/p=q2, that is, the crc code of d2 is r2. (1.2)
Then can be obtained by (1.1) (1.2) conversion:
D1=(p*q1) ^r1; (1.3)
D2=(p*q2) ^r2. (1.4)
(1.3) and (1.4) are carried out XOR, obtain:
D1^d2=(p*q1) ^r1^ (p*q2) ^r2
=(p*q1) ^ (p*q2) ^r1^r2. (1.5)
The crc code of calculation equation (1.5) the right and left, wherein, because (p*q1) ^ (p*q2) can be divided exactly by p, therefore d1^
Crc code=the r1^r2 of d2. (1.6)
Understand, the crc code of former message segment can be obtained by the check code XOR of equal value of each sub- message segment according to (1.6),
The i.e. crc code=x of former message segment1y1Crc code ^x2y2Crc code ^ ... ^xiyiCrc code ... ^xpypCrc code.
In above embodiment, each kernel calculates the check code of equal value of corresponding sub- message segment respectively, and the first kernel will be each
The check code of equal value of sub- message segment carries out the crc code that XOR obtains former message segment.Each kernel calculating etc. is described further below
The method of valency check code.
Crc code by definition, is that to crc constant, message segment is carried out the remainder that complementation obtains.But remainder is grasped
Implement relatively complicated, adopt look-up table to calculate crc code to avoid remainder to operate more in Practical Project.In existing skill
In the look-up table of art, processor needs to safeguard a crc table, the crc table being 256 with existing conventional length in internal memory
As a example: a crc table safeguarded by processor, lists the crc code of 256 kinds of values of single byte in a crc table.Process
Device is according to a crc table, the crc code of the determination message segment of byte-by-byte.Specific determination method is:
(1) by low eight lt 8 of the crc code of f-1 byte,;
(2) high eight-bit of the crc code of f-1 byte and f byte are carried out XOR;
(3) in a crc table, search the result corresponding crc code of (2);
(4) result of (1), (3) is carried out XOR, obtain the crc code of f byte.
Wherein f is the byte number of positive integer and no more than message segment.And define when f value is 1, the crc code of the 0th byte
For 0.After processor performs step (1)~(4) successively to all bytes of message segment, just obtain the crc code of message segment.
To introduce look-up table at this stage only in the application, also may be used in Practical Project taking the crc table that length is 256 as a example
Be 128 or 512 using length crc table to carry out table lookup operation, do not limit in the application.
As can be seen that the look-up table of prior art need not carry out remainder operation from step (1)~(4), implementing
When relatively simple, it is possible to increase the crc computational efficiency of processor.Therefore in embodiments herein three, each kernel can
Calculate the check code of equal value of corresponding sub- message segment with the look-up table providing using prior art.Each kernel calculates corresponding sub- report
The flow process of the check code of equal value of civilian section can refer to step (1)~(4), do not repeat herein.
But, as can be seen that look-up table needs to consult one time first to each byte of message segment from step (3)
Crc table, therefore it is tabled look-up, number of times is equal with the byte number of message segment.And the x in the application1y1~xpypIn message segment, it is singly x1y1
The byte number of message segment is just identical with the byte number of former message segment.If therefore directly using look-up table, single determination x1y1Crc
The number of times of tabling look-up of execution required for code, just suitable with the number of times of tabling look-up needed for the crc code of the former message segment of direct determination, and determine
x1y1~xpypWhole crc codes then will carry out more table lookup operations, considerably increases the burden of processor.
Applicant finds through research, the x of the application1y1~xpypMessage segment has particularity: y1~ypIt is full null sequence.
This end is that the message segment of full null sequence can simplify when calculating crc code further.For this reason, the embodiment of the present application four
Additionally provide a kind of optimization method calculating check code of equal value it is adaptable to each kernel calculates the verification of equal value of corresponding sub- message segment
Code, will describe in detail to the method below:
In the application, first is not zero, the message segment that remaining position is zero is referred to as single-bit ordered sequence.Each report
Civilian section all can split into the XOR of the effective message segment of one or more single byte, and 16 message segments with two bytes are
Example: message segment 1,001 0,000 1001 0000=1000,0,000 0000 000,0^1 0,000 0000 0000^1000 0000^
1 0000.
