JP7247447B2 - CIRCUIT DESIGN DEVICE, CIRCUIT DESIGN METHOD AND PROGRAM - Google Patents

Description

The present invention relates to a circuit design apparatus and circuit design method for designing a circuit of a semiconductor integrated circuit, and a circuit design program for constructing the circuit design apparatus by being executed by a computer.

Design techniques for integrated circuits such as LSI (Large Scale Integration) have been developed.
For example, as a conventional design method, a method has been developed in which a circuit characteristic equation is defined, and a designer with extensive knowledge of circuits formulates a design policy to design an integrated circuit with desired characteristics. In this method, the quality of the circuit obtained by designing depends on the knowledge and experience of the designer, so that it is difficult for a designer with little knowledge and experience to design a high-quality circuit.

On the other hand, assuming the configuration of a target circuit (target circuit), simulations such as SPICE are performed for each transistor that constitutes the circuit, and optimization is performed to obtain a transistor having desired characteristics, and thus the transistor. A design method has been developed for obtaining a target circuit having desired characteristics, which is composed of (see, for example, Patent Document 1).

However, simulation using a genetic algorithm or the like is highly random and results in a lot of wasteful searching. Therefore, if the number of circuit elements constituting a circuit increases, the simulation takes time and design work takes a long time. In addition, there is a problem that the design does not have a learning function, so there is no reproducibility.

JP 2011-065281 A

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to facilitate circuit design in a short time without depending on the knowledge and subjectivity of the designer.

According to the first aspect of the present invention,
an input unit for inputting a circuit configuration of a circuit under design, a circuit characteristic item name of the circuit under design, and an element name of a circuit element constituting the circuit under design;
a simulation unit that generates the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items by simulating the circuit under design while changing the element values of the circuit elements;
In a learning mode, deep learning is performed using the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items, and in the design mode, the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic items, A circuit designing device is provided, comprising a design processing unit that generates at least one of the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements by regression analysis.

Moreover, according to the second aspect of the present invention,
A circuit design method executed by a circuit design device comprising an input unit, a simulation unit, and a design processing unit,
the input unit inputs a circuit configuration of the circuit under design, a circuit characteristic item name of the circuit under design, and an element name of a circuit element constituting the circuit under design;
The simulation unit generates the element values of the circuit elements and the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements by simulating the circuit under design while changing the element values of the circuit elements;
The design processing unit performs deep learning using the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items in the learning mode, and the circuit corresponding to the characteristic values of the circuit characteristic items in the design mode. A circuit design method is provided, wherein at least one of an element value of an element and a characteristic value of the circuit characteristic item corresponding to the element value of the circuit element is generated by regression analysis.

Further, according to a third aspect of the present invention, there is provided a circuit design program for causing a computer to construct the circuit design apparatus.

According to the circuit design apparatus and circuit design method of the present invention, it is possible to easily design a circuit in a short time without depending on the knowledge and subjectivity of the designer.

Further, by executing the circuit design program of the present invention on a computer, it is possible to construct a circuit design apparatus capable of easily designing a circuit in a short time without depending on the knowledge and subjectivity of the designer. can.

1 is a block diagram of a circuit design device according to an embodiment of the present invention; FIG. 4 is a flow chart of the circuit design device according to the embodiment of the present invention; 1 is a circuit configuration diagram of a circuit under design according to an embodiment of the present invention; FIG. 1 is an explanatory diagram of a circuit design device according to an embodiment of the present invention; FIG. 1 is a partial circuit configuration diagram of a circuit design device according to an embodiment of the present invention; FIG. 1 is an explanatory diagram of a circuit design device according to an embodiment of the present invention; FIG. 1 is a partial circuit configuration diagram of a circuit design device according to an embodiment of the present invention; FIG. 1 is an explanatory diagram of a circuit design device according to an embodiment of the present invention; FIG. FIG. 4 is a circuit configuration diagram of a circuit under design according to another embodiment of the present invention; FIG. 10 is an explanatory diagram of a circuit design device according to another embodiment of the present invention;

Hereinafter, a circuit designing apparatus, a circuit designing method, and a circuit designing program for constructing the circuit designing apparatus by being executed by a computer according to embodiments of the present invention will be described with reference to the drawings.
In the embodiments described below, examples of analog circuits will be described, but the present invention can also be applied to digital circuits. Also, in each figure, the same reference numerals are assigned to portions having the same function.

