CN109308360A - Channel parameter finding method, device, storage medium and computer equipment - Google Patents
Channel parameter finding method, device, storage medium and computer equipment Download PDFInfo
- Publication number
- CN109308360A CN109308360A CN201710625321.4A CN201710625321A CN109308360A CN 109308360 A CN109308360 A CN 109308360A CN 201710625321 A CN201710625321 A CN 201710625321A CN 109308360 A CN109308360 A CN 109308360A
- Authority
- CN
- China
- Prior art keywords
- preset
- eye
- quantization parameter
- matrix
- parameter
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Dc Digital Transmission (AREA)
Abstract
The present invention relates to a kind of channel parameter finding methods, comprising: the register parameters of chip are set as to the point value of the preset of pre-emphasis parameters, read the eye high level and eye width values of current demand signal eye figure；The signal quantization parameter of preset is obtained according to the eye high level of reading and eye width values, and detects the signal quantization parameter of preset whether iteration is to extreme value；In the non-iteration of signal quantization parameter to the extreme value of preset, the gradient of preset is obtained according to the eye high level of reading and eye width values, and according to the gradient direction of preset and the mobile preset of preset step-length；In signal quantization parameter iteration to the extreme value of preset, confirm that the point value of matrix corresponding with extreme value is optimum parameter value.The present invention looks for optimum channel parameter along the maximum direction of gradient, can shorten the number of iteration, and then shortens and find the spent time, greatly improves the efficiency, has saved research and development of products cost.The present invention also provides a kind of channel parameter device for searching, storage medium and computer equipments.
Description
Technical field
The present invention relates to field of communication technology, more particularly to a kind of channel parameter finding method, device, storage medium and
Computer equipment.
Background technique
On the veneer of modern communications field switching equipment, private exchange chip and clock and data recovery core are generally comprised
Piece (i.e. CDR).After the completion of sheet making, the cabling of veneer, lamination, thermal parameter of veneer operation etc. influence exchange chip with
Communication channel parameters have determined that between CDR.But the product channel parameter of the different model of homologous series be it is unfixed, often
A product requires to find one's own channel parameter.Simultaneously as technological level, material level of processing limitation, together
Type product also has nuance.Although difference is little, it can make the channel parameter of each of same model individual that can omit
It is different.Therefore in the prior art, it needs largely to test with extracting parameter and finds optimal parameter, so that the channel parameter is full
The quality requirement of all product signals of communication of foot, which is referred to as to find shared parameter, therefore finds the optimization of a veneer
Parameter is very time-consuming and laborious thing, and therefore, it is highly important for optimizing the algorithm.
With the development of modern communication technology, the single channel communications rate between the exchange chip and CDR in industry at present
Have reached 10Gbps even 25Gpbs.So electric signal is transmitted on pcb board with this rate, the radio-frequency component decaying of signal is
It is very violent.Channel is equivalent to a low-pass filter, and signal arrival 10G-25G rate attenuation is very violent, so necessary
Signal is compensated.Compensation technique has preemphasis, postemphasises and balanced.So-called searching channel parameter, actually finds
Reasonable preemphasis is postemphasised and balance parameters, so that communication signal quality optimizes.Parameter after reasonable optimization is set
As shown in Figure 1, being that the signal of input (input) exports the eye figure of (output) after CDR in Fig. 1, eye figure is to measure high speed
One of the important means of SERDES signal quality, it is better that eyes open bigger i.e. signal quality.
It in the prior art, is all that optimal channel ginseng is found using all methods of traversal or ergodic part parameter
Number.Need to wait chip operation to stablize since register being arranged every time, and by sending and receiving data CRC (cyclic redundancy check,
Cyclic Redundancy Check) and eye figure come whether certificate parameter can be used, need to spend 10s or so, all traverse institutes
The time spent is 256*32*10, i.e. 81920s (22.8h).This time is single channel the time it takes.Record is every in this way
Block veneer can parameter, look for 10 such veneers, go the intersection of these parameters, at selected one group at most intermediate one
Parameter is considered relatively good, the shared parameter as all products.Obviously such method is very time-consuming.Therefore, it adopts
With the method traversed entirely the disadvantage is that efficiency is too low, finds suitable parameter and take a long time.In addition, with ergodic part parameter
Method is fixed DFE register, such as is set as 0, only goes to change CTLE register value, although traversal is greatly reduced in this way
The time of all parameters, but the parameter often found is not the parameter optimized, if when unlucky, 10 equipment notes
The parameter of record does not have intersection, i.e., without shared parameter.
Summary of the invention
Based on this, it is necessary in view of the above-mentioned problems, providing a kind of channel parameter finding method using gradient rise method, dress
It sets, storage medium and computer equipment.
A kind of channel parameter finding method, which comprises
The register parameters of chip are set as to the point value of the preset of pre-emphasis parameters, read the eye of current demand signal eye figure
High level and eye width values；The pre-emphasis parameters are matrix；
The signal quantization parameter of the preset is obtained according to the eye high level of reading and eye width values, and detects the preset
Signal quantization parameter whether iteration is to extreme value；
In the non-iteration of signal quantization parameter to the extreme value of the preset, obtained according to the eye high level of reading and eye width values
The gradient of the preset, and according to the gradient direction of the preset and the mobile preset of preset step-length；
In signal quantization parameter iteration to the extreme value of the preset, the matrix corresponding with the extreme value is confirmed
Point value is optimum parameter value.
A kind of channel parameter device for searching, comprising:
Module, the point value of the preset for the register parameters of chip to be set as to pre-emphasis parameters are obtained, reading is worked as
The eye high level and eye width values of front signal eye figure；The pre-emphasis parameters are matrix；
Detection module obtains the signal quantization parameter of the preset for the eye high level and eye width values according to reading, and
Detecting the signal quantization parameter of the preset, whether iteration is to extreme value；
Mobile module is high according to the eye of reading in the non-iteration of signal quantization parameter to the extreme value of the preset
Value and eye width values obtain the gradient of the preset, and described pre- according to the gradient direction of the preset and preset step-length movement
It sets up an office；
Confirmation module, for confirming and the extreme value pair in signal quantization parameter iteration to the extreme value of the preset
The point value for the matrix answered is optimum parameter value.
