Chip module automatic import method

Abstract

This invention discloses an automatic chip module import method. The automatic import program is designed to automatically import chip modules from a database into the required project product. After selecting the closest reference chip, the automatic import program matches the parameter information of the chip modules in the database, and through analysis and calculation, filters the chip modules that are closest to the reference chip and imports them into the required project product. The automatic import program helps designers to filter the chip modules that are closest to the existing mature project product (reference chip) and import them into the required project product. In this way, designers can reduce most of the layout and routing work time in chip design and save project time.

Description

Technical Field

[0001] This invention relates to the field of PCB design technology, and more specifically, to a method for automatically importing chip modules. Background Technology

[0002] From a design perspective, chips can be categorized into general-purpose chips and application-specific chips (ASICs). General-purpose chips typically refer to chips not designed for specific computations and not limited to a particular product or application scenario. The key to general-purpose chips lies in their versatility and flexibility. A general-purpose chip is formed by defining the fixed components of its design concept. Therefore, commonly used CPUs (Central Processing Units), GPUs (Graphics Processing Units), memory chips, and DPUs (Distributed Processing Units) are all common types of general-purpose chips. When these chips of the same type are placed in a circuit, the differences between them and surrounding circuits are minimal; some general-purpose chips differ only in their resistance and capacitance values. Consequently, general-purpose chips of the same type exhibit a very high degree of similarity.

[0003] In this situation, when designing project products, designers often find similar existing project products based on the existing project requirements, reference the common chips of the existing project products, manually import them into the existing project products, and then make the layout and wiring around the common chips the same or similar. This design method is feasible when the number of project products is small, but when the number of products is large, this manual import method will still cause a large workload for designers.

[0004] Moreover, this design approach does not mean that general-purpose chips can be completely replaced. Designers still need to consider the chip's location on the printed circuit board, its network connections, and other attributes when selecting a chip. In addition, the printed circuit board descriptions of the same general-purpose chip may differ in different project products. This means that in finished project products, general-purpose chips with the same description may not actually be the same or similar types of chips, further increasing the workload for designers. Summary of the Invention

[0005] To reduce the workload of designers, this invention provides a method for automatically importing chip modules.

[0006] The technical solution of this invention is as follows:

[0007] A method for automatically importing chip modules includes establishing an automatic import program, associating the program with PCB design software, and connecting the program to a server storing chip modules and reference chips. The usage and operation flow of the automatic import program includes:

[0008] Step S1. Select the reference chip;

[0009] Step S2. Retrieve chip modules of the same type as the benchmark chip;

[0010] Step S3. Automatically analyze and filter the chip modules that are most similar to the benchmark chip from the search results as the target chip modules;

[0011] Step S4. Import the target chip module obtained in step S3 into the current PCB file, and import the layout and routing of the target chip module.

[0012] The above-mentioned automatic chip module import method includes step S3, which includes...

[0013] Step T1. Extract all parameter information from both the individual chip modules and the reference chip;

[0014] Step T2. Compare the parameter information of the chip module with the parameter information of the reference chip to obtain the similarity between the chip module and the reference chip on the same parameter information;

[0015] Step T3. Obtain the similarity value between the chip module and the benchmark chip through preset weighting coefficients, and select the chip module with the highest similarity value as the target chip module to be imported.

[0016] Furthermore, in step T1, all parameter information is stored in text format.

[0017] Step T1 indicates that the automatic import program of this invention analyzes and filters each chip module entirely based on the characters, i.e., text information, of each parameter information. After extracting the parameter information, a corresponding text library is established. During the analysis process of the automatic import program, the text data of the parameter information are unrelated, and each character exists independently. Different splitting and combination can be achieved according to preset analysis templates or patterns.

[0018] Furthermore, the similarity between the chip module and the benchmark chip for the same parameter information is determined using the cosine similarity principle. The cosine similarity calculation formula is as follows:

[0019]

[0020] Among them, X i Y is the calculated value corresponding to the various parameter information of the reference chip. i This refers to the calculated value corresponding to the same parameter information of the chip module.