FIG. 1 is a block diagram of a circuit design device 100 according to an embodiment of the invention.
In FIG. 1 , the circuit design apparatus 100 includes a data processing section 101 , an operation section 102 , a display section 103 and a storage section 104 .
The circuit design device 100 can be configured by a computer having an operation section such as a keyboard and mouse, a display section, a central processing unit (CPU), a storage section, and an input/output section such as a USB port. The computer can function as the circuit design apparatus 100 by causing the computer to execute the circuit design program stored in the program storage area 116 of the storage unit 104 .

An operation unit 102 is operated by a user to input data and instructions, and is composed of a mouse, a keyboard, and the like. The operation unit 102 is used to input data such as the circuit configuration of a circuit to be designed (circuit to be designed), the names of circuit characteristic items of the circuit to be designed, and the element names of circuit elements constituting the circuit to be designed. It is used to instruct the execution of processing, etc.
The data processing unit 101 has a simulation unit 105 and a design processing unit 106 . The simulation unit 105 has a function of generating the element values of the circuit elements and the characteristic values of the corresponding circuit characteristic items by simulating the circuit under design while changing the element values of the circuit elements constituting the circuit under design. have.

The design processing unit 106 performs deep learning using the element values of the circuit elements and the characteristic values of the corresponding circuit characteristic items in the learning mode, and performs deep learning of the circuit elements corresponding to the characteristic values of the circuit characteristic items in the design mode. It has a function of generating at least one of an element value and a characteristic value of the circuit characteristic item corresponding to the element value of the circuit element by regression analysis.
The design processing unit 106 also includes a nondimensionalization unit 107 that nondimensionalizes the characteristic values of the circuit characteristic items generated by the simulation unit 105 and the element values of the circuit elements.

In addition, the design processing unit 106 performs deep learning in the learning mode using the dimensionless characteristic values and the dimensionless element values corresponding thereto, and in the design mode, the characteristic values of the circuit characteristic items and the characteristic value of the circuit characteristic item corresponding to the element value of the circuit element are generated as non-dimensional values by regression analysis.

Further, the design processing unit 106 converts the element values corresponding to the dimensionless characteristic values of the circuit characteristic items from the generation processing unit 108 and the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements into the original dimensions. and a restoration unit 109 for restoring and outputting.
In the present embodiment, the dimensionless unit 107 normalizes the characteristic values of the circuit characteristic items generated by the simulation unit 105 and standardizes the element values of the circuit elements so that the characteristic values of the circuit characteristic items and the circuit element has the function of making the element values of

The characteristic values of the circuit characteristic items are normalized by the relational expression [Yi=(Xi-Xmin)/(Xmax-Xmin)]. Here, Xi is the characteristic value of the circuit characteristic item, Xmin is the minimum characteristic value of the circuit characteristic item, Xmax is the maximum characteristic value of the circuit characteristic item, and Yi is the characteristic value of the circuit characteristic item. is the normalized value.

The element values of circuit elements are standardized according to the relational expression [standardized element value=(element value of relevant circuit element)/(initial value of element value of relevant circuit element)].
In this way, the characteristic values are normalized and the element values are standardized to make them dimensionless. As will be described later, this improves the accuracy.

The characteristic values and element values restored to their original dimensions by the restoration unit 109 are displayed by the display unit 103 and stored in the storage device 104 .
The storage unit 104 has a data storage area 110 that stores various data, and a program storage area 116 that stores various programs such as the circuit design program according to the present embodiment.