A kind of computer readable storage medium, is stored with that computer is executable to be referred on the computer readable storage medium
It enables, when the computer executable instructions are executed by processor, so that the processor executes following steps: by the deposit of chip
Device parameter setting is the point value of the preset of pre-emphasis parameters, reads the eye high level and eye width values of current demand signal eye figure；It is described pre-
Exacerbation parameter is matrix；The signal quantization parameter of the preset is obtained according to the eye high level of reading and eye width values, and detects institute
Stating the signal quantization parameter of preset, whether iteration is to extreme value；The preset the non-iteration of signal quantization parameter to extreme value
When, obtain the gradient of the preset according to the eye high level of reading and eye width values, and according to the gradient direction of the preset with
The mobile preset of preset step-length；In signal quantization parameter iteration to the extreme value of the preset, confirmation and the extreme value
The point value of the corresponding matrix is optimum parameter value.
A kind of computer equipment, including memory and processor store computer-readable instruction in the memory, institute
When stating computer-readable instruction and being executed by the processor, so that the processor executes following steps: by the register of chip
Parameter setting is the point value of the preset of pre-emphasis parameters, reads the eye high level and eye width values of current demand signal eye figure；The pre-add
Weight parameter is matrix；The signal quantization parameter of the preset is obtained according to the eye high level of reading and eye width values, and described in detection
Whether iteration is to extreme value for the signal quantization parameter of preset；In the non-iteration of signal quantization parameter to the extreme value of the preset,
The gradient of the preset is obtained according to the eye high level of reading and eye width values, and according to the gradient direction of the preset and is preset
The mobile preset of step-length；In signal quantization parameter iteration to the extreme value of the preset, confirm corresponding with the extreme value
The matrix point value be optimum parameter value.
Above-mentioned channel parameter finding method, device, storage medium and computer equipment, the algorithm risen using gradient, fastly
Fast iteration simultaneously searches out optimal channel parameter.The present invention looks for optimum channel parameter along the maximum direction of gradient, can be with
There is target to approach optimum channel parameter value with having direction；Meanwhile the number of iteration can be greatly simplified and shorten, and then shorten
The spent time is found, greatly improves the efficiency, has saved research and development of products cost.And compared to fixing a dimensional parameter,
Only change the method for a dimensional parameter, the present invention substantially increases the accuracy for finding extreme value.
Detailed description of the invention
Fig. 1 is the eye figure that input signal exports after CDR in the prior art；
Fig. 2 is the flow chart of channel parameter finding method in one embodiment.
Fig. 3 is the structural block diagram of channel parameter device for searching in one embodiment.
Fig. 4 is the structural block diagram of the acquisition module of channel parameter device for searching in one embodiment.
Fig. 5 is the structural block diagram of the detection module of channel parameter device for searching in one embodiment.
Fig. 6 is the structural block diagram of the mobile module of channel parameter device for searching in one embodiment.
Fig. 7 is the one-dimensional Gaussian Profile schematic diagram of channel parameter finding method in one embodiment.
Fig. 8 is the two-dimensional Gaussian Profile schematic diagram of channel parameter finding method in one embodiment.
Fig. 9 is to be superimposed unit direction matrix schematic diagram in one embodiment in channel parameter finding method.
Figure 10 is that channel parameter seeks trajectory in two-dimensional surface in channel parameter finding method in one embodiment.
Figure 11 is the calculating process of maximum directional derivative in channel parameter finding method in one embodiment.
Figure 12 is that there are the schematic diagrames of multiple channel parameter extreme points in space in one embodiment.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
In the present invention, as shown in Fig. 2, in one embodiment, providing a kind of channel parameter finding method, including step
S10-S40:
In step slo, the register parameters of chip are set as to the point value of the preset of pre-emphasis parameters X, reading is worked as
The eye high level and eye width values of front signal eye figure；The pre-emphasis parameters X is matrix；That is, setting declines for fighting channel first
The pre-emphasis parameters subtracted are matrix X；Wherein, the point value in the matrix X is made of the register parameters of chip；Wherein, described
Register parameters are including but not limited to the CTLE (Continuous Time Linear Equalizer) in exchange chip
The continuous time linear equalization parameter X of registerCTLE, DFE (Decision Feedback Equalizer) in exchange chip
The decision feedback equalization parameter X of registerDFE；In one embodiment, the deposit parameter only includes XCTLEWith XDFETwo ginsengs
Number, then the matrix X is two-dimensional matrix (XCTLE, XDFE).CTLE register 8bit has 256 parameters, and DFE register has
5bit has 32 parameters.In the present embodiment, pre-emphasis parameters X is set first, in the present embodiment, the preemphasis ginseng
Number is matrix X, that is, the matrix X can be for one-dimensional matrix, or two-dimensional matrix or multi-dimensional matrix.It is described pre-
To aggravate parameter be really to increase the radio-frequency component for issuing signal, fights decaying of the channel to radio-frequency component with this, and different pre-
Different gain curves can be corresponded to by aggravating parameter, and the steep of different gains curve also can be different；When the preemphasis of setting is joined
When several gain curves can be very good the attenuation of confrontation channel, recipient can correctly differentiate data.In fact,
In a certain range, gain curve steep it is high or low be all it is feasible, therefore every equipment can there are many channel parameters all
It can satisfy reception to require, but the present invention needs to choose a preferred channels parameter from the channel parameter that these meet the requirements.
If matrix X=X0For optimized parameter, then X=X0When the received bit error rate be it is minimum, the corresponding eye figure for receiving signal is best
's.And work as X=X0When+Δ x, X=X should be greater than by receiving the bit error rate0When the bit error rate, eye figure should slightly almost (be shown at this time
Eye is high, eye is wide smaller).As Δ x → 0, the bit error rate should be no better than X=X0When the bit error rate.