[0021] Furthermore, the automatic import program sets up word segmentation. In step T2, the frequency of each word segmentation in each parameter information is calculated to form a word frequency vector about the parameter information. This vector is then substituted into the cosine similarity calculation formula to calculate the similarity, where X... i Y represents the word frequency vectors corresponding to the various parameter information of the benchmark chip. iThis refers to the word frequency vector corresponding to the same parameter information of the chip module.

[0022] One method for determining the similarity between each chip module and a benchmark chip is based on the cosine similarity principle. In step T1, the characters in each parameter information are split, and according to a preset word segmentation pattern, the characters in the parameter information are combined and matched to determine the number of preset words contained in the parameter information, thus obtaining the frequency of the segmented words. Judging the similarity between a chip module and a benchmark chip on the same parameter information based on the word frequency of the segmented words is based on considering the parameter attribute names possessed by the parameter information. Modules with the same or similar parameter attributes use relatively similar nouns, so the similarity between the two can be judged by the number of nouns with the same parameter attributes. At this point, the word segmentation is equal to the parameter name of a certain attribute.

[0023] Furthermore, the similarity of the chip module and the benchmark chip for the same parameter information is calculated by the proportion of identical character text for that parameter information. The identical character text for the same parameter information between the chip module and the benchmark chip must meet the following conditions:

[0024] Condition 1: Same position; Condition 2: Same value or character.

[0025] When the chip module and the reference chip are dealing with this parameter information, the same character text is the same character located in the same position.

[0026] Similar to parameter information such as type name, according to existing industry rules, two identical or similar attribute names will usually contain a large number of characters with the same format. Then, they are distinguished by local numerical parameters. To improve efficiency and simplify the calculation process, we can consider comparing text with the same characters. This method is usually applied to parameter information such as model and name.

[0027] Furthermore, in step T3, the automatic import program pre-sets weight coefficients corresponding to each parameter information. The similarity value of the chip module relative to the reference chip is equal to the sum of the products of the similarity of each parameter information relative to the reference chip and the corresponding weight coefficient.

[0028] Furthermore, the weighting coefficients are bound to the type of the reference chip. The automatic import program has multiple sets of weighting coefficients built in. After the type of the reference chip is determined in step S1, the automatic import program selects any set containing multiple weighting coefficients according to the type of the reference chip and preset conditions.

[0029] In the above-mentioned automatic chip module import method, in step S3, the chip module and the reference chip are subdivided to form multiple components, the parameter information of each component is obtained, the similarity between the components of each chip module and the components of the reference chip is compared and calculated, and the components with the highest similarity are combined to form a temporary target chip module that is most similar to the reference chip.

[0030] When setting up the server, the imported data is packaged and saved as a unit of chip module. A chip module also includes multiple component contents. When importing and packaging, it can be further split. When comparing chip modules, it is easy to select one component at a time. Finally, chip modules of the same type are combined to form a new chip module, which is then packaged and saved.

[0031] In the above-mentioned method for automatically importing chip modules, in step S4, the automatic import program sets a similarity threshold. The automatic import program imports the target chip module if and only if the similarity of the target chip module obtained in step S3 is higher than the similarity threshold. If the similarity of the target chip module is lower than the similarity threshold, the automatic import program provides a pop-up prompt or a report prompt.

[0032] According to the above-described solution, the beneficial effects of this invention are as follows: This invention provides an automatic chip module import method. An automatic import program is designed to automatically import chip modules from a database into the required project product. After selecting the closest reference chip, the automatic import program matches the parameter information of the chip modules in the database, and through analysis and calculation, filters the chip modules closest to the reference chip, and imports them into the required project product. The automatic import program helps designers filter the chip modules closest to existing mature project products (reference chips) and import them into the required project product. In this way, designers can reduce most of the layout and routing work in chip design, saving project time. Detailed Implementation

[0033] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0034] A method for automatically importing chip modules involves creating an automatic import program and uploading it to PCB design software, enabling the software to reference the program. The automatic import program is associated with a database storing chip modules and a reference chip. In this embodiment, the chip module database is a self-built server. The automatic import program selects an associated server through its options, allowing it to access the corresponding server's database, retrieve saved chip module parameters, analyze, calculate, and filter out the target chip module that best matches the reference chip, and then import this target chip module into the current PCB design file.

[0035] Step S1. Select the reference chip.