The data storage area 110 contains circuit configuration data 111 that is data of the circuit topology of the circuit under design, circuit item data 112 that is data of items related to the circuit under design, and learning data that is data for the generation processing unit 108 to learn. Data 115 is stored at any time or in advance.
The circuit item data 112 includes circuit characteristic item name data 113, which is the name of each circuit characteristic of the circuit to be designed, and circuit element name data 114, which is the names of circuit elements such as MOS transistors and resistors forming the circuit to be designed. is

FIG. 2 is a flow chart showing processing of the circuit design apparatus 100 according to the embodiment of the present invention, and mainly shows processing of the data processing unit 101 .
FIG. 3 is a circuit configuration diagram of a design target circuit 301 to be designed in the embodiment of the present invention, and shows an example of an operational amplifier whose differential input section is composed of PMOS transistors.

In FIG. 3, transistors M1 and M2 are PMOS transistors having the same characteristics. Also, the transistors M3 and M4 are NMOS transistors having the same characteristics. The circuit is composed of substantially six MOS transistors M1 to M8 with different characteristics and one resistor R, and the number of element values of the circuit elements is substantially seven.

In this embodiment, there are 13 items as circuit characteristic items of the circuit under design 301 in FIG. Circuit characteristic item names are current consumption (Idis), power consumption (Pdis), DC gain (DCgain), phase margin (PM), gain bandwidth product (GBP), slew rate (SR), total harmonic distortion (THD) ), common mode rejection ratio (CMRR), power supply rejection ratio (PSRR), output voltage range (OVR), common mode input range (CMIR), output resistance (OR), and input referred noise (IRN).

FIG. 4 is a diagram showing probability distributions used for changing the characteristics of the transistors M1 to M8 when the simulation unit 105 simulates the circuit under design 301. In FIG.
The horizontal axis is the change rate of the variable M, and the vertical axis is the number of occurrences of the variable M. Here, the variable M is a predetermined integral multiple of the gate width W of the transistor. The characteristic is such that there are two peaks centered when the variable M is 1.

By fixing the gate width W and the gate length L of each of the transistors M1 to M8 and changing the value of the variable M, the gate width W is substantially changed. As a result, the element values of the transistors M1 to M8 are changed according to the probability distribution shown in FIG. 4, thereby changing the simulation conditions. Transistors M1 and M2 are changed to the same value, and transistors M3 and M4 are changed to the same value. The simulation is performed by changing the resistance value of the resistor R to a plurality of predetermined values.

FIG. 5 is a diagram showing the configuration of the generation processing unit 108. As shown in FIG. The generation processing unit 108 is configured using a regression analysis model of a feedforward neural network.
In each unit of the interlayer, the activation function Z uses a ReLU (Rectified Linear Unit) (ramp) function. The activation function Z of the output layer is subjected to regression analysis using the identity function.
Relations and outputs between layers are represented by the following equation (1). where x is the input, W is the weight, b is the bias, and u is the total input to each unit.
u=Wx+b, Z=f(u) (1)

As operation modes of the generation processing unit 108, there are a learning mode in which learning is performed by deep learning and a design mode in which circuit design processing is performed. Although the configuration of the generation processing unit 108 shown in FIG. 5 does not change between the learning mode and the design mode, learning and design are performed by switching between the learning mode and the design mode.

When functioning to design element values based on characteristic values, in the learning mode, the characteristic values obtained in the simulation (the characteristic values of 13 types of characteristic items in this embodiment) are input, and the corresponding element values ( In this embodiment, the weight w and the bias b are updated so that the values of substantially seven elements are output. In the design mode after learning, when a characteristic value is input, it functions to generate and output the element value of each circuit element corresponding thereto.

When functioning to design the characteristic values based on the element values, in the learning mode, the weight w and the bias b are updated so that the element values obtained by the simulation become the input and the corresponding characteristic values become the output. In the design mode after learning, when the element value of each circuit element is input, it functions to generate and output the corresponding characteristic value.

The example in FIG. 5 is an example in which the element value is designed based on the characteristic value. In the learning mode, the characteristic value obtained by the simulation is input, and the corresponding element value is output. It works like learning by doing. In the design mode after learning, when a new characteristic value is input, it functions to generate and output the element value of each circuit element corresponding thereto.

FIG. 6 is an explanatory diagram of the circuit design apparatus 100 according to the embodiment of the present invention, and is a table showing results when element values are designed based on characteristic values.
The operation of the embodiment of the present invention will be described in detail below with reference to FIGS. 1 to 6. FIG.