In one embodiment, it is two-dimensional matrix (X that the algorithm of subsequent step, which is matrix X,CTLE, XDFE) the case where.Recognize for the time being
It is X for the continuous time linear equalization parameter of CTLE registerCTLE, DFE register in exchange chip decision feedback equalization
Parameter is XDFEWhen, X at this timeCTLEWith XDFERespectively represent the continuous point value in Fig. 8 on horizontal plane in two reference axis, vertical coordinate axle
Above corresponding is signal quantization parameter Heye(by from the chip read signal eye diagram eye high level and eye width values counted
It calculates, when reading eye high level and eye width values, the continuous time linear equalization parameter of the CTLE register has been set to XCTLE,
The decision feedback equalization parameter of DFE register has been set to XDFE) value.In one embodiment, exchange chip is ready for sending
One group of data, by preemphasis or after postemphasising, signal seems to become more severe, but the variation of this signal just with letter
The filter action in road offsets, and when signal is transferred to the side CDR, signal becomes to be more easier to receive to differentiate level again, and
The characteristic of channel is relatively stable, that is, after the preemphasis for setting exchange chip or parameter of postemphasising, recipient (side CDR) meeting
With metastable eye figure.Therefore it can consider the continuous time linear equalization parameter X of CTLE registerCTLE, DFE register
Decision feedback equalization parameter XDFEWith HeyeIt is one-to-one relationship.
In step S20, the signal quantization parameter of the preset is obtained according to the eye high level of reading and eye width values, and is examined
Surveying the signal quantization parameter of the preset, whether iteration is to extreme value；In one embodiment, scheme to preferably quantify eye
It is bad, the signal quantization parameter of all the points in matrix after being superimposed is calculated according to the following formula；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For eye high level
Weight parameter, θ2For the weight parameter of eye width values.For convenience of calculation, preferably, for the time being by eye high level HEye is highWeight parameter
θ1, eye width values WEye is wideWeight parameter θ2Take 1.Therefore in X0Point Heye(X)Approximate Gaussian distribution (by this present embodiment, the square
The point value of battle array X is not continuous unlimited point, but is set according to register parameters on chip to the point of matrix, because
X is discrete herein for this, therefore in X0Point Heye(X)It is approximate Gaussian Profile).When matrix X is one-dimensional matrix, Fig. 7 is indicated
Its Gaussian Profile approximation situation, and when matrix X is two-dimensional matrix, the case where being referred to Fig. 8, as shown in figure 8, needing to seek
Maximum value H in levelling faceeye, HeyeIntroducing quantified abstract parameter as signal quality.When the matrix X is more higher-dimension
When spending, algorithm is similar, but its probability distribution inconvenience graphically comes out.
It is two-dimensional matrix (X in matrix XCTLE, XDFE) when, if Heye=F (XCTLE,XDFE), setting first needs to find optimal
Two channel parameters be setting be respectively XCTLE, XDFEValue, and the eye height of reading signal eye diagram and eye are wide at this moment, and according toCalculate corresponding signal quantization parameter Heye, and when confirming that it reaches maximum value, recognize
For its iteration to extreme value, and not up to maximum value when, continue iteration and (move the preset along the gradient direction of current point
And rise maximizing), when final current point is moved to extreme point, it can not find the positive gradient (H of current point at this timeeye
Value is maximum), iteration stopping has found optimum parameter value.
In step s 30, in the non-iteration of signal quantization parameter to the extreme value of the preset, according to the eye high level of reading
The gradient of the preset is obtained with eye width values, and mobile described default according to the gradient direction of the preset and preset step-length
Point.And after the movement preset, obtains the signal quantization parameter of the preset after moving and be back to and detect it
Whether iteration is to extreme value.That is, being two-dimensional matrix (X in matrix XCTLE, XDFE) when, if Heye=F (XCTLE,XDFE), F is required first
(XCTLE,XDFE) directional derivative, find maximum directional derivative, i.e. point (XCTLE, XDFE) gradient, and gradient direction be rise
Most fast direction.And it is intelligible, it is not really to use X in this embodimentCTLE, XDFEValue itself goes to carry out the meter of derivative
It calculates, but needs to find optimal channel parameter to be set as above-mentioned value by setting first, and read the eye of eye figure at this moment
It is high and eye is wide goes to be calculated and obtain corresponding signal quantization parameter HeyeAnd at this time the directional derivative of the corresponding point of the parameter and
Gradient direction etc..
It is intelligible, signal quantization parameter H is had verified that in above-mentioned stepseyeWhen non-iteration to extreme value (maximum value), this
When, according to the maximum directional derivative (gradient of the preset) of calculating arrived, the current preset is moved to down
One point.A reasonable numerical value can be taken (in one embodiment, to the preset step-length of next point when mobile
The step-length moved every time is after setting register parameters, to be mapped to signal quantization parameter HeyeCalculating in), it is too small so that
The number of iterations increases, and making the parameter found greatly not very much is optimal (restricted sometimes by chip, stepping length cannot be unlimited
It is small).The above process is then proceeded to, gradient direction a little is found out, it is lasting mobile.
In step s 40, in signal quantization parameter iteration to the extreme value of the preset, confirm corresponding with the extreme value
The matrix X point value be optimum parameter value.That is, in above-mentioned steps S30, when current point is moved to extreme point
It waits, can not find positive gradient at this time, iteration can stop, and algorithm terminates, and have found optimum parameter value (optimum channel parameter).Edge
The maximum direction of gradient look for optimal parameter, can greatly simplify, and shorten iteration number.And then shortens to find and be consumed
The time taken.Compared to the method traversed entirely, which can have target to approach optimum value with having direction.In the present embodiment, Heye
Think that signal quality is better more greatly, HeyeCorresponding X when maximumCTLEWith XDFENumerical value, it is believed that be optimal channel parameter values, this
Shi Ye represents the signal quantization parameter iteration of the preset to extreme value.
In the above-described embodiments, the present invention looks for optimum channel parameter along the maximum direction of gradient, can there is target
Approach optimum channel parameter value with having direction；Meanwhile the number of iteration can be greatly simplified and shorten, and then shorten and find institute
It the time of consuming, greatly improves the efficiency, has saved research and development of products cost.And compared to a dimensional parameter is fixed, only change
The method of one dimensional parameter, the present invention substantially increase the accuracy for finding extreme value.