Assume xiyiThe XOR of multiple single-bit ordered sequences, then x are split intoiyiCrc code be equal to it and split and obtain
The XOR of the crc code of the effective message segment of multiple single byte.Still: message segment 1001 taking 16 message segments in the preceding paragraph as a example
The crc code ^ of the crc code ^1 0,000 0,000 0000 of 0000 1,001 0000 crc code=1,000 0,000 0,000 0000
The crc code of 1000 0000 crc code ^1 0000.
On the premise of sequence length has the upper limit, single-bit ordered sequence is easy to exhaustion, therefore optionally, this Shen
Please in the crc code of whole single-bit ordered sequences is calculated in advance, record, in the 2nd crc table, and is saved at multinuclear
In the internal memory of reason device.It needs to be determined that single-bit ordered sequence crc code when it is only necessary to make a look up in the 2nd crc table
Can.
Table 1 gives an instantiation of the 2nd crc table.In Practical Project, the 2nd crc table can also be other shapes
Formula, if its with table 1 general record the crc code of each single-bit ordered sequence.The length of the sequence of record in 2nd crc table
Degree can determine according to actual needs, do not limit herein.
Table 1
Still taking 16 message segments above as a example: to calculate the crc code of message segment 1,001 0,000 1,001 0000,
Then search respectively in Table 1 the crc code of sequence 1,000 0,000 0,000 0000, the crc code of sequence 1 0,000 0,000 0000,
The crc code of sequence 1,000 0000 and the crc code of sequence 1 0000, then calculate 1,001 0,000 1,001 0000 crc code
The crc code ^1 0000 of the crc code ^1000 0000 of=1,000 0,000 0,000 0000 crc code ^1 0,000 0,000 0000
Crc code.
It is pointed out that a kind of look-up table being only the application is provided from the angle of principle of discussing above enters
Row explain, in actual applications, do not need message segment really splits into multiple single-bit ordered sequences it is only necessary to according to
The value of each bit of message segment consults a crc table.For example to determine message segment 1,001 0,000 1,001 0000
Crc code, then due to message segment right number the 5th, the 8th, the 13rd and the 16th be 1, then in actual applications, only need
To search the 5th crc code, the 8th crc code, the 13rd crc code and the 16th crc code in Table 1.
If it should be understood that determining x by searching the 2nd crc tableiyiCrc code, then search the 2nd crc table number of times
With xiIn non-zero bit number equal.And in embodiment three, search number of times and the x of a crc tableiyiByte number equal.In xi
Shorter and yiWhen longer, determine x by searching the 2nd crc tableiyiCrc code can effectively reduce number of times of tabling look-up.Example
As if xiThere are 2 bytes, wherein have 10 non-zero bits, and yiThere are 20 bytes.The method that then embodiment three provides needs
Consult a 2+20=22 crc table and just can obtain xiyiCrc code, and example IV only needs to just consult 10 the 2nd crc tables
X can be obtainediyiCrc code.And only need in example IV carry out XOR to checking result, for comparing embodiment three
Arithmetic operation is more simple, can lift the crc treatment effeciency of processor.
But consult the 2nd crc table due to what example IV needed bit-by-bit, therefore in xiWhen longer, example IV
Number of times of tabling look-up can be far more than embodiment three.For example, if xiEach bit in 10 bytes, and each byte is had all to be not zero,
Then need to consult 10 × 8=80 time the 2nd crc table by example IV, and the look-up table of embodiment three only needs to consult 10 times
One crc table.
For above-mentioned situation, the application has carried out improvement and has obtained embodiment five, for reducing lookup second to example IV
The number of times of crc table.Its concrete grammar is as follows:
A () kernel searches a crc table according to the look-up table that embodiment three provides, obtain xiCrc code ci, wherein search
The number of times of the first crc table is equal to xiByte number, ciLength be not more than crc constant subtract one;
B () kernel is by ciyiAs the sequence of message of crc code to be asked for, and the look-up table being provided according to example IV is searched
2nd crc table, obtains ciyiCrc code, ciyiCrc code i.e. be equal to xiyiCrc code.Wherein search the number of times of the 2nd crc table
Equal to ciNon-zero bit position number.Due to ciLength shorter, therefore search the 2nd crc table number of times less.