First, the user operates the operation unit 102 to input the circuit configuration of the circuit to be designed, the circuit characteristic item names of the circuit to be designed, and the names of circuit elements constituting the circuit to be designed (see FIG. 2). processing step S201). At this time, the operation unit 102 selects and inputs the circuit configuration of the circuit to be designed from a plurality of circuit configurations included in the circuit configuration data 111 of the storage unit 104 , and selects and inputs the circuit configuration included in the circuit characteristic item name data 113 . A circuit characteristic item name of the circuit to be designed is selected from a plurality of circuit characteristic item names and input, and a circuit element name of the circuit to be designed is selected from a plurality of circuit element names included in the circuit element name data 114. Select and enter.

Here, the operation unit 102 and the storage unit 104 constitute an input unit. The input unit can be configured to include an input interface such as a USB (Universal Serial Bus) port, and can be configured to input various data using this.
Note that the circuit configuration of the circuit to be designed may be generated using data necessary for creating the circuit configuration, such as a netlist and constraint data, stored in the storage unit 104 in advance. .

In the processing target input step of processing step S201, the operational amplifier in FIG. The 13 circuit characteristic item names described above are input as the circuit characteristic item names of the circuit to be designed 301, and the element names of the MOS transistors M1 to M8 and one resistor R (names of nine elements) are input. be.

The simulation unit 105 simulates the circuit under design 301 while changing the element values of the circuit elements (the MOS transistors M1 to M8 and the resistor R) that constitute the circuit under design 301, thereby obtaining each element value and its corresponding value. Learning data consisting of characteristic values of each circuit characteristic item is generated (step S202).

The simulation unit 105 performs simulation while changing the characteristics of the transistors M1 to M8 by changing the variable M using the probability distribution of FIG. At this time, the transistors M1 and M2 are changed to the same value, and the transistors M3 and M4 are also changed to the same value. The simulation is performed by changing the resistance value of the resistor R to a plurality of predetermined values.

(1) When performing a simulation, the simulation unit 105 generates new element values of the transistors M1 to M8 by multiplying a predetermined initial value by the change rate of the variable M, and also sets the resistance value R to a predetermined value. A new resistance element value is generated by changing the initial value by a predetermined amount, and a circuit configuration for simulation of the circuit under design 301 is generated. By performing the above processing while changing the element value of each circuit element, a predetermined number of circuit configurations for simulation of the design target circuit 301 are created.

(2) The simulation unit 105 executes a simulation for each of the predetermined number of circuit configurations created, evaluates each circuit configuration based on a predetermined evaluation formula, and determines the element values of the circuit configurations that satisfy predetermined criteria and Save the corresponding characteristic values as training data.
(3) The simulation unit 105 updates the highest evaluation element value in the generated circuit configuration as a new initial value.

The simulation unit 105 repeatedly executes the simulation processes (1) to (3) to generate learning data.
The learning data generated as described above is stored as learning data 115 in the storage unit 104 . As a result, learning data 115 consisting of the element value of each circuit element and the characteristic value of each circuit characteristic item corresponding thereto is collected.

Thus, the gate width W and the gate length L of each of the transistors M1 to M8 are fixed, and by changing the value of the variable M, the element values of the transistors M1 to M8 are changed, thereby changing the simulation conditions. Therefore, learning data can be efficiently collected. Further, learning data can be efficiently collected by changing the characteristics of the transistors M1 to M8 by changing the variable M using the probability distribution of the double peak characteristics.

Next, when the design processing unit 106 is set to the learning mode by operating the operation unit 102, the design processing unit 106 learns the collected learning data 115 by deep learning (step S203).
At this time, when element values of circuit elements corresponding to the characteristic values of the circuit characteristic item are generated in the later design mode, the characteristic values of the circuit characteristic item are input in the learning mode, and the characteristic values of the circuit characteristic item are input in the learning mode. An instruction is input from the operation unit 102 to perform deep learning by outputting the element value of the circuit element corresponding to .