In one embodiment, the step S10 includes:
The register parameters of the chip are set as to the point value of the preset of the matrix X, and enable the preset folded
Add unit direction matrix, and reads from the chip after superposition the corresponding eye high level of all the points and eye width values in matrix.That is,
Assuming that the matrix X is two-dimensional matrix (XCTLE, XDFE), the preset is (Xi, Xj), wherein the XiFor in exchange chip
CTLE register continuous time linear equalization parameter XCTLECorresponding to the point value of the preset, the XjFor the exchange
The decision feedback equalization parameter X of DFE register in chipDFEPoint value corresponding to the preset；At this point, enabling described default
Point (Xi, Xj) superposition unit direction matrix, and the corresponding H of all the points in superimposed matrix is read from the chipEye is highValue and
WEye is wideValue.
In one embodiment, the step S20 includes:
The signal quantization parameter of all the points in matrix after being superimposed is calculated according to the following formula；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For eye high level
Weight parameter, θ2For the weight parameter of eye width values.Namely, it is assumed that the matrix X is two-dimensional matrix (XCTLE, XDFE), it is described pre-
It sets up an office as (Xi, Xj), it in this step, can be according to the H of readingEye is highValue and WEye is wideValue and following formula obtain the preset
(Xi, Xj) the signal quantization parameter of other points along matrix after superposition:
The signal quantization parameter for comparing all the points in matrix after being superimposed, the signal quantization parameter in the preset are greater than it
When the signal quantization parameter of his all the points, confirm the signal quantization parameter iteration of the preset to extreme value.In the present embodiment,
HeyeThink that signal quality is better more greatly, HeyeCorresponding X when maximumCTLEWith XDFENumerical value, it is believed that be optimal channel parameter
Value, also represents the signal quantization parameter iteration of the preset to extreme value at this time, otherwise non-iteration to extreme value.
In one embodiment, the step S30 includes:
In the non-iteration of signal quantization parameter to the extreme value of the preset, the preset (X is obtainedi, Xj) be superimposed after
The matrix of differences Δ H for the signal quantization parameter that other in matrix are puteye=Heye-Heye2,2·E3*3。
It is led according to the matrix of differences and point away from preset described in the matrix after acquisition superposition and other directions put
Numerical value, and confirm that the maximum value of the directional derivative is the gradient of the preset；Preferably, calculating the directional derivative value
Formula are as follows:Wherein, the LPoint away fromRepresent the preset and other point between point away from,The side of representative
When to derivative value.Preset (X described in the matrix after being superimposed is obtained at this timei, Xj) with other point directional derivative value, and really
Recognize the maximum value of the directional derivativeFor the gradient of the preset.
According to the gradient direction of the preset by the preset (Xi, Xj) movement preset step-length.It is intelligible, one
In a little embodiments, after mobile preset step-length, the signal quantization ginseng that step S20 obtains the preset after movement can be back to
It counts and detects it whether iteration is to extreme value.
Illustrate rising in the above embodiment of the present invention with gradient by taking the chip DS125DF1610 (CDR) of TI as an example below
The method that method quickly finds channel parameter:
In this embodiment, the CDR in switching equipment is the DS125DF1610 of TI.Switching equipment has external
Console debugging interface can read and write the internal register of DS125DF1610, setting, eye including preemphasis and balance parameters
The reading of the eye high level and eye width values of figure checks that sending and receiving data is received and no has CRC.
In the present embodiment, it needs to be arranged two registers: CTLE register (8bit), DFE register (5bit), uses
(XCTLE, XDFE) indicate, construct a two-dimensional surface.If (XCTLE, XDFE) it is continuous point, then, F (XCTLE, XDFE) should be
It can micro- guidable function.Reality is due to (XCTLE, XDFE) be discrete point, therefore the regulating power of chip be also impossible to it is infinitely small.
Although its actual directional derivative cannot be sought, some H with surrounding point approximate can be found outeyeThe difference of value/away from, and with
This maximum directional derivative (gradient direction put) to estimate the point.I.e. the point direction (0,1), (1,1), (1,0),
(1, -1), (0, -1), (- 1, -1), (- 1,0), (- 1,1) amount to 8 reversely upper maximum directional derivatives.
Take point (point value of preset) since median (at this time by the continuous of instruction setting CTLE register for the first time
Linearly balance parameters XCTLEFor the decision feedback equalization parameter X of 128, DFE registerDFE16) namely preset is for
(128,16) (CTLE register 8bit has 256 parameters, and DFE register has 5bit, that is, has 32 parameters, preset
Point value takes the median of the two).
Therefore the position of current point is (128,16).The position of current point is superimposed unit direction matrix, obtains eight sides
To the position of point.That is, as shown in figure 9, after superposition in matrix, eight points in addition to current point be respectively (128,17),
(129,17)、(129,16)、(129,15)、(128,15)、(127,15)、(127,16)、(127,17)。
At this point, sending instruction setting channel parameter (the continuous time linear equalization parameter X of CTLE register respectivelyCTLE,
The decision feedback equalization parameter X of DFE registerDFE) be this eight points point value, after chip operation stablize after, read channel parameter
Respectively each point value when signal eye diagram eye high level HEye is highWith eye width values WEye is wide(function of chip, can be into from the chip
Row is read).In order to eliminate interference, multiple high level H can be continuously read when setting each group of channel parameterEye is highWith eye width values
WEye is wide, and ask its mean value as the eye high level H finally obtainedEye is highWith eye width values WEye is wide.According to the following formula:
Find out 8 points of corresponding H in the position of current point and surrounding in 3*3 matrixeyeValue.The position H of current pointeyeWith it is each around
The difference matrix Δ H of pointeyeIt indicates.So Δ Heye=Heye-Heye2,2·E3*3
The approximate directional derivative of 8 points of surrounding:Find the maximum value of directional derivative
It is exemplified below and how to calculate, 3*3 matrix:
Above data really passes through actual measurement, and is calculated again.LPoint away fromIt is actually the company for thinking CTLE register
Continuous linearly balance parameters XCTLEWith the decision feedback equalization parameter X of DFE registerDFEThe stepping length of the two parameters is 1
It (shows as when generation is mobile, the stepping length that two parameters are modified every time is 1, and the length is mapped to actual HeyeMeter
When in calculation, it may occur that change), can also be according to each chip feature, (stepping of channel parameter is long for the step-length between setting each point
Degree) it is not all 1, practical calculate can be more accurate in this way.LPoint away fromIt can be understood as a dimensionless group, be intended merely to preferably
Quantum chemical method and introduce.For convenience of calculation, think that the stepping length between various discrete point is 1 in citing.Such as Figure 11 institute
Show, above-mentioned calculating finally obtainsAnd its gradient direction is (0, -1).
