JP2004199290A - Noise error decision method and noise error decision program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reflect conditions specific to a cell in an actual semiconductor integrated circuit as faithfully as possible, to more accurately estimate noise resistance of each the cell, and to decide a noise error under a necessary and sufficient standard. <P>SOLUTION: Previously, an output noise amount of the cell is extracted in consideration of output load dependence or shape of generated noise in each the cell and is stored in a library as a noise propagation parameter, and a voltage change permissible amount is stored in the library in each the cell. Based on circuit information, a damaged cell suffering the noise and a succeeding cell connected to an output terminal thereof are recognized, and the noise propagation parameter of the damaged cell and the voltage change permissible amount of the succeeding cell are read from the library. On the condition that the voltage change permissible amount and the noise amount outputted from the damaged cell according to an output load or the shape of the noise generated in input of the damaged cell coincide, a noise resistance parameter is decided, and is compared to a noise analysis result to decide the error. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a noise error determination method and a noise error determination program for analyzing noise generated on signal wiring of a semiconductor integrated circuit and determining whether the noise obtained as a result of the analysis is within an allowable range.
[0002]
In recent years, in a semiconductor integrated circuit, an excessive parasitic capacitance exists between adjacent wirings due to miniaturization of a wiring process, thereby increasing the influence of noise generated on signal wiring on fluctuations in circuit operation and delay. I have. Therefore, when designing a semiconductor integrated circuit, it is necessary to analyze crosstalk noise and determine a noise error, and modify the circuit layout based on the determination result. In particular, the determination of the noise error is a step of determining whether or not the generated noise is of an acceptable level, and is extremely important for determining whether or not the layout needs to be corrected. The presence / absence of the noise error is determined based on the magnitude relationship between the noise resistance and the allowable value of the cell receiving the noise and the noise analysis result for the cell.
[0003]
[Prior art]
FIG. 15 is a flowchart showing a processing procedure of a conventional noise error determination method. As shown in FIG. 15, conventionally, noise immunity of each cell is previously extracted as a cell characteristic and prepared as a noise immunity parameter (step S151). Then, a noise analysis is performed (step S152). The noise error is determined by comparing the noise analysis result with the noise immunity parameter (step S153). If the result of the determination is that there is a problem (step S153: NG), an error is output (step S154). If there is no problem (Step S153: OK), the error determination (Check) is completed (Step S155).
[0004]
FIG. 16 is a conceptual diagram illustrating a conventional noise error determination method. In FIG. 16, reference numeral 1 denotes a damaged wiring affected by noise from the attack wiring 2, and reference numerals 3 to 6 represent cells. In the figure in the circle 7 at the lower left of FIG. 16, as a result of the noise analysis, noise 8 (hereinafter referred to as positive noise) 8 having a peak at the higher potential is applied to the damaged wiring 1. Is represented. The figure in the circle 9 at the lower right of FIG. 16 shows the noise immunity parameters of the cell 6 connected to the damaged wiring 1, which are composed of the high-side and low-side noise margins with respect to the voltage amplitude.
[0005]
2. Description of the Related Art Conventionally, as noise immunity parameters, those expressed using a noise voltage value (hereinafter, referred to as Vp) and a noise width (hereinafter, referred to as Tw) and those expressing only a noise voltage value (Vp) are known. (For example, see Non-Patent Documents 1 and 2).
[0006]
[Non-patent document 1]
Ashok Vittal and one other, "(Crosstalk Reduction for VLSI)", I Triple E Transactions on Computer Aided Design of Integrated Circuits and Systems (IEEE TRANSONSACT) COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS), (USA), March 1997, Vol. 16, No. 3, pp. 290-298.
[Non-patent document 2]
"Timing Library Format Reference," Cadence Design Systems, Inc., (USA), 4.3 Edition, October 2000, p. 85-86
[0007]
[Problems to be solved by the invention]
In the conventional noise error determination method, a temporary determination level and a temporary output load are assumed for each cell, and a noise resistance parameter is obtained using a temporary noise waveform. However, in practice, the determination level of the output of a cell that has received noise differs depending on the noise allowance and input state of the cell connected to the output unit of the cell, and the output waveform of the cell that has received noise is: It differs depending on the load condition of the output section of the cell and how the noise waveform rises and falls.
