Method, apparatus, electronic device, and readable storage medium for trim test

Abstract

The disclosure provides a kind of trimming test method, device, electronic equipment and readable storage medium, it is related to semiconductor testing technical field.Therein, the method for trimming test based on classification model includes: obtaining the historical trimming test data of integrated circuit chip, and the historical trimming test data is configured as classification training sample set;Using classification algorithm, the model training of classification training sample set is carried out, and the classification trimming model is obtained;Trimming file generated based on the classification trimming model is sent to test equipment, to make the test equipment based on the trimming file the circuit parameter of the trimming circuit of target chip is trimmed, and the target chip after trimming is tested.Through the technical scheme of the present disclosure, massive Trim real measurement data can be effectively and reliably utilized, and the target chip to be tested is trimmed based on the generated classification trimming model for trimming test, which is beneficial to simplify the testing process, improve the testing efficiency and reduce the testing cost.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of semiconductor testing, and particularly relates to a method for trim test based on a classification model, an apparatus for trim test based on a classification model, an electronic device and a computer readable storage medium. BACKGROUND

[0002] Trim trim refers to a process of trimming some parameters in a circuit in chip testing, including reference voltage, bias current, band gap voltage and / or oscillator frequency, etc., by correcting the above parameters, so that the parameters are close to the target ideal target value,

[0003] In the related art, for DRAM (Dynamic Random Access Memory), the Trim circuit usually adopts the processing mode of one-time programmable memory efuse, and in the measurement process by ATE (Automatic Test Equipment), the same chip independently performs Trim. Based on the above processing mode, if it is necessary to repeatedly Trim the chip, the ATE needs to reconfigure the chip and measure, and the process is repeated in turn, resulting in a long Trim process, poor efficiency and high test cost.

[0004] In addition, directly discarding a large amount of Trim measured data generated in mass production testing also causes data waste to some extent.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present disclosure is to provide a trim test method, apparatus, electronic device and readable storage medium to simplify the trim test process of the chip, improve the test efficiency and reduce the test cost.

[0007] Other characteristics and advantages of the present disclosure will become apparent from the following detailed description, or will be learned by practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a method for performing trim test based on a classification model is provided, which is applied to a data processing device and includes: obtaining historical trim test data of an integrated circuit chip, and configuring the historical trim test data as a classification training sample set; performing model training on the classification training sample set using a classification algorithm to obtain a classification trim model; and sending a trim file generated based on the classification trim model to a test device, so that the test device performs trim on circuit parameters of a trim circuit of a target chip based on the trim file, and tests the target chip after trim.

[0009] In an example embodiment of the present disclosure, the configuring the historical trim test data as a classification training sample set includes: dividing a wafer corresponding to the historical trim test data into a plurality of regions; determining a home region based on position information of a die of the integrated circuit chip on the wafer; and classifying the historical trim test data based on the home region, and determining a set after classification as the classification training sample set.

[0010] In an example embodiment of the present disclosure, the dividing the wafer corresponding to the historical trim test data into a plurality of regions includes: configuring a number of partitions and a partitioning manner for partitioning the wafer based on process parameters of the wafer; and dividing the plurality of regions based on the number of partitions and the partitioning manner, wherein the plurality of regions include a central region and a plurality of annular regions outside the central region.

[0011] In an example embodiment of the present disclosure, the classification algorithm includes at least one of a random forest, a support vector machine, a K-nearest neighbor algorithm, and a decision tree classification algorithm.

[0012] In an example embodiment of the present disclosure, the classification algorithm includes the K-nearest neighbor algorithm, and the performing model training on the classification training sample set using a classification algorithm to obtain a classification trim model includes: generating a first feature vector based on first position information, a first test value, a first target value, and a first parameter of a to-be-tested sample; generating a plurality of second feature vectors based on second position information, a second test value, a second target value, and the second parameter of a plurality of training samples in the classification training sample set; respectively calculating distances between the first feature vector and the plurality of second feature vectors; selecting K training samples closest to the to-be-tested sample based on the calculation results, to determine a home region of the first position information based on home regions of K second position information, K being a number of the plurality of regions; and performing model optimization based on a relationship between the home region of the first position information and an actual home region of the to-be-tested sample, to generate the classification trim model based on a model optimization result.

[0013] In an example embodiment of the present disclosure, the sending of the trimming file generated based on the classification trimming model to the test device comprises: obtaining a test initial value of the target chip; taking the test initial value as a test sample of the classification trimming model, testing the test sample based on the classification trimming model, obtaining a corresponding trimming parameter, and sending the trimming parameter to the test device as the trimming file; or directly generating the trimming file based on the classification trimming model, and sending the trimming file to the test device.