The total degree that step (a) (b) is tabled look-up is equal to xiByte number+ciNon-zero bit number.Secondary with tabling look-up in example IV
Number is equal to xiIn non-zero bit number compare, in xiUnder the more scene of longer and nonzero digit, the number of times of tabling look-up of embodiment five is far few
Number of times of tabling look-up in example IV.Number of times is equal to x with tabling look-up in embodiment threeiyiByte number compare, in yiSlightly longer (such as big
In two bytes) scene under, the number of times of tabling look-up of embodiment five is far fewer than the number of times of tabling look-up of embodiment three.
In step (b), kernel calculates ciyiCrc code obtain xiyiCrc code, ciyiCrc code be equal to xiyiCrc code
Reason is as follows:
Known ciFor xiCrc code, if crc constant be p, yiByte number be n, then xiC more than/p=q3iIt may be assumed that
xi=(p*q3) ^ci; (2.1)
(2.1) equation the right and left is moved to left simultaneously and obtains after n byte (i.e. 8n bit):
xiyi=[(p*q3) yi]^(ciyi). (2.2)
By (2.2) the right and left simultaneously divided by crc constant p, obtain:
xiyi/ p={ [(p*q3) yi]^(ciyi)}/p. (2.3)
Due to [(p*q3) yi] can be divided exactly by p, therefore { [(p*q3) yi]^(ciyi) remainder of/p and ciyiThe remainder phase of/p
Deng.Therefore, xiyiCrc code be equal to ciyiCrc code.
The crc method of the application offer is provided in above embodiment.The application will be introduced in embodiment six below
The polycaryon processor for realizing above-mentioned crc method providing.
Multiple kernels including the first kernel for the polycaryon processor that the embodiment of the present application six provides, can also include using
In the internal memory of data storage, concrete structure can refer to Fig. 1.Wherein, Fig. 1 only taking have the processor of eight kernels is as a example carried out
Description, polycaryon processor can also have less or more kernels in actual applications, do not limit in the application.
Wherein, the first kernel 101 specifically for: when there being former message segment pending crc verification, by this former message segment cutting
Become p sub- message segment, p is the integer more than or equal to 2.This p sub- message segment is followed successively by the first son according to order from front to back
Message segment x1To pth sub- message segment xp.First kernel 101 determines in this p sub- message segment the corresponding kernel of every sub- message segment,
And the cutting information of this p sub- message segment is sent respectively to corresponding kernel.Wherein this p sub- message segment corresponds to q kernel altogether,
q≤p.First kernel 101, after determining the corresponding kernel of the sub- message segment of p, the cutting information of this p sub- message segment is sent
To corresponding kernel.The side-play amount that each sub- message segment is with respect to former message segment can be included in the cutting information of each sub- message segment,
And the length of each sub- message segment.
This q kernel is believed in the cutting obtaining corresponding sub- message segment specifically for: each kernel in this q kernel
After breath, according to cutting information, obtain corresponding sub- message segment.Each kernel in q kernel obtains corresponding sub- message segment
Afterwards, calculate the check information of corresponding sub- message segment, then the check information of corresponding sub- message segment is sent to the first kernel
101.
First kernel 101 is additionally operable to: receives the verification letter of the sub- message segment of correspondence that each kernel sends in this q kernel
Breath, obtains the check information of this p sub- message segment.Then the check information according to this p sub- message segment, calculates former message segment
Crc code.
In the present embodiment, former message segment is cut into p sub- data segment by the first kernel 101, and by this p sub- message segment
Cutting information is sent to corresponding q kernel.The check information that this q kernel is calculated this p sub- message segment is sent to the
One kernel 101, the first kernel 101 obtains the crc code of former message segment according to the check information of this p sub- message segment.By above-mentioned
Method can be seen that the first kernel 101 in the present embodiment does not need to calculate the crc code of entirely former message segment, but will calculate former
The task of the crc code of message segment distributes to q kernel.In the polycaryon processor that the present embodiment provides, the first kernel 101 only needs to
Carry out the arithmetic operation of the check information of simply former message segment cutting operation and p sub- message segment.Thus will be former
The calculating operation of the crc code of message segment shares to undertake to q kernel, greatly reduces the amount of calculation of the first kernel 101, processes
In device, the load of each kernel more equalizes, and the efficiency carrying out crc verification is higher.