When generating characteristic values of circuit characteristic items corresponding to the element values of the circuit elements in the design mode, the element values of the circuit elements are input in the learning mode, and the circuit characteristics corresponding to the element values of the circuit elements are generated in the learning mode. An instruction is input from the operation unit 102 to perform deep learning with the characteristic value of the item as an output.

The non-dimensionalizing unit 107 non-dimensionalizes the element values and the corresponding characteristic values using the learning data 115 composed of the element values and the corresponding characteristic values, and inputs them to the generation processing unit 108 . As described above, the dimensionless unit 107 normalizes the characteristic values and standardizes the element values to make the dimensionless.

The generation processing unit 108 performs deep learning using the dimensionless element values and the corresponding characteristic values from the dimensionless unit 107 .
At this time, when the element value corresponding to the characteristic value is generated in the design mode, the generation processing unit 108 receives the characteristic value in the learning mode and outputs the element value corresponding to the characteristic value. Do deep learning. Further, when the characteristic value corresponding to the element value is generated in the design mode, the generation processing unit 108 receives the element value in the learning mode and outputs the characteristic value corresponding to the element value. do the learning.

Next, when designing the circuit under design 301, the operation unit 102 sets the design processing unit 106 to the design mode, and the characteristic values of the new circuit characteristic items of the circuit under design 301 stored in the storage unit 104, Alternatively, the element value of the new circuit element is input to the design processing unit 106, and circuit design processing is performed as follows (step S204).

Here, when generating element values of circuit elements corresponding to the characteristic values of the circuit characteristic items, new characteristic values of the circuit characteristic items are input to the input unit of the design processing unit 106, and the element values corresponding to the circuit element values are generated. When generating characteristic values of circuit characteristics, new element values of circuit elements are input.
The dimensionless unit 107 renders the input characteristic value or element value dimensionless and inputs the dimensionless value to the generation processing unit 108 .

When generating element values corresponding to characteristic values in the design mode, the generation processing unit 108 generates element values corresponding to the characteristic values input from the dimensionless unit 107 and inputs them to the restoration unit 109 . The restoration unit 109 restores the dimensionless element value from the generation processing unit 108 to the original unit value and outputs it to the display unit 103 .
The display unit 103 displays the element value of each circuit element corresponding to the characteristic value of each circuit characteristic item of the circuit to be designed. Further, the element value of each circuit element corresponding to the characteristic value of each circuit characteristic item of the circuit to be designed is stored in the data area 110 of the storage unit 104 .

When generating characteristic values corresponding to element values in the design mode, the generation processing unit 108 generates characteristic values corresponding to the element values input from the dimensionless unit 107 and inputs them to the restoration unit 109. . The restoration unit 109 restores the dimensionless characteristic value from the generation processing unit 108 to the original unit value and outputs it to the display unit 103 .
The display unit 103 displays the characteristic values of each circuit characteristic item corresponding to the element value of each circuit element of the circuit to be designed. Further, the characteristic value of each circuit characteristic item corresponding to the element value of each circuit element of the circuit to be designed is stored in the data area 110 of the storage unit 104 .

In the present embodiment, instructions for simulation, learning mode and design mode, data input, etc. are performed by instructions from the operation unit 102. However, a series of processes from simulation to completion of the design mode can be performed automatically. It is possible to make changes such as In either case, the design processing unit 106 performs deep learning using the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items in the learning mode, and performs deep learning on the characteristic values of the circuit characteristic items in the design mode. At least one of the element value of the corresponding circuit element and the characteristic value of the circuit characteristic item corresponding to the element value of the circuit element is automatically generated by regression analysis.

By the way, in the embodiment of the present invention, the generation processing unit 108 normalizes the characteristic values by the dimensionless unit 107 and normalizes the element values, so that the element values and the characteristic values are obtained by using the dimensionless values. It is configured to perform the generation process of
FIG. 6 shows the error rate of the element value (right column) obtained when the characteristic value is normalized and the element value is standardized and the design process is performed as in this embodiment, and both the characteristic value and the element value is a chart comparing the error rates (middle column) of the element values obtained when design processing is performed by normalizing by the above formula. As is clear from FIG. 6, it can be seen that the precision is improved by the dimensionless processing in the embodiment of the present invention.