[0014] In an example embodiment of the present disclosure, further comprising: receiving feedback information sent by the test device based on the trimming file; detecting that the feedback information includes indication information of model updating, updating the classification trimming model; or extracting a cycle number of trimming test in the feedback information, and updating the classification trimming model when it is detected that the cycle number is greater than or equal to a cycle threshold.

[0015] In an example embodiment of the present disclosure, the updating of the classification trimming model comprises: obtaining a test result sent by the test device; updating the classification training sample set based on the test result; and updating the classification trimming model based on the updated classification training sample set.

[0016] In an example embodiment of the present disclosure, further comprising: in response to the obtained model adjustment parameter, adjusting the classification trimming model based on the model adjustment parameter, wherein the model adjustment parameter comprises at least one of a process adjustment parameter, a material adjustment parameter, and a device adjustment parameter.

[0017] According to another aspect of the present disclosure, a method for trimming test based on a classification model is provided, applied to a test device, comprising: generating a trimming parameter based on an obtained trimming file, the trimming file being generated based on a classification trimming model; sending the trimming parameter to a target chip to trim a trimming circuit of the target chip, and obtaining a test result of the target chip after trimming.

[0018] In an example embodiment of the present disclosure, further comprising: determining that the trimming parameter does not meet trimming requirements based on the test result, obtaining a revised trimming parameter, and re-performing trimming test on the target chip based on the revised trimming parameter, the revised trimming parameter being generated by revising the trimming parameter based on the classification trimming model; generating feedback information based on the result of the trimming test, and sending the feedback information to a data processing device to perform model adjustment based on the feedback information by the data processing device, wherein the data processing device is configured to send the trimming file to the test device.

[0019] In an example embodiment of the present disclosure, the determining that the trimming parameter does not meet the trimming requirement based on the test result, obtaining a revised trimming parameter, and re-performing the trimming test on the target chip based on the revised trimming parameter include: the test result includes an actual test value of the target chip, and it is detected that a deviation between the actual test value and a reference target value is greater than a deviation threshold, and then it is determined that the trimming parameter does not meet the trimming requirement, and the revised trimming parameter is obtained; the trimming test is performed on the target chip based on the revised trimming parameter, and when the revised trimming parameter still does not meet the trimming requirement, the number of execution times of the trimming test is counted; it is detected that the execution times is less than an allowed number of times, the revised trimming parameter is re-obtained to perform the trimming test on the target chip; it is detected that the execution times is greater than or equal to the allowed number of times, it is determined that a test period is completed, the number of periods of the trimming test is incremented by 1, and the execution times is re-counted when the trimming test is performed based on the revised trimming parameter next time.

[0020] In an example embodiment of the present disclosure, the obtaining the revised trimming parameter includes: determining a difference between the trimming parameter and a revision threshold as a third parameter; determining a sum of the trimming parameter and the revision threshold as a fourth parameter; obtaining the classification trimming model, and determining one of the third parameter and the fourth parameter as the revised trimming parameter based on the classification trimming model.

[0021] In an example embodiment of the present disclosure, the determining one of the third parameter and the fourth parameter as the revised trimming parameter based on the classification trimming model includes: outputting a first model target value corresponding to the third parameter based on the classification trimming model; outputting a second model target value corresponding to the fourth parameter based on the classification trimming model; determining a first deviation value between the first model target value and a reference target value, and a second deviation value between the second model target value and the reference target value; if the first deviation value is less than the second deviation value, determining the third parameter as the revised trimming parameter; if the second deviation value is less than the first deviation value, determining the fourth parameter as the revised trimming parameter.

[0022] In an example embodiment of the present disclosure, the generating the feedback information based on the result of the trimming test comprises: determining, based on the result of the trimming test, that the number of cycles of the trimming test is greater than or equal to a cycle threshold, generating the feedback information comprising indication information of model updating; or sending, based on the result of the trimming test, the number of cycles of the trimming test to the data processing device, and updating the classification trimming model when the data processing device detects that the number of cycles is greater than or equal to the cycle threshold.

[0023] In an example embodiment of the present disclosure, the generating the trimming parameter based on the obtained trimming file comprises: extracting the trimming parameter included in the trimming file; or the trimming file is generated based on the classification trimming model, and the trimming parameter is obtained by testing the initial value of the target chip based on the classification trimming model.

