Model parameter extracting apparatus and model parameter extracting program for semiconductor device model

1. A model parameter extracting apparatus comprising:

a binning processor configured to carry out a binning process for limiting an effective range of a semiconductor device model used in a circuit simulation; and

a model parameter extractor configured to extract a model parameter used in the circuit simulation, for each of a plurality of bins formed by the binning process,

wherein

the model parameter extractor selects one of the plurality of bins as a target bin,

the model parameter extractor extracts a first model parameter for predicting an electric characteristic of a semiconductor device at a first end portion of the target bin,

based on the first model parameter, the model parameter extractor sets up a candidate for a second model parameter for predicting an electric characteristic of the semiconductor device at a second end portion of the target bin,

based on the first model parameter and the candidate for the second model parameter, the model parameter extractor calculates a limited curve representing the electric characteristic of the semiconductor device,

the model parameter extractor finds a start-point-side gradient representing a gradient of a tangent line of the limited curve including a start point; and an end-point-side gradient representing a gradient of a tangent line of the limited curve including an end point, and

based on a result of a comparison between the start-point-side gradient and the end-point-side gradient, the model parameter extractor extracts the second model parameter.

2. The model parameter extracting apparatus according to claim 1 , wherein

the model parameter extractor sets up a plurality of candidate parameters as the candidates for the second model parameter,

for the respective plurality of candidate parameters, the model parameter extractor generates a plurality of graphs, each representing an electric characteristic of the semiconductor device at an intermediate portion between the first end portion and the second end portion,

based on a result of a comparison between the start-point-side gradient and the end-point-side gradient of each of the plurality of graphs, the model parameter extractor determines a range from which to extract the second model parameter, and

the model parameter extractor extracts the second model parameter within the range thus determined.

3. The model parameter extracting apparatus according to claim 2 , wherein

based on the result of the comparison, the model parameter extractor holds allowable range information to be referred to when the range from which to extract the second model parameter is determined, and

based on information obtained from the result of the comparison and on the allowable range information, the model parameter extractor determines the range from which to extract the second model parameter.

4. The model parameter extracting apparatus according to claim 3 , wherein

the model parameter extractor receives setup range information indicative of a range of the plurality of candidate parameters, and thus sets up the plurality of candidate parameters one-by-one on the basis of the setup range information,

the model parameter extractor calculates an interpolation parameter for interpolating between each of the plurality of candidate parameters and the first model parameter,

based on the interpolation parameters respectively corresponding to the plurality of candidate parameters, the model parameter extractor generates a plurality of the limited curves, respectively, and

for each of the plurality of limited curves, the model parameter extractor finds the start-point-side gradient and the end-point-side gradient, and thus determines a range suitable for optimizing the second model parameter.

5. The model parameter extracting apparatus according to claim 3 , wherein

the model parameter extractor receives gradient range information as the allowable range information, the gradient range information indicating a range of a difference between the start-point-side gradient and the end-point-side gradient, and

based on the result of the comparison and the gradient range information, the model parameter extractor determines the range suitable for optimizing the second model parameter.

6. The model parameter extracting apparatus according to claim 1 , wherein

before extracting the first model parameter, the model parameter extractor extracts a global model parameter for reproducing the characteristic of the semiconductor device for all the plurality of bins, and

the model parameter extractor extracts the first model parameter on the basis of the global model parameter.

7. The model parameter extracting apparatus according to claim 1 , wherein

the model parameter extractor uses the second model parameter as a new first model parameter,

the model parameter extractor selects a bin adjacent to the target bin as a new target bin, and

on the basis of the new target bin and the new first model parameter, the model parameter extractor extracts a new second model parameter.

8. The model parameter extracting apparatus according to claim 7 , wherein

the binning processor forms the plurality of bins in such a manner that the distribution is divided in a matrix, and

the model parameter extractor sequentially selects a new target bin from bins arrayed in a first row of the matrix, and

if there is no new target bin yet to be selected from the first row of the matrix, the model parameter extractor extracts the first model parameter and the second model parameter for each of bins arrayed in a second row of the matrix.