[0008]
Therefore, there is a problem that the noise immunity parameter of each cell does not accurately reflect the noise immunity of the actual circuit, and a problem that the output waveform of the cell receiving the noise is not accurate. Further, conventionally, it is necessary to set a criterion for error determination more strictly than necessary, and there is a problem that unnecessary layout correction and unnecessary re-design are caused.
[0009]
The present invention has been made in view of the above-described problems, and reflects a unique condition of a cell in an actual circuit as accurately as possible to more accurately estimate the noise immunity of each cell, and provides a necessary and sufficient standard. It is another object of the present invention to provide a noise error determination method and a noise error determination program capable of determining a noise error.
[0010]
[Means for Solving the Problems]
To achieve the above object, the present invention is characterized by the following. That is, for each cell, the amount of output noise of the cell is extracted in advance in consideration of the shape of the generated noise and output load dependency, and stored in a library as a noise propagation parameter. In addition, the voltage change allowable amount is stored in the library for each cell in advance. At this time, the voltage change allowable amount is prepared for each input terminal of the cell, and if it differs depending on circuit conditions, it is also prepared for each condition.
[0011]
Then, based on the circuit information, the damaged cell receiving the noise and the subsequent cell connected to the output terminal are recognized, and the noise propagation parameter of the damaged cell and the allowable voltage change of the subsequent cell are read from the library. The noise tolerance parameter is determined under the condition that the read allowable voltage change amount and the noise amount output from the damaged cell according to the shape and output load of the noise generated at the input of the damaged cell. At this time, the noise resistance parameter can be determined in consideration of the noise waveform shape, the output load of the damaged cell, and the like. Then, an error is determined by comparing the noise analysis result with the noise resistance parameter.
[0012]
According to the present invention, it is possible to determine the noise error under the condition unique to the cell while considering the influence on the subsequent cell and at the same time, considering the shape of the generated noise and the output load condition. By reflecting the cell-specific conditions in the circuit as faithfully as possible, the noise tolerance of each cell can be more accurately estimated, and the noise error can be determined based on necessary and sufficient criteria.
[0013]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a noise error determination method and a noise error determination program according to an embodiment of the present invention will be described in detail with reference to the drawings.
[0014]
(Embodiment 1)
FIG. 1 is a flowchart illustrating a processing procedure of the noise error determination method according to the first embodiment of the present invention, and FIG. 2 is a conceptual diagram illustrating the determination method. As shown in FIG. 1, characteristics are extracted for each cell in advance, and a noise propagation parameter described later is obtained (step S11). Further, a voltage change allowable amount described later is set for each input terminal of each cell (step S12). Then, for example, the noise propagation parameter and the allowable voltage change amount are stored as a library in a storage device of the computer system.
[0015]
Next, based on the circuit information, a damaged cell receiving noise and a subsequent cell connected to the output terminal of the damaged cell are extracted from the circuit connection information (step S13). In the conceptual diagram shown in FIG. 2, a cell connected to the damaged wiring 11 is a damaged cell 16, and a cell connected to an output of the damaged cell 16 is a subsequent cell 21. Then, the noise propagation parameter of the damaged cell 16 and the allowable voltage change amount of the input terminal of the succeeding cell 21 are read from the library, and the allowable voltage change amount is compared with the output voltage value of the noise propagation parameter to obtain the noise of the damaged cell 16. A resistance parameter is generated (Step S14).
[0016]
Also, a noise analysis is performed, and a noise voltage value Vp and a noise width Tw are obtained from the analysis result (step S15). The values of Vp and Tw obtained as a result of the analysis are compared with the noise resistance parameters generated in step S14, and a noise error determination is performed (step S16). If the result of the determination is that there is a problem (step S16: NG), an error is output (step S17). If there is no problem (Step S16: OK), the error determination (Check) is completed (Step S18). Although not particularly limited, Tw is a half width of noise.