It is presumed that this is due to the following reasons.
That is, normalization is a calculation for judging whether data groups with different units (with different size ranges) are good or bad on the same scale. To normalize the circuit characteristics,
(1) If data with a large value indicates good performance, set the data with the largest value to 1 and the data with the smallest value to 0.
(2) If data with a small value indicates good performance, set the data with the smallest value to 1 and the data with the largest value to 0.
Normalize each. This makes it possible to evaluate circuit characteristics.

On the other hand, in the case of element values, if the element values are normalized by defining them as in (1) as described above, it will be judged that the performance is good when the values are large. However, in the case of element values, larger values are not necessarily better. Therefore, it is not possible to judge whether it is good or bad if it is made dimensionless by normalization.

In addition, standardization is suitable for dimensionless element values because it uses the difference from the reference value to make dimensionless and does not judge whether it is good or bad. Therefore, in the embodiment of the present invention, element values are made dimensionless using normalization.
In addition, the standardization used in neural networks is generally processed so that the input data has an average of 0 and a variance of 1, and the data is generally limited to a range of -1 to 1, but this time (learning The value of the data)/(initial value of the learning data) yielded good results, so this is used.

FIG. 7 is a diagram showing the configuration of the generation processing unit 108 corresponding to FIG. 5, and is an example of a case where it functions to design characteristic values based on element values.
In the learning mode, the generation processing unit 108 updates the weight w and the bias b so that the element values obtained by the simulation are input and the corresponding characteristic values are output. In the design mode after learning, when the element value of each circuit element is input, it functions to generate and output the corresponding characteristic value. Other processes such as dimensionless processing and restoration processing are the same as those described above.

FIG. 8 is a chart showing the average matching rate of characteristic values generated based on the element values by causing the generation processing unit 108 to function as shown in FIG.
When calculating the average matching rate, (1) the characteristic value A of the test data not used for learning is input to the design processing unit 106 to generate the corresponding element value, and (2) the element value is subjected to design processing. Input to the unit 106 to generate the corresponding characteristic value B, and (3) calculate the matching rate of the characteristic value B with respect to the characteristic value A.

The above (1) to (3) are performed for a predetermined number of test data, and the average value of the obtained match rates is taken as the average match rate. Good results have been achieved with an average accuracy of 90% or more.
By using the learned model, the next design process does not require the learning time, so the design process can be performed in only the time required to generate the desired characteristic values and element values.

FIG. 9 is a circuit configuration diagram showing an example of another design object circuit 901, showing an example of an operational amplifier whose input section is configured by NMOS transistors. In FIG. 9, transistors M1 and M2 are NMOS transistors with the same characteristics, and transistors M3 and M4 are PMOS transistors with the same characteristics. It is composed substantially of six MOS transistors M1 to M8 with different characteristics and one resistor R. FIG.

FIG. 10 is a chart showing the average matching rate of the characteristic values obtained in the same manner as described above by causing the generation processing unit 108 to function as shown in FIG. Also in this example, an average accuracy of 90% or more is realized, and good results are obtained.

As described above, according to the embodiment of the present invention, the input unit (operation unit 102, storage unit 104);
a simulation unit 105 that generates the element values of the circuit elements and the characteristic values of the circuit characteristic items corresponding thereto by simulating the circuit under design while changing the element values of the circuit elements;
In a learning mode, deep learning is performed using the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items, and in the design mode, the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic items, and a design processing unit 106 for generating at least one characteristic value of the circuit characteristic item corresponding to the element value of the circuit element by regression analysis.

Here, when generating the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic item in the design mode, the design processing unit 106 inputs the characteristic values of the circuit characteristic item in the learning mode. Deep learning is performed so that the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic items are output, and in the design mode, the characteristic values of the circuit characteristic items are input and the characteristic values of the circuit characteristic items are obtained. It may be configured to generate an element value for the corresponding circuit element.