9. A non-transitory computer-readable storage medium storing a model parameter extracting program instructing a procedure for causing a computer to function as a model parameter extracting apparatus for extracting a model parameter used in a circuit simulation for predicting an electric characteristic of a semiconductor device,

the program, when executed by the computer, performing steps comprising:

a binning processing step of carrying out a binning process for limiting an effective range of a semiconductor device model used in the circuit simulation; and

a model parameter extracting step of extracting the model parameter used in the circuit simulation, for each of a plurality of bins formed by the binning process,

wherein the model parameter extracting step comprises the steps of:

(a) selecting one of the plurality of bins as a target bin, and thus extracting a first model parameter for predicting an electric characteristic of the semiconductor device at a first end portion of the target bin;

(b) based on the first model parameter, setting up a candidate for a second model parameter for predicting the electric characteristic of the semiconductor device at a second end portion of the target bin;

(c) based on the first model parameter and the candidate for the second model parameter, calculating a limited curve representing the electric characteristic of the semiconductor device;

(d) finding a start-point-side gradient representing a gradient of a tangent line of the limited curve including a start point, and an end-point-side gradient representing a gradient of a tangent line of the limited curve including an end point; and

(e) on the basis of a result of a comparison between the start-point-side gradient and the end-point-side gradient, extracting the second model parameter.

10. The non-transitory computer-readable storage medium according to claim 9 , wherein

the step (c) includes the steps of:

setting up a plurality of candidate parameters as the candidates for the second model parameter; and

for the respective plurality of candidate parameters, generating a plurality of graphs, each representing an electric characteristic of the semiconductor device at an intermediate portion between the first end portion and the second end portion, and

the step (e) includes the steps of:

based on a result of a comparison between the start-point-side gradient and the end-point-side gradient of each of the plurality of graphs, determining a range from which to extract the second model parameter; and

extracting the second model parameter within the range thus determined.

11. The non-transitory computer-readable storage medium according to claim 10 , wherein the step (e) includes the steps of:

based on the result of the comparison, reading allowable range information to be referred to when the range from which to extract the second model parameter is determined; and

based on information obtained from the result of the comparison and on the allowable range information, determining the range from which to extract the second model parameter.

12. The non-transitory computer-readable storage medium according to claim 11 , wherein

the step (c) further includes the steps of:

receiving setup range information indicative of a range of the plurality of candidate parameters, and thus setting up the plurality of candidate parameters step-by-step on the basis of the setup range information;

calculating an interpolation parameter for interpolating between each of the plurality of candidate parameters and the first model parameter; and

based on the interpolation parameters respectively corresponding to the plurality of candidate parameters, generating a plurality of the limited curves, respectively, and

the step (e) further includes a step of, for each of the plurality of limited curves, finding the start-point-side gradient and the end-point-side gradient, and thus determining a range suitable for optimizing the second model parameter.

13. The non-transitory computer-readable storage medium according to claim 11 , wherein

the step (e) includes the steps of:

receiving gradient range information as the allowable range information, the gradient range information indicating a range of a difference between the start-point-side gradient and the end-point-side gradient; and

based on the result of the comparison and the gradient range information, determining the range suitable for optimizing the second model parameter.

14. The non-transitory computer-readable storage medium according to claim 9 ,

further comprising a step of, before extracting the first model parameter, extracting a global model parameter for reproducing the characteristic of the semiconductor device for all the plurality of bins, wherein

the step (a) includes a step of extracting the first model parameter on the basis of the global model parameter.

15. The non-transitory computer-readable storage medium according to claim 9 , further comprising the steps of:

(g) using the second model parameter as a new first model parameter, and a bin adjacent to the target bin as a new target bin; and

(h) based on the new target bin and the new first model parameter, extracting a new second model parameter.

16. The non-transitory computer-readable storage medium according to claim 15 , wherein

the binning processing step includes a step of forming the plurality of bins in such a manner that the distribution is divided in a matrix,

the step (g) includes a step of sequentially selecting the new target bin from bins arrayed in a first row of the matrix,

the step (h) includes a step of if there is no new target bin yet to be selected from the first row of the matrix, the model parameter extractor, extracting the first model parameter and the second model parameter for each of bins arrayed in a second row of the matrix.