JP3127876B2 - Model parameter extraction device and method, and storage medium storing the program - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a model parameter extracting apparatus and method for extracting model parameters of a model formula of an element model used in circuit simulation, and a storage medium storing the program.

[0002]

2. Description of the Related Art A highly integrated LSI (Large Scale Integr
For the development of an ated circuit), a circuit simulation by a computer is indispensable. Also, in order to perform circuit simulation, MOS (Metal Oxide
Semiconductor) FET (Field Effect Transistor)
There is a need for an element model that expresses characteristics such as BJT (Bipolar Juction Transistor) and the like with a model formula. Here, the element model is, for example, in the case of a MOSFET, the element shape (channel length L, channel width W), bias (drain-source voltage V _ds, gate-source voltage V _gs, bulk-source voltage V _bs ) and model The drain current I _ds for an arbitrary shape and an arbitrary bias can be calculated. In order to execute the circuit simulation with high accuracy, the parameters (model parameters) of the model formula corresponding to the device model are optimized so that the characteristics of the device model match the electrical characteristics of the experimental results or the process device simulation results. (Hereinafter, this is referred to as model parameter extraction).

In general, the operation of extracting model parameters for circuit simulation is performed in accordance with an extraction procedure including a plurality of extraction steps for optimizing a model parameter using characteristics of a transistor.
Conventionally, there is known modeling software for circuit simulation in which such a predetermined extraction procedure or an extraction procedure customized by a user can be executed as a batch job. For example, ICCA of HP (Hewllet Packard)
In P, the pre-installed extraction step is called PEL
The extraction procedure can be freely rearranged and executed as a batch job using an original customization language called.

Normally, when performing model parameter extraction work, a plurality of model parameter sets, which are sets of initial values of an enormous number of model parameters, are set as work conditions (hereinafter referred to as condition assignment). To perform model parameter extraction based on the model parameters assigned as conditions as a batch job, and to employ the most accurate model parameter set among a plurality of extracted model parameter sets. Extraction of model parameters with high accuracy, and TAT (Turn Around Time) of extraction work
Is being shortened.

A conventional model parameter extraction system will be described below with reference to FIGS. FIG. 6 is a block diagram showing an electrical configuration of a conventional model parameter extracting device, FIG. 7 is a flowchart showing a process of a batch job execution control unit in the device, and FIG. 8 is a process of an optimization processing unit in the device. It is a flowchart which illustrates.
As shown in FIG. 6, a conventional model parameter extraction device includes a model parameter initial value information 21, a model parameter extraction procedure information 22, a model parameter extraction control unit 23, a batch job execution control unit 24, an interactive execution control It is roughly composed of a unit 25, an optimization processing unit 26, characteristic measurement data information 27, characteristic model formula information 28, and extracted model parameter information 29.

The model parameter initial value information 21 is a file that stores information on initial values of a plurality of model parameter sets that are conditionally assigned. The model parameter extraction procedure information 22 is a file that stores information on an extraction procedure for extracting model parameters. The model parameter extraction control unit 23 uses the information on the number of conditional assignments of the model parameter initial values and the information on the model parameter extraction procedure to perform the model parameter extraction work in the batch job execution control unit 24 or the interactive execution control unit 25. Control. The batch job execution control unit 24 executes a model parameter extracting operation by batch job processing.
The interactive execution control unit 25 executes a model parameter extracting operation by an interactive process. The optimization processing unit 26 uses the information of the actual measurement data of the characteristic, the information of the characteristic model formula, and the information of the initial value of the model parameter to generate a model in the processing of the batch job execution control unit 24 or the interactive execution control unit 25. Perform parameter optimization processing. The characteristic measurement data information 27 is a file that stores characteristic measurement data of various elements whose model parameters are to be extracted. Characteristic model formula information 28
Is a file that stores model expressions representing the characteristics of various elements for which model parameters are to be extracted. The extracted model parameter information 29 is stored in the optimization processing unit 26.
Is a file that stores information on the optimized model parameters extracted by the processing of FIG.