The position for updating current point, is moved to next point position, is moved to new point by the former preset (128,16)
(128,15).The method for calculating the value newly put are as follows: origin is added according to the direction vector being calculated in above-mentioned steps
(0, -1) obtains newly putting coordinate.It should be noted that in above-mentioned 3*3 matrix, although 6.3 be obtained maximum HeyeBut if
Position (129,15), X are moved to by position (128,16)CTLEWith XDFE1 is both increased, since the two is considered as orthogonality relation,
So the distance between two o'clock R=(12+12) 1/2=1.414.Farther out relative to distance for point (128,15), therefore, 6.3
Although for obtained maximum Heye, but by formulaEstimate that obtained directional derivative is less than a little instead
(128,15).Therefore the moving direction of next point is to (0, -1) namely point (128,15).But in practice, various discrete parameter is
Increase by 1, but the variation of preemphasis or equalizer curve is not the relationship of a linear change.Therefore in order to allow the number of iterations it is less,
Optimum point is quickly found, the distance of various discrete parameter point can be adjusted accordingly according to demand.
As shown in Figure 10, current location point HeyeWhen value is not up to maximum value, the movement that repeats the above steps is described pre-
Set up an office (direction shown in arrow in Figure 10), finally makes the H of current pointeyeWhen value is maximum value, stop iteration.So far it has sought
Find the optimum value of the IA High Speed Channel parameter of the equipment veneer.
The present invention most begins setting up mathematical model, in order to which simplification is easy to understand, therefore, it is considered that HeyeProbability distribution entire
Only one extreme point in space, and be maximum value.But in practice, HeyeIt should be that there are extreme points in subrange, such as scheme
Shown in 12, CTLE shown in figure represents the continuous time linear equalization parameter X of CTLE registerCTLE, DFE generation shown in figure
The decision feedback equalization parameter X of table DFE registerDFE,There are multiple extreme points in this space, and attached in each extreme point
Closely meet the probability distribution of foregoing description.Therefore, time extreme point, the point are only found on a veneer as soon as not ensuring that
For maximum value.Intelligible, in the present invention, same type product is generally also looked for not in order to which the channel parameter searched out is reliable
(such as 10) equipment of same preset quantity, while finding the channel parameter.It is estimated from probability to obtain several best ginsengs
Several points.Again in the next steps by the point of obtained optimal parameter, according to the step in above-described embodiment in remaining equipment
Verify whether it is also available parameter value in the extreme point or certain area in region.The parameter point verified again changes
Generation number will be substantially less that random snap shot finds extreme point.In order to avoid unlucky situation occurs, multiple devices all be can not find
Maximum value.So should allow multiple devices initial point random distribution planar, allow its find path be varied.It is appreciated that
, described above, only the present invention is better illustrates algorithm, and illustrates the process calculated, and non-limiting of the invention
Protection scope.The realization of the algorithm does not limit any programming language.
The above embodiment of the present invention is utilized the probability Estimation to actual signal parameter, after mathematical modeling, using in gradient
The algorithm risen, iteratively faster simultaneously search out optimum channel parameter.Method of the invention is compared to traditional whole parameters of traversal
Method greatly improves the efficiency, and has saved research and development of products cost.Relative to a dimensional parameter is fixed, only change a dimension
The method of parameter substantially increases the accuracy for finding extreme value.The present invention looks for optimum channel along the maximum direction of gradient
Parameter can have target to approach optimum channel parameter value with having direction；Meanwhile the number of iteration can be greatly simplified and shorten,
And then shorten and find the spent time, it greatly improves the efficiency, has saved research and development of products cost.And compared to fixing one
Dimensional parameter, only changes the method for a dimensional parameter, and the present invention substantially increases the accuracy for finding extreme value.
As shown in figure 3, in one embodiment, providing a kind of channel parameter device for searching, comprising:
Module 31 is obtained, the point value of the preset for the register parameters of chip to be set as to pre-emphasis parameters is read
The eye high level and eye width values of current demand signal eye figure；The pre-emphasis parameters are matrix；
Detection module 32 obtains the signal quantization parameter of the preset for the eye high level and eye width values according to reading,
And it detects the signal quantization parameter of the preset whether iteration is to extreme value；
Mobile module 33, in the non-iteration of signal quantization parameter to the extreme value of the preset, according to the eye of reading
High level and eye width values obtain the gradient of the preset, and according to the gradient direction of the preset and preset step-length movement
Preset；
Confirmation module 34, for confirming and the extreme value in signal quantization parameter iteration to the extreme value of the preset
The point value of the corresponding matrix is optimum parameter value.
As shown in figure 4, in one embodiment, the acquisition module 31 includes:
Reading submodule 311, the point value of the preset for the register parameters of the chip to be set as to the matrix,
And the preset superposition unit direction matrix is enabled, and the corresponding eye of all the points is high in matrix after reading superposition from the chip
Value and eye width values.
As shown in figure 5, in one embodiment, the detection module 32 includes:
First acquisition submodule 321, for calculating the signal quantization ginseng of all the points in matrix after superposition according to the following formula
Number；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For eye high level
Weight parameter, θ2For the weight parameter of eye width values；
First confirmation submodule 322 is preset for compare the signal quantization parameter of all the points in matrix after superposition described
When the signal quantization parameter of point is greater than the signal quantization parameter of other all the points, confirm that the signal quantization parameter of the preset changes
In generation, is to extreme value.
As shown in fig. 6, in one embodiment, the mobile module 33 includes:
Second acquisition submodule 331, for obtaining institute in the non-iteration of signal quantization parameter to the extreme value of the preset
State a little be superimposed after in matrix the signal quantization parameters of other points matrix of differences Δ Heye；
Second confirmation submodule 332 is used for according to the matrix of differences and point away from after acquisition superposition described in the matrix
The directional derivative value of preset and other points, and confirm that the maximum value of the directional derivative is the gradient of the preset；
Preset step-length is moved for the gradient direction according to the preset in the preset by mobile submodule 333.