[0017]
In FIG. 2, reference numeral 12 denotes an attack wire, and reference numerals 13 to 15 denote cells. The figure in the circle 17 at the lower left of FIG. 2 shows a state where the positive noise 18 is applied to the damaged wiring 11 as a result of the noise analysis. This positive noise 18 is input to the damaged cell 16. The figure in the circle 31 at the lower center of FIG. 2 represents the noise propagation parameters of the damaged cell 16. The figure in the circle 41 at the lower right of FIG. 2 indicates the allowable voltage change amount of the input terminal of the succeeding cell 21. The figure inside the ellipse 51 in the upper right of FIG. 2 represents a noise resistance parameter unique to the damaged cell 16.
[0018]
Dvp is a voltage value of the positive noise 18. Dvop is the peak voltage of the output waveform with respect to the positive noise 18. Dvon is a peak voltage of an output waveform with respect to noise having a peak at a lower potential (hereinafter referred to as negative noise). In the present embodiment, Dvop and Dvon are, for example, differences between the peak potential of the output waveform and the power supply potential affected by the noise. PNL and NNL are voltage change allowances for positive noise and negative noise, respectively.
[0019]
Next, an example of a method for extracting noise propagation parameter characteristics will be described. As shown in FIG. 3, a circuit is prepared in which a load capacitor 62 is connected to an output terminal of an extraction target cell (hereinafter, referred to as an extraction target cell) 61. Then, a noise waveform 63 of a combination in which Vp and Tw are variously changed is input to the input terminal of the extraction target cell 61. At this time, the peak voltage (Dvop and Dvon) of the noise waveform 64 output from the output terminal is obtained and stored as a noise propagation parameter. The graph on the right side of FIG. 3 is a noise propagation graph 32 showing the relationship between Dvop and Vp and Tw.
[0020]
As shown in FIG. 4, the noise input to the extraction target cell 61 is a noise 65 having a peak at a lower potential from the positive power supply potential VDD and a noise having a peak at a lower potential from the negative power supply potential VSS. 66, a noise 67 having a peak at a higher potential from the positive power supply potential VDD, and a noise 68 having a peak at a higher potential from the negative power supply potential VSS are prepared. Then, characteristics of each type of noise are extracted.
[0021]
In the present embodiment, the case of positive noise is described as an example, but the same applies to the case of negative noise. Also, as shown in FIG. 3, the attached drawing shows the case of positive noise, but the same applies to the case of negative noise.
[0022]
Further, as shown in FIG. 5, for example, when extracting the noise propagation parameter characteristics, for each of the noise waveforms 63 in which Vp and Tw are variously changed, the generation time of the noise peak with respect to the full width Ttw of the noise waveform 63 is represented by Tnp1. , Tnp2,... Then, as shown in FIG. 6, various noise waveforms 63 having different noise peak occurrence times Tnp are input to the input terminal of the extraction target cell 61, and the peak voltage (Dvop and Dvop) of the noise waveform 64 output from the output terminal is output. Dvon). Thus, a noise propagation parameter when noise having various waveform shapes is input to the extraction target cell 61 is obtained. The graph on the right side of FIG. 6 is a noise propagation graph 33 showing the relationship between Dvop and Tnp.
[0023]
Also, as shown in FIG. 7, for example, when extracting the noise propagation parameter characteristics, the capacitance CL of the load capacitor 62 connected to the output terminal of the extraction target cell 61 is changed variously while the noise is applied to the input terminal of the extraction target cell 61. A noise waveform 63 having various waveform shapes is input, and peak voltages (Dvop and Dvon) of the noise waveform 64 output from the output terminal are acquired. Thus, a noise propagation parameter when the output load of the cell is different can be obtained. The graph on the right side of FIG. 7 is a noise propagation graph 34 showing the relationship between Dvop and CL. As shown in the graph 34, Dvop (and Dvon) has a characteristic that it increases when the output load is small, and decreases when the output load is large.
[0024]
Dvop and Dvon have a correlation with Tnp and CL described above. Therefore, based on the characteristic extraction results for various input noise waveform shapes and the characteristic extraction results for various output loads, Dvop and Dvon are represented as functions using Tnp, CL, and a dependency coefficient for those parameters. You can also.