Optionally, the x in this p sub- message segmentiCutting information include: xiOriginal position with respect to former message segment
Side-play amount offset_i, and xiLength length_i, 1≤i≤p.xiCorresponding kernel is used for according to xiCorresponding cutting letter
Breath, obtains x from the former message segment of shared driveiFor: from the beginning of the offset_i byte of former message segment, length is
The sequence of message of length_i.
Optionally, this q kernel is used for determining xiCheck information, particularly for determine xiCheck code of equal value.xi's
Check code of equal value is equal to: in xiEnd fill the sequences y of complete zeroiSequence of message x afterwardsiyiCrc code.Wherein yi
Byte number be equal to: length_sum-offset_i length_i.First kernel 101 is according to the verification of this p sub- message segment
Information determines the crc code of former message segment, specially the check code of equal value of this p sub- message segment is carried out XOR, obtains former message
The crc code of section.
Optionally, this q kernel is used for determining the check code of equal value of corresponding sub- message segment, specifically could be for first really
The crc code of fixed corresponding sub- message segment, further according to the crc code of corresponding sub- message segment, determines the verification of equal value of corresponding sub- message segment
Code.
Optionally, each kernel in this q kernel specifically for: determine corresponding sub- message by searching a crc table
The crc code of section.The crc code of the sequence of message that multiple length are a byte is listed in wherein the first crc table.Then each kernel
2nd crc table is searched: each non-zero bit position corresponding single-bit ordered sequence of the crc code of corresponding sub- message segment
Crc code.Wherein, list the crc code of multiple single-bit ordered sequences in the 2nd crc table, single-bit ordered sequence is used for representing
First bit be 1 and other bit be 0 sequence.Then each kernel will be right for each non-zero bit position of corresponding sub- message segment
The crc code of the single-bit ordered sequence answered carries out XOR, has just obtained the crc code of corresponding sub- message segment.
Optionally, polycaryon processor also includes internal memory 109, for preserving former message segment.This q kernel is additionally operable to: according to
The cutting information of corresponding sub- message segment, obtains corresponding sub- report in the former message segment from internal memory by way of shared drive
Civilian section.
The associated description of said apparatus can correspond to and be understood refering to the associated description of embodiment of the method part and effect,
This place does not do and excessively repeats.
It should be understood that disclosed system in several embodiments provided herein, apparatus and method are permissible
Realize by another way.For example, device embodiment described above is only schematically, for example, described unit
Divide, only a kind of division of logic function, actual can have other dividing mode when realizing, for example multiple units or assembly
Can in conjunction with or be desirably integrated into another system, or some features can be ignored, or does not execute.Another, shown or
The coupling each other discussing or direct-coupling or communication connection can be by some interfaces, the indirect coupling of device or unit
Close or communicate to connect, can be electrical, mechanical or other forms.
In addition, can be integrated in a processing unit in each functional unit in each embodiment of the application it is also possible to
It is that unit is individually physically present it is also possible to two or more units are integrated in a unit.Above-mentioned integrated list
Unit both can be to be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.
If described integrated unit is realized and as independent production marketing or use using in the form of SFU software functional unit
When, can be stored in a computer read/write memory medium.Based on such understanding, the technical scheme of the application is substantially
The part in other words prior art being contributed or all or part of this technical scheme can be in the form of software products
Embody, this computer software product is stored in a storage medium, including some instructions with so that a computer
Equipment (can be personal computer, server, or network equipment etc.) executes the complete of each embodiment methods described of the application
Portion or part steps.And aforesaid storage medium includes: u disk, portable hard drive, read only memory (rom, read-only
Memory), random access memory (ram, random access memory), magnetic disc or CD etc. are various can store journey
The medium of sequence code.
The above, above example only in order to the technical scheme of the application to be described, is not intended to limit;Although with reference to front
State embodiment the application has been described in detail, it will be understood by those within the art that: it still can be to front
State the technical scheme described in each embodiment to modify, or equivalent is carried out to wherein some technical characteristics;And these
Modification or replacement, do not make the essence of appropriate technical solution depart from the spirit and scope of the application each embodiment technical scheme.