[0036] A specific chip is selected as the reference chip, and data information from that reference chip is obtained.

[0037] Step S2. Retrieve a chip module of the same type as the reference chip.

[0038] In step S2, the retrieval process is the search process. The automatic import program searches all chip modules through the selected database and then extracts the data information of each chip module. This step is essentially the process of obtaining the data information of the chip modules.

[0039] In step S2, preliminary screening can be performed. In this embodiment, preliminary screening is performed based on the type of chip module. The chip module type is typically categorized by its purpose, such as computing chips, storage chips, sensing chips, and power chips. When a chip module is imported into the self-built server, a relevant attribute card is created, and all relevant information and data for that chip module are packaged and saved. During the preliminary screening, the type of the baseline chip is used as the matching condition to filter chip modules of the same type.

[0040] Step S3. Automatically analyze and filter the most similar chip modules from the search results as the target chip modules.

[0041] Step T1 extracts the character content from the parameter information of each chip module and the reference chip.

[0042] In step T1, the program extracts various parameter information from the chip module. This parameter information is associated with the chip module name and collectively packaged and stored in the database. For the database, each parameter information of the chip module is equivalent to a stored module / set. In this embodiment, the parameter information of the chip module includes five items: Reference, PackageSymbol, Device Type, Value, and Pin IO Information. Other parameter information can be added to meet other requirements, thereby further filtering to obtain more suitable chip modules.

[0043] After extracting the information, the character content of the parameter information in the chip module is obtained. The character content of the parameter information obtained here is in text format, which simplifies the comparison and calculation of the program and speeds up the operation.

[0044] Step T2. Compare the parameter information of the chip module with the parameter information of the reference chip to obtain the similarity between the chip module and the reference chip on the same parameter information.

[0045] Obtain the character content of the same parameter information of the chip module and the benchmark chip in text format, and compare the two to obtain the character similarity between the chip module and the benchmark chip in that parameter information.

[0046] Example 1:

[0047] In this embodiment, the cosine similarity principle is used to determine the similarity between the two.

[0048] The cosine of the angle between two vectors in a vector space is used as a measure of the difference between the two individuals. The closer the value is to 1, the closer the angle is to 0°, meaning the two vectors are more similar. This is called cosine similarity, and it is the basis of the cosine similarity principle. The specific calculation formula is:

[0049]

[0050] Among them, X i Y is the calculated value corresponding to the various parameter information of the reference chip. iThis refers to the calculated value corresponding to the same parameter information of the chip module. In this embodiment, the character text extracted from the parameter information is segmented and its frequency is calculated to form a word frequency vector, and the cosine similarity of the word frequency vector is calculated. Specifically, the automatic import program sets common chip terms or expressions as word segments, sets a word segmentation list, and then expands the text of the parameter information value. The word segmentation content is matched cyclically for the text of the parameter information value, or the text of the parameter information value is divided into fields and matched cyclically with the word segmentation content. Finally, the expanded text of the parameter information value is matched with word segments one by one. The remaining text can be combined and matched according to computer language. Then, the frequency of each word in the text of the parameter information value is marked according to the order of the word segmentation list. In this way, vectors composed of the frequency of each word in the text of the parameter information value corresponding to the benchmark chip and the chip module can be obtained respectively. Then, the cosine similarity between the two vectors is calculated. In this way, the similarity between the chip module and the benchmark chip for this parameter information is obtained.

[0051] Example 2:

[0052] In this embodiment, the similarity between the reference chip and the chip module for the same parameter information is obtained by iteratively matching the text of the reference chip and the chip module. Specifically, according to the text position order, the first text character of the reference chip's parameter information value is compared with the first text character of the chip module's same parameter information value, the second text character of the reference chip's parameter information value is compared with the second text character of the chip module's same parameter information value, the third text character of the reference chip's parameter information value is compared with the third text character of the chip module's same parameter information value, and so on. The number of text characters that are the same between the two is divided by the total number of text characters of the reference chip's parameter information value to obtain the similarity between the chip module and the reference chip for that parameter information.

[0053] In addition to the two embodiments described above, other calculation methods such as matrix principles and second-order matrices can be used to calculate the similarity between the values ​​of the chip module parameter information and the values ​​of the reference chip parameter information.