[0025]
Next, the voltage change allowable amount will be described. The voltage change allowable amount is a maximum change amount of the input voltage when the input voltage to the cell is changed within a range where the output voltage from the cell does not change. For each cell, the potential changes from the positive power supply potential VDD to a lower potential, the negative power supply potential VSS changes to a lower potential, and the positive power supply potential VDD changes to a higher potential. A voltage change allowable amount is set for each of the case and the case where the potential changes from the negative power supply potential VSS to the higher one. In a cell in which the output voltage is sensitive to a change in the input voltage, the allowable voltage change has a small value. Conversely, in a cell in which the output voltage hardly changes with respect to the input voltage change, the voltage change allowable amount has a large value. Note that the value of the voltage change allowable amount can be intentionally controlled according to the characteristics of the cell.
[0026]
Also, even when the cells are the same, if the allowable voltage change is different depending on the operating conditions, the allowable voltage change is set for each operating condition. For example, when the input terminal of the cell is formed of a transfer gate, the allowable voltage change amount of the input terminal differs depending on the state of the transfer gate. As shown in FIG. 8, when the transfer gate 69 is non-conductive (closed), an allowable voltage change amount when the transfer gate 69 is closed is set. As shown in FIG. 9, when the transfer gate 69 is conductive (open), the allowable voltage change amount of the circuit (cell 61 in FIG. 9) connected to the transfer gate 69 is set.
[0027]
Next, a method for generating a cell-specific noise immunity parameter will be described. In generating the noise immunity parameter, it is recognized which of the four types of noise 65, 66, 67, 68 shown in FIG. The noise propagation parameter and the voltage change allowance corresponding to the type are handled. Here, for convenience of explanation, a case of positive noise in true logic will be described, but the same applies to other cases.
[0028]
First, as shown as step 1 in FIG. 10, circuit information is recognized, a circuit network for performing a noise error determination is selected, and a damaged cell 16 and a subsequent cell 21 connected to its output terminal are detected. Next, in step 2, the noise propagation parameters of the damaged cell 16 are read. Although not particularly limited, in the illustrated noise propagation graph 35, the noise propagation parameter is a characteristic extraction result under the condition that Vp is four and Tw is three.
[0029]
Next, in step 3, the voltage change allowable amount of the succeeding cell 21 is read. Then, the value of Vp and the value of Tw in which the noise propagation parameter Dvop and the allowable voltage change amount PNL match are obtained. At that time, if necessary, the values of Vp and Tw that match the PNL are calculated using two-point interpolation or an interpolation function. The obtained value is a noise immunity parameter, and is represented by a characteristic graph 52 as shown in Step 4. When the parameter value is a representative value, interpolation between two points or an interpolation function is used.
[0030]
The noise resistance parameter of the damaged cell 16 generated in this way is unique to the damaged cell 16 because the input state of the succeeding cell 21 and the allowable noise amount are taken into consideration. Therefore, when the allowable voltage change amount of the succeeding cell 21 changes depending on conditions, or when the succeeding cell 21 differs, even if the damaged cell 16 is the same cell, the noise immunity parameter may have a different value.
[0031]
Next, an error determination method will be described. As shown in FIG. 11, a characteristic graph 52 of the noise resistance parameter is used. The characteristic graph line of the graph 52 is a threshold value Vth for error determination. Therefore, when the values of the noise voltage value Vp and the noise width Tw of the output waveform obtained from the noise analysis result fall below the characteristic graph line of the noise immunity parameter, the noise is within the allowable range and no error occurs. If it exceeds the characteristic graph line, an error occurs because the noise exceeds the allowable range.
[0032]
A program for causing a computer to execute the above-described noise error determination method is recorded on a computer-readable recording medium, and is realized by the computer. Further, the noise error determination program can be distributed via a computer-readable recording medium via a network such as the Internet. FIG. 14 is a block diagram illustrating a configuration of an example of a computer provided for performing the above-described noise error determination method.
[0033]
The computer includes, for example, CPU 101, ROM 102, RAM 103, HDD (hard disk drive) 104, FDD (flexible disk drive) 106, display 108, keyboard 109, mouse and the like (including various pointing devices) 110, printer 111, and CD-ROM. Drives 112 are connected to each other via a bus 100. A program for causing a computer to execute the above-described noise error determination method is recorded on the FD 107 or the CD-ROM 113. The programs recorded on the FD 107 and the CD-ROM 113 are stored in the HD 105 and executed.