Next, a model parameter extracting process by a batch job in a conventional model parameter extracting apparatus will be described with reference to FIGS. When performing model parameter extraction by a batch job, the model parameter extraction control unit 23 reads the model parameter initial value information and the model parameter extraction procedure information, and instructs the batch job execution control unit 24 to execute the extraction as a batch job. Pass information about the procedure. Since the model parameter initial value information stores all the initial values of the model parameter set subjected to the condition assignment, the extraction procedure defined in the model parameter extraction procedure information is executed the number of times corresponding to the number of condition assignments. The batch job execution control unit 24 controls the batch job to be executed.

As shown in FIG. 7, the batch job execution control section 24 first reads information on the extraction procedure and the initial values of the first model parameter set to which conditions are assigned as information on the first batch job to be executed (FIG. 7). 7 steps S
1). Then, the extraction step to be executed first is read (step S2 in FIG. 7), the optimization process of the read extraction step is executed (step S3 in FIG. 7), and it is determined whether the executed extraction step is the last of the extraction procedure (step S3). If it is not the last, the extraction step to be executed next is read (step S5 in FIG. 7), and the processing is repeated from the execution of the optimization processing of the read extraction step (step S3 in FIG. 7). When the executed extraction step is at the end of the extraction procedure (step S4 in FIG. 7), it is determined whether the executed batch job is the last batch job (step S6 in FIG. 7). Read the information of the batch job to be executed (step S in FIG. 7)
7) The batch job processing is repeatedly executed from the reading of the extraction step to be executed first (step S2 in FIG. 7). If the batch job to be executed is the last batch job (step S6 in FIG. 7), the processing is executed. To end.

As shown in FIG. 8, the optimization processing unit 26 reads actual measurement data to be optimized from the actual measurement data information for optimization stored in the actual measurement data information 27 of the characteristic (step P1 in FIG. 8). The model formula to be optimized is read from the characteristic model formula information 28 (step P2 in FIG. 8), and the initial value of the model parameter to be optimized from the model parameter initial value information stored in the model parameter initial value information 21 (Step P in FIG. 8)
3) The model formula used for optimization is calculated (step P4 in FIG. 8), the data of the calculation result of the model formula and the measured data are evaluated, and RMS (Root Mean Square) error and MA
X (Maximum) error is calculated (step P5 in FIG. 8),
Whether the RMS error has become smaller than the value specified as the end condition 1 (step P6 in FIG. 8), whether the MAX error has become smaller than the value specified as the end condition 2 (step P7 in FIG. 8), or as the end condition 3. Whether the rate of change of the RMS error has become smaller than the specified value (step P8 in FIG. 8) or whether the rate of change of the parameter has become smaller than the value specified as the end condition 4 (step P8 in FIG. 8).
9) Or, if it is determined that the number of repetitions has exceeded the value specified as the end condition 5 (step P10 in FIG. 8), and if none of the end conditions is satisfied, a change process of incrementing the model parameter value by, for example, a predetermined value The process is repeated (step P11 in FIG. 8), and the process is repeated again from the calculation of the model formula used for optimization (step P4 in FIG. 8). When any one of the end conditions is satisfied, the optimized model parameter information in the extracted model parameter information 29 is updated (step P12 in FIG. 8).
The process ends.