The channel parameter device for searching of above-described embodiment looks for optimum channel parameter along the maximum direction of gradient, can
To there is target to approach optimum channel parameter value with having direction；Meanwhile the number of iteration can be greatly simplified and shorten, and then shorten
The spent time is found, greatly improves the efficiency, has saved research and development of products cost.And join compared to a dimension is fixed
Number, only changes the method for a dimensional parameter, and the present invention substantially increases the accuracy for finding extreme value.
In one embodiment, a kind of computer readable storage medium is also provided, on the computer readable storage medium
Computer executable instructions are stored with, when the computer executable instructions are executed by processor, so that the processor executes
The register parameters of chip: being set as the point value of the preset of pre-emphasis parameters by following steps, read current demand signal eye figure
Eye high level and eye width values；The pre-emphasis parameters are matrix；The preset is obtained according to the eye high level of reading and eye width values
Signal quantization parameter, and detect the signal quantization parameter of the preset whether iteration is to extreme value；Signal in the preset
When the non-iteration of quantization parameter to extreme value, the gradient of the preset is obtained according to the eye high level of reading and eye width values, and according to institute
State gradient direction and the mobile preset of preset step-length of preset；The preset signal quantization parameter iteration extremely
When value, confirm that the point value of the matrix corresponding with the extreme value is optimum parameter value.
In one embodiment, when computer executable instructions are executed by processor, processor is also made to execute following step
It is rapid: the register parameters of the chip being set as to the point value of the preset of the matrix, and enable the preset superposition unit
Direction matrix, and read from the chip after superposition the corresponding eye high level of all the points and eye width values in matrix.
In one embodiment, when computer executable instructions are executed by processor, processor is also made to execute following step
It is rapid: to calculate the signal quantization parameter of all the points in matrix after being superimposed according to the following formula；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For eye high level
Weight parameter, θ2For the weight parameter of eye width values；
The signal quantization parameter for comparing all the points in matrix after being superimposed, the signal quantization parameter in the preset are greater than it
When the signal quantization parameter of his all the points, confirm the signal quantization parameter iteration of the preset to extreme value.
In one embodiment, when computer executable instructions are executed by processor, processor is also made to execute following step
It is rapid: in the non-iteration of signal quantization parameter to the extreme value of the preset, obtain the point be superimposed after other points in matrix
The matrix of differences of signal quantization parameter；According to the matrix of differences and point away from preset described in the matrix after acquisition superposition
The directional derivative value put with other, and confirm that the maximum value of the directional derivative is the gradient of the preset；According to described pre-
The gradient direction to set up an office is by the mobile preset step-length in the preset, and the signal quantization parameter of the preset after obtaining movement is simultaneously
Detecting it, whether iteration is to extreme value.
Above-mentioned computer readable storage medium looks for optimum channel parameter along the maximum direction of gradient, can there is mesh
Approach optimum channel parameter value with indicating direction；Meanwhile the number of iteration can be greatly simplified and shorten, and then shorten searching
It the spent time, greatly improves the efficiency, has saved research and development of products cost.And compared to a dimensional parameter is fixed, only become
Change the method for a dimensional parameter, the present invention substantially increases the accuracy for finding extreme value.
In one embodiment, a kind of computer equipment, including memory and processor are also provided, stored up in the memory
Have computer-readable instruction, when the computer-readable instruction is executed by the processor so that the processor execute with
The register parameters of chip: being set as the point value of the preset of pre-emphasis parameters by lower step, reads the eye of current demand signal eye figure
High level and eye width values；The pre-emphasis parameters are matrix；The letter of the preset is obtained according to the eye high level of reading and eye width values
Number quantization parameter, and detect the signal quantization parameter of the preset whether iteration is to extreme value；Semaphore in the preset
When changing the non-iteration of parameter to extreme value, the gradient of the preset is obtained according to the eye high level of reading and eye width values, and according to described
The mobile preset of the gradient direction and preset step-length of preset；The preset signal quantization parameter iteration to extreme value
When, confirm that the point value of the matrix corresponding with the extreme value is optimum parameter value.
In one embodiment, when computer-readable instruction is executed by processor, processor is also made to execute following steps:
The register parameters of the chip are set as to the point value of the preset of the matrix, and enable preset superposition unit direction
Matrix, and read from the chip after superposition the corresponding eye high level of all the points and eye width values in matrix.
In one embodiment, when computer-readable instruction is executed by processor, processor is also made to execute following steps:
The signal quantization parameter of all the points in matrix after being superimposed is calculated according to the following formula；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For eye high level
Weight parameter, θ2For the weight parameter of eye width values；
The signal quantization parameter for comparing all the points in matrix after being superimposed, the signal quantization parameter in the preset are greater than it
When the signal quantization parameter of his all the points, confirm the signal quantization parameter iteration of the preset to extreme value.
In one embodiment, when computer-readable instruction is executed by processor, processor is also made to execute following steps:
In the non-iteration of signal quantization parameter to the extreme value of the preset, obtain the point be superimposed after in matrix other points signal
The matrix of differences of quantization parameter；According to the matrix of differences and point away from obtaining after superposition preset described in the matrix and its
The directional derivative value that he orders, and confirm that the maximum value of the directional derivative is the gradient of the preset；According to the preset
Gradient direction by the mobile preset step-length in the preset, obtain it is mobile after the signal quantization parameter of the preset simultaneously detect
Whether iteration is to extreme value for it.
Above-mentioned computer equipment looks for optimum channel parameter along the maximum direction of gradient, can have target to have direction
Ground approaches optimum channel parameter value；Meanwhile the number of iteration can be greatly simplified and shorten, and then shorten spent by searching
Time greatly improves the efficiency, and has saved research and development of products cost.And compared to a dimensional parameter is fixed, only tie up for variation one
The method of parameter is spent, the present invention substantially increases the accuracy for finding extreme value.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, described program can be stored in a computer-readable storage medium
In, in the embodiment of the present invention, which be can be stored in the storage medium of computer system, and by the computer system
At least one processor executes, and includes the process such as the embodiment of above-mentioned each method with realization.Wherein, the storage medium can be
Magnetic disk, CD, read-only memory (Read-OnlX MemorX, ROM) or random access memory (Random Access
MemorX, RAM) etc..