[0034]
According to the first embodiment, it is possible to determine the noise error under the condition unique to the cell while considering the influence on the succeeding cell and the shape of the generated noise and the output load condition at the same time. In addition, it is possible to estimate the noise tolerance of each cell more accurately by reflecting the cell-specific conditions in an actual circuit as faithfully as possible, and to judge a noise error based on necessary and sufficient criteria. Therefore, the reliability of the semiconductor integrated circuit can be improved.
[0035]
Further, according to the first embodiment, the criterion can be stricter for cells with high noise sensitivity, while the criterion can be relaxed for cells with low noise sensitivity. It is possible to avoid re-designing or the like. Therefore, the number of steps for developing the semiconductor integrated circuit can be reduced, and the development time can be reduced.
[0036]
(Embodiment 2)
In the second embodiment, in addition to Vp and Tw, parameters other than Vp and Tw, such as waveform shape parameter (Tnp) and output load capacitance (CL), are considered as parameters for changing the output waveform of noise. It is. When considering these parameters, Dvop is represented by a correlation characteristic and a correlation coefficient with Tnp and CL. Therefore, assuming that the correction coefficients are Ktnp and Kcl for Tnp and CL, respectively, Dvop obtained in consideration of Vp and Tw as in the first embodiment is newly calculated by the following equation (1). The peak voltage Dvop 'of the output waveform can be obtained.
[0037]
Dvop '= Dvop × Ktnp × ΔTnp × Kcl × ΔCL (1)
[0038]
When the noise resistance parameter is obtained, in step 3 shown in FIG. 10, the correction value (peak voltage Dvop ') calculated from the above equation (1) is used instead of the noise propagation parameter Dvop. The same applies to Dvon.
[0039]
FIG. 12 is a flowchart illustrating a processing procedure of the noise error determination method according to the second embodiment of the present invention. As shown in FIG. 12, characteristics are extracted for each cell in advance, and the relationship between Vp and Tw and Dvop or Dvon is acquired, and this is stored in a library as a noise propagation parameter (step S121). Further, the allowable voltage change amounts PNL and NNL are set for each input terminal of each cell and stored in the library (step S122).
[0040]
Next, based on the circuit information, a damaged cell and a subsequent cell connected to the output of the damaged cell are extracted from the connection information of the circuit (step S123), and the noise propagation parameters of the damaged cell 16 and the subsequent cell are extracted from the library. The voltage change allowable amount of the input terminal 21 is read. In addition, a noise analysis is performed (step S124), and a waveform shape parameter Tnp and an output load capacity CL of the damaged cell 16 are acquired from the analysis result. Then, the peak voltage Dvop of the output waveform is corrected based on the equation (1), and the allowable voltage change is compared with the correction value Dvop 'of the peak voltage of the output waveform to generate a noise resistance parameter of the damaged cell 16. (Step S125). The same applies to the case of negative noise.
[0041]
Next, a noise voltage value Vp and a noise width Tw are obtained from the noise analysis result, and the values of Vp and Tw are compared with the noise tolerance parameter generated in step S125 to determine a noise error ( Step S126). If the result of the determination is that there is a problem (step S126: NG), an error is output (step S127). If there is no problem (Step S126: OK), the error determination Check) ends (Step S128). Note that the method of extracting the characteristics of the noise propagation parameters, the setting contents of the allowable voltage change amount, and the method of determining an error are the same as those in the first embodiment.
[0042]
According to the second embodiment, the noise immunity parameter can be generated in consideration of the waveform shape of the noise and the output load of the cell. Therefore, the unique condition of the cell in the actual circuit can be reflected more faithfully. The noise resistance of each cell can be more accurately estimated, and the noise error can be determined based on necessary and sufficient criteria.
[0043]
(Embodiment 3)
In the third embodiment, error determination is performed without using a noise resistance parameter. For example, the peak voltage Dvop (or Dvon) of the noise output waveform is defined as a function of, for example, the noise voltage value Vp, the noise width Tw, the waveform shape parameter Tnp, and the output load capacitance CL, and the following equation (2) is derived. I do.
[0044]
Dvop = f (Vp, Tw, Tnp, CL) (2)
[0045]
Then, the value of Dvop obtained from the equation (2) is compared with the value of the allowable voltage change amount PNL. As a result, if PNL> Dvop, it is determined that an error has not occurred. If PNL <Dvop, it is determined that an error has occurred. The same applies to Dvon.