[0010]

However, in the model parameter extraction processing by a batch job in the conventional model parameter extraction device, the optimization processing unit 2
No. 6 sequentially executes each extraction step in accordance with the extracted extraction procedure, regardless of the type of the termination condition of the optimization. Therefore, no matter what the termination condition of the optimization process executed in the extraction step, the optimization process is performed for all the extraction steps defined in the read model parameter extraction procedure, and the batch job execution control is further performed. The unit 24 performs such processing for all batch jobs. However, even if the optimization end condition is satisfied in the extraction step of a certain batch job and the optimization processing is ended, for example, the condition that the number of optimization repetitions exceeds the defined extraction in the end condition 5 in FIG. If the process is completed with the conditions satisfied, it cannot be said that the model parameters have been optimized. In this case, even if the model parameters are optimized in the next and subsequent extraction steps, it cannot be said that the high-precision model parameters have been optimized in the batch job. On the other hand, depending on the type of characteristics to be optimized in the extraction step of the batch job, high-precision processing is performed by satisfying the condition that the number of optimization repetitions exceeds a defined value, as shown in the end condition 5. May not affect the extraction of model parameters in the next and subsequent extraction steps. in this way,
Regardless of what termination condition is satisfied by the optimizing process, there are some batch jobs for which the next and subsequent extraction steps should be executed.

In a conventional model parameter extraction processing system, when generating a batch job and extracting model parameters by the number of conditional assignments of the initial values of the model parameters, in the extraction step executed by the batch job, Whether or not subsequent extraction steps can be executed cannot be determined based on the optimization end condition. Therefore, even if the model parameters are not sufficiently optimized in a certain extraction step in a batch job, Regardless of the conditions, all of the subsequent extraction steps are executed, and such processing is performed in all batch jobs. For this reason, there is a problem that computer resources are wasted, and that highly accurate model parameter extraction cannot be performed in a short time.

The present invention has been made in view of the above circumstances, and in a model parameter extraction system, it is possible to reduce the cost of using a computer and execute highly accurate model parameter extraction in a short time. It is an object of the present invention to provide an apparatus and method for extracting model parameters and a recording medium on which the program is recorded.

[0013]

According to a first aspect of the present invention, there is provided a model parameter extracting apparatus for generating a batch job corresponding to a model parameter extracting procedure including a plurality of extracting steps. First control means generated for each initial value of a model parameter set to which a condition corresponding to a model expression is assigned, and an extraction step for executing an optimization process of a model parameter for each extraction step constituting the batch job. It is determined whether the end condition satisfies the condition for executing the next extraction step,
When the condition is satisfied, the execution of the next extraction step is continued, and when the condition is not satisfied, the process of stopping the execution of the batch job and shifting to the execution of the next batch job is performed for all the batch jobs. And control means.

According to a second aspect of the present invention, there is provided the model parameter extracting device according to the first aspect, wherein the second control means calculates a model equation of the element according to an initial value of the model parameter. When the evaluation result of the value and the measured value of the characteristic of the element does not satisfy any of the plurality of termination conditions, the same processing is repeated by changing the value of the model parameter, and the evaluation result satisfies any of the termination conditions. And a processing unit for optimizing the model parameters for extracting the model parameter values of the above as the optimum values.

According to a third aspect of the present invention, there is provided a model parameter extracting method, wherein a batch job corresponding to a model parameter extracting procedure including a plurality of extracting steps is conditionally assigned according to a model equation of a transistor element. A first step generated for each initial value of the above, and an evaluation result of a value obtained by calculating the model formula of the transistor element according to the initial value of the model parameter and an actually measured value of the transistor characteristic of the element are used for a plurality of termination conditions. If none of the conditions is satisfied, the same process is repeated by changing the value of the model parameter, and the model parameter optimization process of extracting the model parameter value as the optimum value when the evaluation result satisfies one of the termination conditions is performed in the batch. The conditions for ending extraction steps when executed for each extraction step that composes a job are as follows: It is determined whether or not a condition for executing the extraction step is satisfied. When the condition is satisfied, the execution of the next extraction step is continued. When the condition is not satisfied, the execution of the batch job is stopped and the next batch is executed. And a second step of performing processing for shifting to job execution for all batch jobs.