Each technical characteristic of embodiment described above can be combined arbitrarily, for simplicity of description, not to above-mentioned reality
It applies all possible combination of each technical characteristic in example to be all described, as long as however, the combination of these technical characteristics is not deposited
In contradiction, all should be considered as described in this specification.
The embodiments described above only express several embodiments of the present invention, and the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to protection of the invention
Range.Therefore, the scope of protection of the patent of the invention shall be subject to the appended claims.
Claims (10)
1. a kind of channel parameter finding method, which is characterized in that the described method includes:
The register parameters of chip are set as to the point value of the preset of pre-emphasis parameters, read the eye high level of current demand signal eye figure
With eye width values；The pre-emphasis parameters are matrix；
The signal quantization parameter of the preset is obtained according to the eye high level of reading and eye width values, and detects the letter of the preset
Whether iteration is to extreme value for number quantization parameter；
In the non-iteration of signal quantization parameter to the extreme value of the preset, according to the eye high level of reading and the acquisition of eye width values
The gradient of preset, and according to the gradient direction of the preset and the mobile preset of preset step-length；
In signal quantization parameter iteration to the extreme value of the preset, the point value of the matrix corresponding with the extreme value is confirmed
For optimum parameter value.
2. the method according to claim 1, wherein described be set as preemphasis ginseng for the register parameters of chip
The point value of several presets reads the eye high level and eye width values of current demand signal eye figure, comprising:
The register parameters of the chip are set as to the point value of the preset of the matrix, and enable the preset superposition unit
Direction matrix, and read from the chip after superposition the corresponding eye high level of all the points and eye width values in matrix.
3. according to the method described in claim 2, it is characterized in that, described according to the eye high level of reading and the acquisition of eye width values
The signal quantization parameter of preset, and detect the signal quantization parameter of the preset whether iteration is to extreme value, comprising:
The signal quantization parameter of all the points in matrix after being superimposed is calculated according to the following formula；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For the power of eye high level
Weight parameter, θ2For the weight parameter of eye width values；
The signal quantization parameter for comparing all the points in matrix after being superimposed, the signal quantization parameter in the preset are greater than other institutes
When signal quantization parameter a little, confirm the signal quantization parameter iteration of the preset to extreme value.
4. according to the method described in claim 3, it is characterized in that, the non-iteration of signal quantization parameter in the preset
When to extreme value, the gradient of the preset is obtained according to the eye high level of reading and eye width values, and according to the gradient of the preset
Direction and the mobile preset of preset step-length, comprising:
In the non-iteration of signal quantization parameter to the extreme value of the preset, obtain the point be superimposed after other points in matrix
The matrix of differences of signal quantization parameter；
According to the matrix of differences and point away from the directional derivative of preset and other points described in the matrix after acquisition superposition
Value, and confirm that the maximum value of the directional derivative is the gradient of the preset；
Preset step-length is moved into the preset according to the gradient direction of the preset.
5. a kind of channel parameter device for searching characterized by comprising
Module is obtained, the point value of the preset for the register parameters of chip to be set as to pre-emphasis parameters reads current letter
The eye high level and eye width values of number eye figure；The pre-emphasis parameters are matrix；
Detection module, the signal quantization parameter of the preset is obtained for the eye high level and eye width values according to reading, and is detected
Whether iteration is to extreme value for the signal quantization parameter of the preset；
Mobile module, in the non-iteration of signal quantization parameter to the extreme value of the preset, according to the eye high level of reading and
Eye width values obtain the gradient of the preset, and mobile described default according to the gradient direction of the preset and preset step-length
Point；
Confirmation module, for confirming corresponding with the extreme value in signal quantization parameter iteration to the extreme value of the preset
The point value of the matrix is optimum parameter value.
6. device according to claim 5, which is characterized in that the acquisition module includes:
Reading submodule, the point value of the preset for the register parameters of the chip to be set as to the matrix, and enable institute
Preset superposition unit direction matrix is stated, and reads from the chip after superposition the corresponding eye high level of all the points and eye in matrix
Width values.
7. device according to claim 6, which is characterized in that the detection module includes:
First acquisition submodule, for calculating the signal quantization parameter of all the points in matrix after superposition according to the following formula；
Wherein, HeyeFor signal quantization parameter, HEye is highFor the eye high level of reading, WEye is wideFor the eye width values of reading, θ1For the power of eye high level
Weight parameter, θ2For the weight parameter of eye width values；
First confirmation submodule, for compare be superimposed after in matrix all the points signal quantization parameter, letter in the preset
When number quantization parameter is greater than the signal quantization parameter of other all the points, the signal quantization parameter iteration of the preset is confirmed extremely
Value.
8. device according to claim 7, which is characterized in that the mobile module includes:
Second acquisition submodule, in the non-iteration of signal quantization parameter to the extreme value of the preset, obtain the point with
The matrix of differences for the signal quantization parameter that other in matrix are put after superposition；
Second confirmation submodule is used for according to the matrix of differences and point away from preset described in the matrix after acquisition superposition
The directional derivative value put with other, and confirm that the maximum value of the directional derivative is the gradient of the preset；
Preset step-length is moved for the gradient direction according to the preset in the preset by mobile submodule.
9. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Executable instruction, when the computer executable instructions are executed by processor, so that the processor perform claim requires 1 to 4
Any one of the method the step of.