Claims (14)
1. a kind of cyclic redundancy check (CRC) crc method is it is adaptable to polycaryon processor is it is characterised in that described polycaryon processor includes
No less than two kernels, methods described includes:
Former message segment to be verified is cut into p sub- message segment by the first kernel, and described p sub- message segment is followed successively by from front to back
First sub- message segment, the second sub- message segment ... the sub- message segment of pth, p is the positive integer more than or equal to 2;
Described first kernel determines the q kernel for processing described p sub- message segment respectively, and by individual for described p message segment
Cutting information be sent respectively to corresponding kernel, wherein, the sub- message segment that each kernel in described q kernel is processed with it
Corresponding, and in described q kernel, each kernel corresponds to one or more sub- message segments, every height report in described p sub- message segment
Civilian section corresponds to a kernel, q≤p;
Described q kernel, according to the cutting information of corresponding sub- message segment, obtains corresponding sub- message segment;
Described q kernel determines the check information of corresponding sub- message segment, and the check information of described p sub- message segment is sent
To described first kernel;
Described first kernel, according to the check information of described p sub- message segment, determines the crc code of described former message segment.
2. crc method according to claim 1 is it is characterised in that in described p sub- message segment, the cutting of the i-th sub- message segment
The information of cutting includes:
Side-play amount offset_i with respect to described former message segment for the described i-th sub- message segment and the length of described i-th sub- message segment
Degree length_i, 1≤i≤p.
3. crc method according to claim 2 is it is characterised in that described q kernel determines the school of corresponding sub- message segment
The information of testing includes:
Described q kernel determines the check code of equal value of corresponding sub- message segment, wherein, the verification of equal value of described i-th sub- message segment
Code is equal to: the crc code of the message segment obtaining after the described i-th sub- message segment end polishing j 0, former message segment described in wherein j=
Length length_sum-offset_i length_i;
Described first kernel, according to the check information of described p sub- message segment, determines that the crc code of described former message segment includes:
The check code of equal value of described p sub- message segment is carried out XOR by described first kernel, obtains the crc of described former message segment
Code.
4. crc method according to claim 3 it is characterised in that described q kernel determine corresponding sub- message segment etc.
Valency check code includes:
Described q kernel determines the crc code of corresponding sub- message segment;
Described q kernel, according to the crc code of described corresponding sub- message segment, determines the check code of equal value of corresponding sub- message segment.
5. crc method according to claim 4 is it is characterised in that described q kernel determines corresponding sub- message segment
Crc code includes:
Described q kernel passes through to search a crc table, determines the crc code of corresponding sub- message segment, wherein, a described crc table
Include the crc code of the sequence that multiple length are a byte;
Described q kernel, according to the crc code of described corresponding sub- message segment, determines the check code bag of equal value of corresponding sub- message segment
Include:
Described q kernel, according to the crc code of described corresponding sub- message segment, is searched the 2nd crc table, is obtained corresponding sub- message segment
Check code of equal value, wherein, described 2nd crc table includes the crc code of multiple single-bit ordered sequences, and described single-bit is effective
Sequence be the first bit be 1 and other bit be 0 sequence.
6. crc method according to claim 5 is it is characterised in that the described crc according to described corresponding sub- message segment
Code, searches the 2nd crc table, and the check code of equal value obtaining described corresponding sub- message segment includes:
Determine each bit corresponding single-bit ordered sequence of the crc code of corresponding sub- message segment;
In described 2nd crc table, search the crc code of described each bit corresponding single-bit ordered sequence;
Described in finding, the crc code of each bit corresponding single-bit ordered sequence carries out XOR, obtains described correspondence
Sub- message segment check code of equal value.
7. the crc method according to any one of claim 2 to 6 is it is characterised in that in described polycaryon processor also includes
Deposit, described internal memory is used for preserving described former message segment;
Described q kernel, according to the cutting information of corresponding sub- message segment, obtains corresponding sub- message segment and includes:
Described q kernel, according to the cutting information of corresponding sub- message segment, by shared drive mode, obtains from described internal memory
Corresponding sub- message segment.