[0016]

According to a fourth aspect of the present invention, there is provided a computer-readable storage medium storing a program for executing the model parameter extracting method according to the third aspect of the present invention. In a first step, a batch job corresponding to a model parameter extraction procedure including a plurality of extraction steps is generated for each initial value of a model parameter set to which a condition is assigned corresponding to a model equation of a transistor element. Changing the value of the model parameter when the evaluation result of the calculated value of the model formula of the transistor element according to the initial value of the model parameter and the measured value of the transistor characteristic of the element does not satisfy any of a plurality of termination conditions. And repeat the same process, and if the evaluation result When the model parameter optimizing process for extracting the model parameter value at the time of addition is performed for each of the extracting steps constituting the batch job, the ending condition of the extracting step satisfies the condition for executing the next extracting step Processing for determining whether or not the condition is satisfied, and continuing the execution of the next extraction step when the condition is satisfied, and stopping the execution of the batch job when not satisfying the condition, and shifting to the execution of the next batch job Is recorded for all batch jobs.

[0018]

[0019]

According to the structure of the present invention, a batch job corresponding to a model parameter extraction procedure including a plurality of extraction steps is generated for each initial value of a model parameter set corresponding to a model equation of an element, and each batch job constituting the batch job is generated. It is determined whether or not the end condition of the extraction step when the model parameter optimization process is executed for each extraction step satisfies a condition for executing the next extraction step. When the condition is satisfied, execution of the next extraction step is performed. When this condition is not satisfied, the execution of the batch job is stopped and the process of shifting to the execution of the next batch job is performed for all the batch jobs, so that unnecessary processing can be omitted. . In this case, the optimization process of the model parameter does not satisfy any of the plurality of termination conditions, the evaluation result of the value calculated from the model formula of the element according to the initial value of the model parameter and the measured value of the characteristic of the element. When the model parameter value is updated and the same processing is repeated, the evaluation result, the model parameter value when one of the end conditions is satisfied is extracted as the optimum value, and the next extraction step is executed according to the end condition at this time. This is performed by updating an end condition for judging whether or not this is possible.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. The description will be made specifically using an embodiment. FIG. 1 is a block diagram showing an electrical configuration of a model parameter extracting device according to an embodiment of the present invention, FIG. 2 is a flowchart showing processing of a batch job execution control unit in the device, and FIG. FIG. 4 is a flowchart illustrating the processing of the optimization processing unit.
FIG. 5 is a diagram illustrating an example of an extraction procedure in the device, and FIG. 5 is a diagram illustrating an example of extraction step execution determination information in the device. As shown in FIG. 1, the model parameter extraction device of this example includes a model parameter initial value information 1, a model parameter extraction procedure information 2, a model parameter extraction control unit 3, a batch job execution control unit 4, an extraction step It is roughly composed of execution determination information 5, optimization processing unit 6, actual measurement data information 7, characteristic model formula information 8, extracted model parameter information 9, and optimization end condition information 10. .

The model parameter initial value information 1 is a file storing information on initial values of a plurality of model parameter sets that are conditionally assigned. The model parameter extraction procedure information 2 is a file that stores information on an extraction procedure for extracting model parameters. The model parameter extraction control unit 3 controls the model parameter extraction work in the batch job execution control unit 4 using the information on the number of conditions of the model parameter initial value and the information on the model parameter extraction procedure. The batch job execution control unit 4
Execute model parameter extraction work by batch job processing. The extraction step execution determination information 5 is a file for storing information indicating whether or not to execute the next model parameter extraction step in each model parameter extraction step for each end condition. The optimization processing unit 6 uses the information on the measured characteristic data, the information on the characteristic model formula, and the information on the initial values of the model parameters to perform the optimization of the model parameters in the processing of the batch job execution control unit 4. Perform Characteristic measurement data information 7
Is a file that stores actual measurement data of the characteristics of various elements whose model parameters are to be extracted. The characteristic model formula information 8 is a file that stores model formulas that represent the characteristics of various elements whose model parameters are to be extracted. The extracted model parameter information 9 is a file that stores the information of the optimized model parameters extracted by the processing of the optimization processing unit 6. The optimization end condition information 10 is a file that stores the end condition of the optimization processing in the optimization processing unit 6.