Further, when generating the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements in the design mode, the design processing unit 106 inputs the element values of the circuit elements in the learning mode and Deep learning is performed with the characteristic values of the circuit characteristic items corresponding to the element values of the elements as outputs, and in the design mode, the circuit corresponding to the element values of the circuit elements with the element values of the circuit elements as inputs. It can be configured to generate characteristic values for characteristic items.

Therefore, it is possible to easily design a circuit in a short time without depending on the designer's knowledge and subjectivity.
In addition, a circuit design apparatus capable of easily designing a circuit in a short time without depending on the knowledge and subjectivity of the designer by having the computer execute the circuit design program according to the embodiment of the present invention. becomes possible to build.

It can be used for inventions related to circuit design of analog integrated circuits and digital integrated circuits.

DESCRIPTION OF SYMBOLS 100... Circuit design apparatus 101... Data processing part 102... Operation part 103... Display part 104... Storage part 105... Simulation part 106... Design processing part 107... None Dimensionalization unit 108 Generation processing unit 109 Restoration unit 110 Data area 111 Circuit configuration data 112 Circuit item data 113 Circuit characteristic item name data 114 Circuit element name data 115 Learning data 116 Programs 301, 901 Circuits to be designed M1 to M8 MOS transistor R Resistor

Claims (5)

an input unit for inputting a circuit configuration of a circuit under design, a circuit characteristic item name of the circuit under design, and an element name of a circuit element constituting the circuit under design;
a simulation unit that generates the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items by simulating the circuit under design while changing the element values of the circuit elements;
In a learning mode, deep learning is performed using the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items, and in the design mode, the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic items, and a design processing unit that generates at least one of the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements by regression analysis. The design processing unit
When generating the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic item in the design mode, in the learning mode, the characteristic values of the circuit characteristic item are input and correspond to the characteristic values of the circuit characteristic item. Deep learning is performed so that the element value of the circuit element to be output is the element value of the circuit element, and in the design mode, the element value of the circuit element corresponding to the characteristic value of the circuit characteristic item is input with the characteristic value of the circuit characteristic item. generate and
When the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements are generated in the design mode, the element values of the circuit elements are input in the learning mode, and the element values of the circuit elements are generated in the learning mode. Deep learning is performed with the characteristic value of the circuit characteristic item as an output, and in the design mode, the element value of the circuit element is used as an input to generate the characteristic value of the circuit characteristic item corresponding to the element value of the circuit element. 2. The circuit design apparatus according to claim 1, wherein: A circuit design method executed by a circuit design device comprising an input unit, a simulation unit, and a design processing unit,
the input unit inputs a circuit configuration of the circuit under design, a circuit characteristic item name of the circuit under design, and an element name of a circuit element constituting the circuit under design;
The simulation unit generates the element values of the circuit elements and the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements by simulating the circuit under design while changing the element values of the circuit elements;
The design processing unit performs deep learning using the element values of the circuit elements and the corresponding characteristic values of the circuit characteristic items in the learning mode, and the circuit corresponding to the characteristic values of the circuit characteristic items in the design mode. A circuit design method, wherein at least one of an element value of an element and a characteristic value of the circuit characteristic item corresponding to the element value of the circuit element is generated by regression analysis. The design processing unit
When generating the element values of the circuit elements corresponding to the characteristic values of the circuit characteristic item in the design mode, in the learning mode, the characteristic values of the circuit characteristic item are input and correspond to the characteristic values of the circuit characteristic item. Deep learning is performed so that the element value of the circuit element to be output is the element value of the circuit element, and in the design mode, the element value of the circuit element corresponding to the characteristic value of the circuit characteristic item is input with the characteristic value of the circuit characteristic item. generate and
When the characteristic values of the circuit characteristic items corresponding to the element values of the circuit elements are generated in the design mode, the element values of the circuit elements are input in the learning mode, and the element values of the circuit elements are generated in the learning mode. Deep learning is performed with the characteristic value of the circuit characteristic item as an output, and in the design mode, the element value of the circuit element is used as an input to generate the characteristic value of the circuit characteristic item corresponding to the element value of the circuit element. 4. The circuit design method according to claim 3, wherein: A circuit design program for causing a computer to construct the circuit design apparatus according to claim 1 or 2.

Images