8. a kind of polycaryon processor is it is characterised in that multiple kernels including the first kernel for the described polycaryon processor;
Described first kernel is used for: former message segment to be verified is cut into p sub- message segment, described p sub- message segment is in the past
Be followed successively by backward the first sub- message segment, the second sub- message segment ... the sub- message segment of pth, p is the positive integer more than or equal to 2;
Described first kernel is additionally operable to: determine the q kernel for processing described p sub- message segment respectively, and sub by described p
The cutting information of message segment is sent respectively to corresponding kernel, wherein, the son that each kernel in described q kernel is processed with it
Message segment corresponds to, and in described q kernel, each kernel corresponds to one or more sub- message segments, every in described p sub- message segment
Individual sub- message segment corresponds to a kernel, q≤p;
Described q kernel is used for: according to the cutting information of corresponding sub- message segment, obtains corresponding sub- message segment;
Described q kernel is additionally operable to: determines the check information of corresponding sub- message segment, and the verification by described p sub- message segment
Information is sent to described first kernel;
Described first kernel is additionally operable to: according to the check information of described p sub- message segment, determines the crc code of described former message segment.
9. polycaryon processor according to claim 8 is it is characterised in that in described p sub- message segment, the i-th sub- message segment
Cutting information include:
Side-play amount offset_i with respect to described former message segment for the described i-th sub- message segment and the length of described i-th sub- message segment
Degree length_i, 1≤i≤p.
10. polycaryon processor according to claim 9 is it is characterised in that described q kernel is additionally operable to: determines corresponding
The check code of equal value of sub- message segment, wherein, the check code of equal value of described i-th sub- message segment is equal to: at the described i-th sub- message segment end
The crc code of the message segment obtaining after tail polishing j 0, length length_sum-offset_i of former message segment described in wherein j=
length_i;
Described first kernel is additionally operable to: the check code of equal value of described p sub- message segment is carried out XOR, obtains described former message segment
Crc code.
11. polycaryon processors according to claim 10 are it is characterised in that described q kernel is additionally operable to:
Determine the crc code of corresponding sub- message segment;
According to the crc code of described corresponding sub- message segment, determine the check code of equal value of corresponding sub- message segment.
12. polycaryon processors according to claim 11 are it is characterised in that described q kernel is additionally operable to:
By searching a crc table, determine the crc code of corresponding sub- message segment, wherein, a described crc table includes multiple
Length is the crc code of the sequence of a byte;
According to the crc code of described corresponding sub- message segment, determine that the check code of equal value of corresponding sub- message segment includes:
According to the crc code of described corresponding sub- message segment, search the 2nd crc table, obtain the verification of equal value of corresponding sub- message segment
Code, wherein, described 2nd crc table includes the crc code of multiple single-bit ordered sequences, and described single-bit ordered sequence is first
Bit be 1 and other bit be 0 sequence.
13. polycaryon processors according to claim 12 are it is characterised in that described q kernel is additionally operable to:
Determine each bit corresponding single-bit ordered sequence of the crc code of corresponding sub- message segment;
In described 2nd crc table, search the crc code of described each bit corresponding single-bit ordered sequence;
Described in finding, the crc code of each bit corresponding single-bit ordered sequence carries out XOR, obtains described correspondence
Sub- message segment check code of equal value.