Next, the model parameter extracting process in this example will be described with reference to FIGS. The model parameter extraction control unit 3 shown in FIG. 1 includes model parameter initial value information and an extraction procedure shown in, for example, an extraction step A, an extraction step B, an extraction step C, an extraction step D, and an extraction step E as shown in FIG. Is read, and information of an extraction procedure to be executed as a batch job is passed to the batch job execution control unit 24. In the model parameter initial value information, all the initial values of the model parameter set subjected to the condition assignment are stored, so the extraction procedure defined in the model parameter extraction procedure information is performed by the number of times corresponding to the number of this condition assignment. A batch job to be executed is controlled to be executed by the batch job execution control unit 4.

As shown in FIG. 2, the batch job execution control unit 4 first reads the extraction procedure and the initial values of the first model parameter set to which conditions are assigned as information on the batch job to be executed first (FIG. 2). Step Q1). Then, the first extraction step to be executed is read (step Q2), the optimization processing of the extracted extraction step is executed (step Q3), and whether or not the next extraction step should be executed according to the termination condition of the executed extraction step From the extraction step execution determination information 5 (step Q
4) Check whether the end condition of the optimization process does not satisfy the condition for executing the next extraction step (step Q
5) If it satisfies, the extraction step to be executed next is read (step Q6), and the optimization processing of the read extraction step is executed (step Q3), and the processing is repeated again. If the end condition of the optimization process does not satisfy the condition for executing the next extraction step (step Q5), it is determined whether the executed batch job is the last batch job (step Q7) and the last batch job is executed. If not, the information of the batch job to be executed next is read (step Q8), and the process of the batch job is executed by repeating the process again from the reading of the extraction step to be executed first (step Q2). If it is the batch job (step Q7), the process ends.

In the optimization processing unit 6, as shown in FIG.
The actual measurement data to be optimized is read from the actual measurement data information for optimization stored in the characteristic actual measurement data information 7 (step R1), and the model expression to be optimized is read from the characteristic model expression information 8 (step R1). R2), the initial values of the model parameters to be optimized are read from the information on the initial values of the model parameters stored in the model parameter initial value information 1 (step R3), and the model formula used for the optimization is calculated (step R4). ), The data of the calculation result of the model formula and the actually measured data are evaluated, and the RMS error and M
The AX error is calculated (step R5), and whether the RMS error has become smaller than the value specified as the end condition 1 (step R6) or whether the MAX error has become smaller than the value specified as the end condition 2 (step R7) Or whether the change rate of the RMS error has become smaller than a value specified as the end condition 3 (step R8), or the end condition 4
If the change rate of the parameter is smaller than the specified value (step R9) or whether the number of repetitions exceeds the specified value as the end condition 5 (step R10), if none of the end conditions is satisfied, Then, a change process for incrementing the value of the model parameter by, for example, a predetermined value is performed (step R11), and the process is repeated again from the calculation of the model formula used for optimization (step R4). When any one of the end conditions is satisfied, the optimization end condition information in the optimization end condition information 10 is updated (step R12), and the optimized model parameter information in the extracted model parameter information 9 is updated. Then (step R13), the process ends.

The information stored in the extraction step execution determination information 5 as to whether or not to execute the next extraction step based on the termination condition of the executed extraction step is as shown in FIG. That is, as shown in each column of FIG. 5, the extraction step A, the extraction step B, the extraction step C, the extraction step D, and the extraction step E shown in FIG.
For each extraction step, GO and STO are determined according to which one of the end condition 1, the end condition 2, the end condition 3, the end condition 4, and the end condition 5 terminates the optimization process.
P. In the table shown in FIG. 5, in a certain extraction step, when the optimization processing has been completed under the termination condition indicated by GO, the processing of the next extraction step in the extraction step shown in FIG. 4 is executed. , STOP is terminated under the displayed termination condition, the processing of the batch job is terminated without executing the processing of the next extraction step in the extraction step shown in FIG.