[0046]
Alternatively, the noise voltage value may be used as a reference for error determination. In this case, the noise voltage value Vp is, for example, a function of the noise width Tw, the waveform shape parameter Tnp, the output load capacitance CL, and the allowable voltage change amount PNL, and a correlation equation of the following equation (3) is derived.
[0047]
Vp = f (Tw, Tnp, CL, PNL) (3)
[0048]
When the noise voltage value obtained from the noise analysis result is expressed as Vp ′, this Vp ′ is compared with the value of Vp obtained from Expression (3). As a result, if Vp> Vp ′, it is determined that there is no error, and if Vp <Vp ′, it is determined that there is an error. The same applies to NNL. In addition, functions such as Tw, Tnp, and CL can be converted into functions and used as criteria for error determination.
[0049]
FIG. 13 is a flowchart illustrating a processing procedure of the noise error determination method according to the third embodiment of the present invention. As shown in FIG. 13, characteristics are extracted in advance for each cell and stored in a library as noise propagation parameters (step S131). Further, an allowable voltage change amount is set for each input terminal of each cell and stored in the library (step S132).
[0050]
Next, based on the circuit information, the damaged cell 16 and the subsequent cell 21 connected to the output of the damaged cell 16 are extracted from the circuit connection information (step S133), and the noise propagation parameters of the damaged cell 16 are extracted from the library, The voltage change allowable amount of the input terminal of the succeeding cell 21 is read. Further, noise analysis is performed (step S134), and the analysis result is obtained. Then, a noise error determination is performed using the above equation (2) or (3) (step S135).
[0051]
If the result of the determination is that there is a problem (step S135: NG), an error is output (step S136). If there is no problem (Step S135: OK), the error determination (Check) is completed (Step S137). The method of extracting the characteristics of the noise propagation parameters and the setting contents of the allowable voltage change amount are the same as those in the first embodiment.
[0052]
According to the third embodiment, the noise error determination can be performed without using the noise tolerance parameter. Therefore, even when the data amount of the noise propagation parameter increases, the graph reading method and the two-point interpolation method can be used. The effect on data processing time and memory consumption can be reduced.
[0053]
In the above, the present invention is not limited to the above-described embodiment, but can be variously modified. For example, the noise width may be the full width Ttw of the noise instead of the half width Tw of the noise. Further, the parameter representing the noise waveform shape is not limited to the noise peak occurrence time Tnp. Further, the output load is not limited to the capacity, but may be a resistance or the like.
[0054]
(Supplementary Note 1) In a noise error determination method for analyzing noise generated on signal wiring of a semiconductor integrated circuit and determining whether noise obtained as a result of the analysis is within an allowable range,
Based on the circuit information, a step of extracting a damaged cell receiving the noise and a subsequent cell connected to an output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. A step of determining an error based on the noise analysis result based on the allowable voltage change amount;
A noise error determination method comprising:
[0055]
(Supplementary Note 2) For each cell, when various types of noise are input to a circuit in which an output load is connected to the cell, the noise output from the cell is determined in advance, and the relationship between input noise and output noise is determined. The noise error determination method according to claim 1, wherein the noise error determination method is configured as a library as a propagation parameter.
[0056]
(Supplementary note 3) The noise error determination method according to supplementary note 2, wherein the noise propagation parameter is obtained by variously changing the peak voltage value of the input noise.
[0057]
(Supplementary note 4) The noise error determination method according to supplementary note 2, wherein the noise propagation parameter is obtained by variously changing a peak voltage value and a noise width of the input noise.
[0058]
(Supplementary note 5) The noise error determination method according to supplementary note 2, wherein the noise propagation parameter is obtained by variously changing a peak voltage value, a noise width, and a noise waveform shape of the input noise.
[0059]
(Supplementary note 6) The noise error determination method according to supplementary note 2, wherein the noise propagation parameter is obtained by variously changing a peak voltage value, a noise width, a noise waveform shape, and the output load of the input noise.