[0054] Step T3. Obtain the similarity value between each chip module and the benchmark chip using preset weighting coefficients, and select the chip module with the highest similarity value as the target chip module to be imported.

[0055] The automatic import program sets the weight coefficients corresponding to each parameter information in advance according to the requirements. In step T2, the similarity of the chip module to the reference chip in each parameter information is obtained. The similarity value is calculated as similarity value × corresponding weight coefficient. Then, the sum of the similarity values ​​of all parameter information is calculated. The chip module with the largest similarity value is the closest and most similar to the reference chip, and is selected as the target chip module for importing the current layout and routing.

[0056] Example 3:

[0057] In this embodiment, the parameter information of the reference chip includes Reference, Package Symbol, Device Type, Value, and Pin IO Information. Weighting coefficients for Reference, Package Symbol, Device Type, Value, and Pin IO Information are preset, with each of these five parameters having a weighting coefficient of 20%. The Pin IO Information also includes pin_Number, pin_Type, pin_SigNoise Model, and pin_Net, which are also part of the parameter information. Similarity calculation is performed in step T2, and the weight ratio of these four is the average of the four weighting coefficients of the Pin IO Information. After obtaining the similarity information of the seven parameters (Reference, Package Symbol, Device Type, Value, Pin Number, Pin Type, Pin Noise Model, and Pin Net) in chip module A to the reference chip [Similarity of Reference (Reference Number) is A1, Package Symbol (Package Name) is A2, Device Type (Package Type) is A3, Value (Package Model) is A4, Pin Number (Pin Number) is A5, Pin Type (Pin Type) is A6, Pin Noise Model (Pin Noise Model) is A7, and Pin Net (Pin Net) is A8], the...

[0058] Chip module similarity value

[0059] =A1×20%+A2×20%+A3×20%+A4×20%+A5×20%×25%+A6×20%×25%+A7×20%×25%+A8×20%×25%.

[0060] That is, the sum of the product of the similarity of each parameter information and its corresponding weight coefficient, and the one with the largest similarity value is regarded as the closest to the benchmark chip.

[0061] In addition, since the weight coefficients are preset content, they can be bound to the type of the reference chip. After obtaining the type of the reference chip in step S1, the weight coefficients of its corresponding parameter information can be obtained. The automatic import tool packages and saves the weight coefficients according to different needs, and selects and uses them according to the type of the reference chip.

[0062] Step S4. Import the target chip modules obtained in step S3 into the current PCB design file, and import the layout and routing according to the calculation method.

[0063] In step T3, the target chip module that is closest to the reference chip in the server is obtained. Before step S4, an import threshold can be set, that is, only chip modules whose similarity value is greater than the import threshold can be imported into the existing layout and routing. If the target chip module in the server is significantly different from the reference chip, it is not suitable to be imported into the existing layout and routing. This is because a lot of layout and routing content needs to be modified later. Setting an import threshold achieves the minimum similarity of the target chip module to be imported.

[0064] In step S4, the automatic import tool uses the `exportModule / getRouting` function to obtain the placement and routing information of the target chip module and directly imports the target chip module into the existing placement and routing. Alternatively, the automatic import tool uses the `doPlacement / doRouting` function to import the placement and routing of the target chip module. During the import process, the placement and routing information of the target chip module is first compared with that of the reference chip. Any discrepancies are excluded from the import, leaving blank areas in the existing placement and routing. Then, based on the blank areas, the reference chip's corresponding portions are selected from the server based on the highest similarity values. This similarity can also be filtered using information similarity, i.e., from overall similarity to local similarity. The discrepancies in the target chip module here do not mean completely identical images, but rather differences in values ​​and information such as network connections, port connections, and number of lines, since lines can be extended.