Specific examples of the case where a difference is provided in the termination condition of the optimization of the model parameters include, for example, steps 1 to 5 for optimizing the model parameters for determining the threshold value of the transistor, There is the following batch job, which comprises an extraction procedure including steps 6 and 7 for optimizing model parameters for determining Step of optimizing parameters for determining threshold value Extraction step 1: Optimization of threshold voltage parameters A, B, C Extraction step 2: Optimization of threshold voltage parameters D, E, F Extraction step 3: Optimization of threshold voltage parameters G, H, I Extraction step 4: optimization of threshold voltage parameters J, K, L Extraction step 5: optimization of threshold voltage parameters M, N, O Determine drain current Step of optimizing parameters Extraction step 6: Optimization of current parameters P and Q Extraction step 7: Optimization of current parameters R and S In this case, voltage parameters J, K and L of extraction step 4
In these optimizations, even if these parameters are not optimized, the accuracy of the threshold is hardly affected. On the other hand, if the voltage parameters and current parameters other than the extraction step 4 are not optimized with sufficient accuracy, the parameters cannot be optimized with high accuracy even if the next extraction step is executed.

As described above, according to this example, the batch job execution control unit 4 executes the optimization processing of the extraction step for all the batch jobs, and then executes the extraction step execution determination information 5 corresponding to the extraction step. Is compared with the end condition of the optimization processing in this extraction step in the extracted model parameter information 9 to determine whether to execute the next extraction step. For example, if the end condition of the optimization is the end condition 5 after executing the extraction step A in a certain batch job, the extraction steps after the extraction step B are not executed as a batch job. On the other hand, if the end condition is one of end conditions 1 to 4, the next extraction step B is executed. Therefore,
In the model parameter extraction system of this example, the number of batch jobs is generated by the number of conditions of the initial values of the model parameter set, and in the extraction steps constituting the extraction procedure in all the batch jobs, Since the execution condition of the extraction step to be executed next is controlled according to the end condition of the extraction step, there is no waste of computer resources, the cost of computer use can be reduced, and high-precision model parameter extraction can be performed. In a short time.

Although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and there are design changes and the like without departing from the gist of the present invention. Even this is included in the present invention. For example, the model parameter extraction device of this example may be entirely or partially of a hardware configuration or a software configuration. A part of the model parameter extraction device of this example (for example, the model parameter extraction control unit 3, the batch job execution control unit 4, the optimization processing unit 6, etc.) includes a CPU (central processing unit) and an internal storage such as a ROM or a RAM. Device and FDD
(Flexible disk driver), HDD (hard disk driver), a computer having an external storage device such as a CD-ROM driver, an input device such as a keyboard and a mouse, and a display device are preferable. On a recording medium such as an FD (flexible disk), HD (hard disk), or CD-ROM, a model parameter extraction program for causing a computer to execute the processing of the above-described embodiment (FIGS. 1, 2, and 3) is recorded. Is preferred. The recording medium may be a semiconductor memory or another recording medium. In such a configuration, the model parameter extraction program is read into the computer from the recording medium and controls the operation of the computer. When the model parameter extraction program runs, the computer executes the function of each unit described above.

[0029]

As described above, according to the present invention, in a batch job for sequentially executing each extraction step defined in the model parameter extraction procedure information, the number of conditions of the initial value of the model parameter set is reduced. When a batch job is generated and executed by the number of minutes, depending on the end condition of the optimization performed in each extraction step of each batch job,
It is determined whether or not to execute the next extraction step.If the termination condition is not sufficiently optimized for the model parameters, the process of stopping the execution of the next extraction step is performed in all cases. Therefore, it is possible to reduce the use cost of the computer constituting the system and to perform highly accurate model parameter extraction in a short time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an electrical configuration of a model parameter extracting device according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a processing procedure of a batch job execution control unit in the apparatus.

FIG. 3 is a flowchart illustrating a process of an optimization processing unit in the apparatus.