[0060]
(Supplementary note 7) In a noise error determination method for analyzing noise generated on signal wiring of a semiconductor integrated circuit and determining whether or not noise obtained as a result of the analysis is within an allowable range,
Extracting, from the circuit information, a damaged cell receiving the noise and a subsequent cell connected to an output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. Based on the permissible voltage change, generating a noise immunity parameter specific to the state of the damaged cell connected to the subsequent cell,
Comparing the noise analysis result with the noise immunity parameter to determine an error; and
A noise error determination method comprising:
[0061]
(Supplementary Note 8) For each cell, when various kinds of noise are input to a circuit in which an output load is connected to the cell, the noise output from the cell is obtained in advance, and the relationship between the input noise and the output noise is determined. 7. The noise error determination method according to claim 7, wherein a library is stored as a propagation parameter.
[0062]
(Supplementary note 9) The noise error determination method according to supplementary note 8, wherein the noise propagation parameter is obtained by variously changing the peak voltage value of the input noise.
[0063]
(Supplementary note 10) The noise error determination method according to supplementary note 8, wherein the noise propagation parameter is obtained by variously changing a peak voltage value and a noise width of the input noise.
[0064]
(Supplementary note 11) The noise error determination method according to supplementary note 8, wherein the noise propagation parameter is obtained by variously changing the peak voltage value, the noise width, and the noise waveform shape of the input noise.
[0065]
(Supplementary note 12) The noise error determination method according to supplementary note 8, wherein the noise propagation parameter is obtained by variously changing the peak voltage value, the noise width, the noise waveform shape, and the output load of the input noise.
[0066]
(Supplementary note 13) The noise error determination method according to any one of Supplementary notes 7 to 12, wherein the noise resistance parameter is generated in consideration of a noise waveform shape obtained from a noise analysis result.
[0067]
(Supplementary note 14) The noise error determination method according to any one of Supplementary notes 7 to 12, wherein the noise tolerance parameter is generated in consideration of an output load of the damaged cell obtained from a noise analysis result.
[0068]
(Supplementary note 15) The noise according to any one of Supplementary notes 7 to 12, wherein the noise tolerance parameter is generated in consideration of a noise waveform shape obtained from a noise analysis result and an output load of the damaged cell. Error determination method.
[0069]
(Supplementary Note 16) Any one of Supplementary Notes 1 to 15, characterized in that for each cell, an allowable voltage change amount corresponding to each of various circuit conditions and characteristics of the input section of the cell is obtained and stored in a library. The noise error determination method described in (1).
[0070]
(Supplementary note 17) The noise error determination method according to any one of Supplementary notes 1 to 16, further comprising a step of outputting an error when a result of the determination indicates that the noise analysis result exceeds an allowable range. .
[0071]
(Supplementary Note 18) In a noise error determination program that analyzes noise generated on signal wiring of a semiconductor integrated circuit and determines whether noise obtained as a result of the analysis is within an allowable range,
Based on the circuit information, a step of extracting a damaged cell receiving the noise and a subsequent cell connected to an output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. A step of determining an error based on the noise analysis result based on the allowable voltage change amount;
A noise error determination program that causes a computer to execute the program.
[0072]
(Supplementary Note 19) In a noise error determination program that analyzes noise generated on signal wiring of a semiconductor integrated circuit and determines whether noise obtained as a result of the analysis is within an allowable range,
Extracting, from the circuit information, a damaged cell receiving the noise and a subsequent cell connected to the output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. Based on the voltage change allowance, a step of generating a noise immunity parameter specific to the state in which the subsequent cell is connected to the damaged cell,
Comparing the noise analysis result and the noise immunity parameter, and performing an error determination;
A noise error determination program that causes a computer to execute the program.
[0073]
【The invention's effect】
According to the present invention, it is possible to more accurately reflect the cell-specific conditions in an actual circuit, more accurately estimate the noise immunity of each cell, and determine a noise error based on a necessary and sufficient criterion. Therefore, the reliability of the semiconductor integrated circuit can be improved. Further, according to the present invention, unnecessary layout correction and unnecessary re-designing can be avoided, so that the number of steps for developing a semiconductor integrated circuit can be reduced and the development time can be shortened. .
[Brief description of the drawings]
FIG. 1 is a flowchart illustrating a processing procedure of a noise error determination method according to a first embodiment of the present invention;
FIG. 2 is a conceptual diagram illustrating a noise error determination method according to the first embodiment of the present invention.