14. polycaryon processors according to any one of claim 9 to 13 it is characterised in that described polycaryon processor also
Including internal memory, described internal memory is used for preserving described former message segment;
Described q kernel is additionally operable to: according to the cutting information of corresponding sub- message segment, by shared drive mode, from described interior
Deposit the corresponding sub- message segment of middle acquisition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610698751.4A CN106354581B (en) | 2016-08-22 | 2016-08-22 | A kind of cyclic redundancy check method and multi-core processor |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610698751.4A CN106354581B (en) | 2016-08-22 | 2016-08-22 | A kind of cyclic redundancy check method and multi-core processor |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106354581A true CN106354581A (en) | 2017-01-25 |
CN106354581B CN106354581B (en) | 2019-08-23 |
Family
ID=57844263
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610698751.4A Active CN106354581B (en) | 2016-08-22 | 2016-08-22 | A kind of cyclic redundancy check method and multi-core processor |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106354581B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109787713A (en) * | 2017-11-15 | 2019-05-21 | 华为技术有限公司 | A kind of cyclic redundancy check (CRC) calculation method and device |
CN111061242A (en) * | 2018-10-16 | 2020-04-24 | 联合汽车电子有限公司 | Calibration system and method for motor controller of electric vehicle |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101656677A (en) * | 2009-09-18 | 2010-02-24 | 杭州迪普科技有限公司 | Message diversion processing method and device |
CN102739258A (en) * | 2011-04-14 | 2012-10-17 | 普天信息技术研究院有限公司 | Method and apparatus for calculating cyclic redundancy check code |
CN102801635A (en) * | 2012-08-30 | 2012-11-28 | 北京傲天动联技术有限公司 | Packet ordering method used in multi-core processor system |
CN102946297A (en) * | 2012-11-10 | 2013-02-27 | 华中科技大学 | Nested CRC (cyclic redundancy check) code generation method and device for data transmission error control |
-
2016
- 2016-08-22 CN CN201610698751.4A patent/CN106354581B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101656677A (en) * | 2009-09-18 | 2010-02-24 | 杭州迪普科技有限公司 | Message diversion processing method and device |
CN102739258A (en) * | 2011-04-14 | 2012-10-17 | 普天信息技术研究院有限公司 | Method and apparatus for calculating cyclic redundancy check code |
CN102801635A (en) * | 2012-08-30 | 2012-11-28 | 北京傲天动联技术有限公司 | Packet ordering method used in multi-core processor system |
CN102946297A (en) * | 2012-11-10 | 2013-02-27 | 华中科技大学 | Nested CRC (cyclic redundancy check) code generation method and device for data transmission error control |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109787713A (en) * | 2017-11-15 | 2019-05-21 | 华为技术有限公司 | A kind of cyclic redundancy check (CRC) calculation method and device |
CN111061242A (en) * | 2018-10-16 | 2020-04-24 | 联合汽车电子有限公司 | Calibration system and method for motor controller of electric vehicle |
Also Published As
Publication number | Publication date |
---|---|
CN106354581B (en) | 2019-08-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108121810A (en) | A kind of data duplicate removal method, system, central server and distributed server | |
CN112910806B (en) | Joint channel estimation and user activation detection method based on deep neural network | |
CN108366036B (en) | Sparse code multiple access system-oriented modulation codebook design method | |
CN108768656B (en) | Data verification method based on Hash algorithm | |
CN103716056B (en) | Data compression method, uncompressing data and equipment | |
WO2007075106A1 (en) | Fast low-density parity-check code encoder | |
AU2022203072B2 (en) | Node grouping method, apparatus and electronic device | |
CN109214512B (en) | Deep learning parameter exchange method, device, server and storage medium | |
CN114281793A (en) | Data verification method, device and system | |
WO2023040197A1 (en) | Cross-node communication method and apparatus, device, and readable storage medium | |
CN106354581A (en) | Cyclic redundancy check method and multi-core processor | |
CN110518945A (en) | A kind of MIMO detection method and device based on deep learning Yu SDR algorithm | |
CN109189792A (en) | A kind of parallel Hash tables look-up architecture and method | |
CN115395961A (en) | Data lossless compression and encrypted transmission method based on joint middleware | |
CN104270287A (en) | Message disorder detecting method and device | |
EP3840272A1 (en) | Resource mapping method and apparatus | |
CN111597139A (en) | Communication method, system, equipment and medium of GPU | |
CN116562373A (en) | Data mining method, device, equipment and medium | |
CN103516478B (en) | Method for sending information and equipment | |
CN107229635A (en) | A kind of method of data processing, memory node and coordinator node | |
CN109799483A (en) | A kind of data processing method and device | |
CN108833156A (en) | A kind of appraisal procedure and system of the simulation performance index for power telecom network | |
CN111966404B (en) | GPU-based regular sparse code division multiple access SCMA high-speed parallel decoding method | |
CN114039698A (en) | Data sending and receiving method, device and equipment based on multiple frequencies | |
CN112712167A (en) | Memory access method and system supporting acceleration of multiple convolutional neural networks |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 |