FIG. 4 is a diagram showing an example of an extraction procedure in the same device.

FIG. 5 is a diagram showing an example of extraction step execution determination information in the same device.

FIG. 6 is a block diagram showing an electrical configuration of a conventional model parameter extraction device.

FIG. 7 is a flowchart illustrating a processing procedure of a batch job execution control unit in the apparatus.

FIG. 8 is a flowchart illustrating a process of an optimization processing unit in the apparatus.

[Explanation of symbols]

1 Model parameter initial value information 2 Model parameter extraction procedure information 3 Model parameter extraction control unit (first control means) 4 Batch job execution control unit (second control means) 5 Extraction step execution determination information 6 Optimization processing unit ( Processing means) 7 Characteristic measured data information 8 Characteristic model formula information 9 Extracted model parameter information 10 Optimization end condition information

Continuation of the front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 17/50 662 G06F 9/46 340 JICST file (JOIS)

Claims (4)

(57) [Claims] 1. A first control means for generating a batch job corresponding to a model parameter extraction procedure including a plurality of extraction steps for each initial value of a model parameter set in which a condition is assigned corresponding to a model equation of an element. When it is determined whether or not the end condition of the extraction step when the optimization process of the model parameter is executed for each extraction step constituting the batch job satisfies the condition for executing the next extraction step, and the condition is satisfied. A second control means for continuing the execution of the next extraction step and, when the condition is not satisfied, stopping the execution of the batch job and shifting to the execution of the next batch job for all the batch jobs. A model parameter extraction device, comprising: 2. An evaluation result of a value obtained by calculating a model formula of an element according to an initial value of a model parameter and an actually measured value of a characteristic of the element is given to the second control means by one of a plurality of end conditions. The same process is repeated by changing the value of the model parameter when the condition is not satisfied, and extracting the model parameter value as the optimum value when the evaluation result satisfies one of the termination conditions 2. The model parameter extracting device according to claim 1, further comprising: means. 3. A first step of generating a batch job corresponding to a model parameter extraction procedure including a plurality of extraction steps for each initial value of a model parameter set in which a condition corresponding to a model formula of a transistor element is assigned; When the evaluation result between the value calculated from the model formula of the transistor element according to the initial value of the parameter and the measured value of the transistor characteristic of the element does not satisfy any of a plurality of termination conditions, the value of the model parameter is changed. The same process is repeated, and when the evaluation result satisfies any one of the termination conditions, the model parameter value is extracted as an optimal value. The model parameter optimizing process is performed for each of the extracting steps constituting the batch job. It is determined whether or not the condition for ending the step satisfies the condition for executing the next extraction step. When the condition is satisfied, the execution of the next extraction step is continued, and when the condition is not satisfied, the process of stopping the execution of the batch job and shifting to the execution of the next batch job is performed for all the batch jobs. And a step for extracting model parameters. 4. A computer-readable recording medium on which a program for executing the model parameter extracting method according to claim 3 is recorded, wherein in the first step,
A batch job corresponding to a model parameter extraction procedure including a plurality of extraction steps is generated for each initial value of a model parameter set to which a condition is assigned corresponding to a model formula of a transistor element. In a second step, an initial value of a model parameter is set. When the evaluation result of the calculated value of the model formula of the transistor element and the measured value of the transistor characteristic of the element does not satisfy any of a plurality of termination conditions, the same processing is repeated by changing the value of the model parameter. And an ending condition of the extracting step when the model parameter optimizing process for extracting the model parameter value as the optimum value when the evaluation result satisfies any one of the ending conditions is executed for each of the extracting steps constituting the batch job. Judge whether the condition for executing the next extraction step is satisfied or not. In addition, a program for continuing the execution of the next extraction step when the condition is not satisfied, and executing a process of stopping the execution of the batch job and shifting to the execution of the next batch job when all the conditions are not satisfied is recorded. A storage medium storing a model parameter extraction program characterized by the following.