[0065] Preferably, in steps T5 and T4, not only can similarity matching be performed on the parameter information of the chip module / package, but further subdivision and filtering can be performed on the internal structure, layout, and routing of the chip module. For example, the chip module can be split into multiple components, i.e., the module can be further subdivided into modules, each component possessing various parameter information. Components can be categorized by function, by interconnection network, by location, by connection point, etc. This further subdivides the content of the reference chip and the chip module. Using the same filtering principle, similarity calculations are performed on the refined components, ultimately forming a temporary target chip module with the highest similarity value between each component and the reference chip. This temporary target chip module is then imported into the existing layout and routing. Minor local differences may occur at this point, which can be corrected through fitting or manual modification. The components of this temporary target chip module can originate from different chip modules; it is not necessary to consider whether the chip modules are from the same product during the filtering process.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A chip module automatic import method characterized by, An automatic import program is established, which is associated with the PCB design software and connected to a server storing chip modules and reference chips. The usage and operation flow of the automatic import program includes: Step S1. Select the reference chip; Step S2. Retrieve chip modules of the same type as the benchmark chip; Step S3. Automatically analyze and filter the chip modules that are most similar to the benchmark chip from the search results as the target chip modules; Step S3 includes Step T1: Extract all parameter information from both the individual chip modules and the reference chip; Step T2. Compare the parameter information of the chip module with the parameter information of the reference chip to obtain the similarity between the chip module and the reference chip on the same parameter information; Step T3. Obtain the similarity value between the chip module and the benchmark chip through preset weighting coefficients, and select the chip module with the highest similarity value as the target chip module to be imported. In step S3, the subdivided chip module and the reference chip form multiple components, the parameter information of each component is obtained, the similarity between the components of each chip module and the components of the reference chip is compared and calculated, and the components with the highest similarity are used to form a temporary target chip module that is most similar to the reference chip. Step S4. Import the target chip module obtained in step S3 into the current PCB file, and import the layout and routing of the target chip module. In step S4, the automatic import program sets a similarity threshold. The automatic import program imports the target chip module if and only if the similarity of the target chip module obtained in step S3 is higher than the similarity threshold. If the similarity of the target chip module is lower than the similarity threshold, the automatic import program will display a pop-up prompt or report prompt. In step S4, the automatic import program uses the exportModule / getRouting function to obtain the relevant information of the target chip module's layout and routing, and directly imports the target chip module's layout and routing into the existing layout and routing. Alternatively, the automatic import program imports the target chip module's layout and routing through the doPlacement / doRouting function. During the import process, the layout and routing information of the target chip module is first compared with the layout and routing information of the reference chip. If there are discrepancies, the layout and routing of that part is not imported, resulting in blank areas in the existing layout and routing. Then, based on the part of the reference chip corresponding to the blank area, the part with the highest similarity value is selected in the server.

2. The method for automatically importing chip modules according to claim 1, characterized in that, In step T1, all parameter information is stored in text format.

3. The method for automatically importing chip modules according to claim 1, characterized in that, The cosine similarity principle is used to determine the similarity between a chip module and a benchmark chip for the same parameter information. The cosine similarity calculation formula is as follows: Wherein, X i is the calculated value corresponding to each parameter information of the reference chip, Y i is the calculated value corresponding to the same parameter information of the chip module.

4. The method for automatically importing chip modules according to claim 3, characterized in that, The automatic import program sets word segmentation, calculates the frequency of each word in each parameter information in step T2, forms a word frequency vector about the parameter information and substitutes it into the cosine similarity calculation formula to calculate the similarity, wherein, X i is the word frequency vector corresponding to each parameter information of the reference chip, i Y is the word frequency vector corresponding to the same parameter information of the chip module.

5. The method for automatically importing chip modules according to claim 1, characterized in that, The similarity of the parameter information is determined by the proportion of identical character text between the chip module and the benchmark chip for the same parameter information. For identical character text between the chip module and the benchmark chip for the same parameter information, the following conditions must be met: Condition 1: Same position; Condition 2: Same value.

6. The method for automatically importing chip modules according to claim 1, characterized in that, In step T3, the automatic import program pre-sets weight coefficients corresponding to each parameter information. The similarity value of the chip module relative to the reference chip is equal to the sum of the products of the similarity of each parameter information relative to the reference chip and the corresponding weight coefficient.

7. The method for automatically importing chip modules according to claim 6, characterized in that, The weighting coefficients are bound to the type of the reference chip. The automatic import program has multiple sets of weighting coefficients built in. After the type of the reference chip is determined in step S1, the automatic import program selects any set containing multiple weighting coefficients according to the type of the reference chip and preset conditions.

Images