10. a kind of computer equipment, which is characterized in that including memory and processor, store computer in the memory
Readable instruction, when the computer-readable instruction is executed by the processor, so that the processor perform claim requires 1 to 4
Any one of the method the step of.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710625321.4A CN109308360A (en) | 2017-07-27 | 2017-07-27 | Channel parameter finding method, device, storage medium and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710625321.4A CN109308360A (en) | 2017-07-27 | 2017-07-27 | Channel parameter finding method, device, storage medium and computer equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109308360A true CN109308360A (en) | 2019-02-05 |
Family
ID=65202295
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710625321.4A Pending CN109308360A (en) | 2017-07-27 | 2017-07-27 | Channel parameter finding method, device, storage medium and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109308360A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111352504A (en) * | 2019-12-25 | 2020-06-30 | 晶晨半导体(上海)股份有限公司 | Memory interface timing sequence analysis method and system |
| CN113779744A (en) * | 2020-06-10 | 2021-12-10 | 英业达科技有限公司 | Method for determining continuous time linear equalizer settings |
| CN115150303A (en) * | 2022-07-29 | 2022-10-04 | 苏州浪潮智能科技有限公司 | Switch port testing method, system, equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1605274A1 (en) * | 2004-06-08 | 2005-12-14 | Bruker BioSpin MRI GmbH | Automatic k-space trajectory calibration for MRI |
| US20060203899A1 (en) * | 2005-02-03 | 2006-09-14 | Gee David M | Method of equalising a channel and apparatus therefor |
| US20080007854A1 (en) * | 2006-07-10 | 2008-01-10 | Broadcom Corporation, A California Corporation | Guided target search for hard disk drive applications |
| CN101771637A (en) * | 2010-01-08 | 2010-07-07 | 南京邮电大学 | Non-Gauss noise-against blind equalization method |
| CN102298779A (en) * | 2011-08-16 | 2011-12-28 | 淮安盈科伟力科技有限公司 | Image registering method for panoramic assisted parking system |
| CN103885911A (en) * | 2012-12-20 | 2014-06-25 | 辉达公司 | Multipass approach for performing channel equalization training |
| CN103885907A (en) * | 2012-12-20 | 2014-06-25 | 辉达公司 | Equalization Coefficient Search Algorithm |
| CN104281553A (en) * | 2013-07-03 | 2015-01-14 | 辉达公司 | Adaptation of crossing DFE tap weight |
| CN106155950A (en) * | 2015-03-23 | 2016-11-23 | 中兴通讯股份有限公司 | Parameter processing method and device |
-
2017
- 2017-07-27 CN CN201710625321.4A patent/CN109308360A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1605274A1 (en) * | 2004-06-08 | 2005-12-14 | Bruker BioSpin MRI GmbH | Automatic k-space trajectory calibration for MRI |
| US20060203899A1 (en) * | 2005-02-03 | 2006-09-14 | Gee David M | Method of equalising a channel and apparatus therefor |
| US20080007854A1 (en) * | 2006-07-10 | 2008-01-10 | Broadcom Corporation, A California Corporation | Guided target search for hard disk drive applications |
| CN101771637A (en) * | 2010-01-08 | 2010-07-07 | 南京邮电大学 | Non-Gauss noise-against blind equalization method |
| CN102298779A (en) * | 2011-08-16 | 2011-12-28 | 淮安盈科伟力科技有限公司 | Image registering method for panoramic assisted parking system |
| CN103885911A (en) * | 2012-12-20 | 2014-06-25 | 辉达公司 | Multipass approach for performing channel equalization training |
| CN103885907A (en) * | 2012-12-20 | 2014-06-25 | 辉达公司 | Equalization Coefficient Search Algorithm |
| CN104281553A (en) * | 2013-07-03 | 2015-01-14 | 辉达公司 | Adaptation of crossing DFE tap weight |
| CN106155950A (en) * | 2015-03-23 | 2016-11-23 | 中兴通讯股份有限公司 | Parameter processing method and device |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111352504A (en) * | 2019-12-25 | 2020-06-30 | 晶晨半导体(上海)股份有限公司 | Memory interface timing sequence analysis method and system |
| CN111352504B (en) * | 2019-12-25 | 2021-11-26 | 晶晨半导体(上海)股份有限公司 | Memory interface timing sequence analysis method and system |
| CN113779744A (en) * | 2020-06-10 | 2021-12-10 | 英业达科技有限公司 | Method for determining continuous time linear equalizer settings |
| CN113779744B (en) * | 2020-06-10 | 2023-07-18 | 英业达科技有限公司 | Method for determining continuous time linear equalizer settings |
| CN115150303A (en) * | 2022-07-29 | 2022-10-04 | 苏州浪潮智能科技有限公司 | Switch port testing method, system, equipment and storage medium |
| CN115150303B (en) * | 2022-07-29 | 2023-08-08 | 苏州浪潮智能科技有限公司 | A switch port testing method, system, device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111597753B (en) | Data depth change characteristic self-adaptive two-dimensional resistivity inversion method and system | |
| CN107490800B (en) | A satellite navigation fast positioning method, device and satellite navigation receiver | |
| WO2019085709A1 (en) | Pooling method and system applied to convolutional neural network | |
| CN109308360A (en) | Channel parameter finding method, device, storage medium and computer equipment | |
| CN109035363B (en) | Line-circle optimal fitting method for rapid iteration | |
| CN103885876B (en) | Method of testing and equipment | |
| CN106300418B (en) | Photovoltaic DC-to-AC converter control parameter discrimination method based on adaptive differential evolution algorithm | |
| CN113759311B (en) | A positioning method, device and storage medium | |
| WO2024140593A1 (en) | Information processing method and apparatus, signal positioning method and apparatus, device and medium | |
| CN106909539A (en) | Image indexing system, server, database and related methods | |
| CN105763269A (en) | Method for calibrating antenna and calibration signal processing device and system | |
| CN111122222B (en) | A method and system for determining the location of a sample point | |
| CN113392857B (en) | Target detection method, device and equipment terminal based on yolo network | |
| EP2991234A1 (en) | Method, apparatus, and system for configuring high-speed serial bus parameter | |
| CN112737713A (en) | PCIe link equalization coefficient automatic adjustment method and device | |
| CN110286275A (en) | A method and device for determining system stability when equipment is connected to the grid | |
| CN104462788B (en) | A kind of correction method for finite element model based on reverse minor structure | |
| CN104714964B (en) | A kind of physiological data Outliers Detection method and device | |
| TW201818703A (en) | Modulation mode detection method and device | |
| CN104182407A (en) | Data processing system for narrowing data search range | |
| CN104778149B (en) | The computational methods of current-carrying capacity inverse problem | |
| CN106452621A (en) | Peak power determining method, peak-to-average power ratio determining method, peak power determining device, and peak-to-average power ratio determining device | |
| CN113312768B (en) | Optimization design method for electromagnet multi-quality characteristic parameters based on Taguchi and BBD | |
| CN115118296A (en) | A real-time error compensation method, system, terminal and medium for a multi-channel receiver | |
| CN107045410A (en) | A kind of method and system for infrared touch panel high precision computation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190205 |
|
| WD01 | Invention patent application deemed withdrawn after publication |