FIG. 3 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 4 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 5 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 6 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 7 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 8 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 9 is a conceptual diagram illustrating a method of extracting a noise propagation parameter.
FIG. 10 is a conceptual diagram illustrating a method of generating a noise immunity parameter unique to a cell.
FIG. 11 is a conceptual diagram illustrating an error determination method.
FIG. 12 is a flowchart illustrating a processing procedure of a noise error determination method according to the second embodiment of the present invention;
FIG. 13 is a flowchart illustrating a processing procedure of a noise error determination method according to the third embodiment of the present invention;
FIG. 14 is a block diagram illustrating a hardware configuration of a computer used for executing a noise error determination method according to the present invention.
FIG. 15 is a flowchart showing a processing procedure of a conventional noise error determination method.
FIG. 16 is a conceptual diagram illustrating a conventional noise error determination method.
[Explanation of symbols]
16 Damaged cell
21 Subsequent cells
61 cells (cells to be extracted)
62 Output load (load capacity)

Claims (10)

In a noise error determination method for analyzing noise generated on signal wiring of a semiconductor integrated circuit and determining whether noise obtained as a result of the analysis is within an allowable range,
Based on the circuit information, a step of extracting a damaged cell receiving the noise and a subsequent cell connected to an output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. A step of determining an error based on the noise analysis result based on the allowable voltage change amount;
A noise error determination method comprising: When various noises are input to a circuit in which an output load is connected to a cell for each cell, a noise output from the cell is obtained in advance, and a relationship between the input noise and the output noise is used as a noise propagation parameter as a library. 2. The noise error determination method according to claim 1, wherein the noise error is determined. The noise error determination method according to claim 2, wherein the noise propagation parameter is obtained by variously changing a peak voltage value, a noise width, a noise waveform shape, and the output load of the input noise. In a noise error determination method for analyzing noise generated on signal wiring of a semiconductor integrated circuit and determining whether noise obtained as a result of the analysis is within an allowable range,
Extracting, from the circuit information, a damaged cell receiving the noise and a subsequent cell connected to an output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. Based on the permissible voltage change, generating a noise immunity parameter specific to the state of the damaged cell connected to the subsequent cell,
Comparing the noise analysis result with the noise immunity parameter to determine an error; and
A noise error determination method comprising: When various noises are input to a circuit in which an output load is connected to a cell for each cell, a noise output from the cell is obtained in advance, and a relationship between the input noise and the output noise is used as a noise propagation parameter as a library. The noise error determination method according to claim 4, wherein: 6. The noise error determination method according to claim 5, wherein the noise propagation parameter is obtained by variously changing a peak voltage value, a noise width, a noise waveform shape, and the output load of the input noise. 7. The noise error determination method according to claim 4, wherein the noise tolerance parameter is generated in consideration of a noise waveform shape obtained from a noise analysis result and an output load of the damaged cell. . 8. A library according to claim 1, wherein for each cell, a voltage change allowable amount corresponding to each of various circuit conditions and characteristics of an input portion of the cell is obtained and stored in a library. Noise error determination method. In a noise error determination program that analyzes noise generated on signal wiring of a semiconductor integrated circuit and determines whether noise obtained as a result of the analysis is within an allowable range,
Based on the circuit information, a step of extracting a damaged cell receiving the noise and a subsequent cell connected to an output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. A step of determining an error based on the noise analysis result based on the allowable voltage change amount;
A noise error determination program that causes a computer to execute the program. In a noise error determination program that analyzes noise generated on signal wiring of a semiconductor integrated circuit and determines whether noise obtained as a result of the analysis is within an allowable range,
Extracting, from the circuit information, a damaged cell receiving the noise and a subsequent cell connected to the output terminal of the damaged cell;
A noise propagation parameter indicating a relationship between an output noise from the damaged cell and an input noise to the damaged cell, and a maximum change amount of an input to the subsequent cell within a range in which an output from the subsequent cell does not change. Based on the voltage change allowance, a step of generating a noise immunity parameter specific to the state in which the subsequent cell is connected to the damaged cell,
Comparing the noise analysis result and the noise immunity parameter, and performing an error determination;
A noise error determination program that causes a computer to execute the program.

Images