[0024] In an example embodiment of the present disclosure, further comprising: determining, based on the test result, that the trimming parameter meets the trimming requirement, and then generating a corresponding trimming log.

[0025] According to still another aspect of the present disclosure, there is provided a device for trimming test based on a classification model, applied to a data processing device, comprising: a configuration module configured to obtain historical trimming test data of an integrated circuit chip and configure the historical trimming test data as a classification training sample set; a training module configured to use a classification algorithm to train the classification training sample set to obtain a classification trimming model; and a sending module configured to send a trimming file generated based on the classification trimming model to a test device, so that the test device adjusts a circuit parameter of a trimming circuit of a target chip based on the trimming file, and tests the target chip after trimming.

[0026] According to still another aspect of the present disclosure, there is provided a device for trimming test based on a classification model, applied to a test device, comprising: a generating module configured to generate a trimming parameter based on an obtained trimming file, the trimming file being generated based on a classification trimming model; and a testing module configured to send the trimming parameter to a target chip to trim a trimming circuit of the target chip, and obtain a test result after trimming the target chip.

[0027] According to still another aspect of the present disclosure, there is provided a system for trim test based on a classification model, comprising: a network storage configured to store historical trim test data of integrated circuit chips; a data processing device communicatively connected to the network storage, configured to acquire the historical trim test data, and perform model training based on the historical trim test data to generate a classification trim model; and a plurality of test devices communicatively connected to the data processing device, configured to acquire a trim file generated based on the classification trim model, and send trim parameters generated based on the trim file to a corresponding target chip to trim a trim circuit of the target chip, and acquire a test result of the target chip after being trimmed.

[0028] According to still another aspect of the present disclosure, there is provided an electronic device, comprising: a processor; and a memory configured to store executable instructions of the processor; wherein the processor is configured to perform the method for trim test based on a classification model according to any one of the first aspect.

[0029] According to still another aspect of the present disclosure, there is provided a computer readable medium having stored thereon a computer program, the program being executed by a processor to implement the method for trim test based on a classification model according to any one of the above embodiments.

[0030] The method for trim test based on a classification model provided by the embodiments of the present disclosure can collect historical trim test data of integrated circuit chips as a classification training sample set, and perform model training on the classification training sample set by using a classification algorithm to obtain a trained classification trim model, so that when a target chip is tested, the actual test value of the target chip can be input into the classification trim model to output corresponding trim parameters based on position information of the target chip, and the circuit parameters of a trim circuit of the target chip can be trimmed based on the trim parameters to make the actual test value after trimming match the trim target value, thereby effectively and reliably utilizing a large amount of Trim test data, generating a classification trim model, and further performing trim test on a target chip to be tested by using the classification trim model, which is conducive to simplifying the test process, improving the test efficiency, and reducing the test cost.

[0031] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and are not limiting to the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0032] The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate embodiments consistent with the present disclosure and, together with the description, further serve to explain the principles of the disclosure. It is to be understood that the drawings are designed solely for purposes of illustration to be used in conjunction with the description in envisioning the embodiments. It is to be understood that the drawings are designed solely for purposes of illustration to be used in conjunction with the description in envisioning the embodiments.

[0033] Figure 1 A system diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0034] Figure 2 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0035] Figure 3 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0036] Figure 4 A wafer classification diagram for a trim test scheme based on a classification model is shown according to the present disclosure;

[0037] Figure 5 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0038] Figure 6 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0039] Figure 7 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0040] Figure 8 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0041] Figure 9 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0042] Figure 10 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0043] Figure 11 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0044] Figure 12 A method diagram for performing trim test based on a classification model is shown according to the present disclosure;

[0045] Figure 13 A schematic block diagram of an apparatus for performing trim test based on a classification model is shown according to the present disclosure;

[0046] Figure 14 A schematic block diagram of another apparatus for performing trim test based on a classification model is shown according to the present disclosure;

[0047] Figure 15 A structural schematic diagram of a computer system of an electronic device suitable for implementing an embodiment of the present disclosure is provided for an embodiment of the present disclosure. DETAILED DESCRIPTION

[0048] Example implementations will now be described more fully with reference to the accompanying drawings. Example implementations may, however, be implemented in many different forms and should not be construed as limited to the implementations set forth herein; rather, these implementations are provided so that this disclosure will be thorough and complete, and will fully convey the inventive concept to those skilled in the art. Features, structures, or characteristics described in connection with one implementation can be combined in any suitable manner with features, structures, or characteristics of other implementations.

[0049] In addition, the drawings are only schematic and are non-limiting. Like references signs denote like parts throughout the drawings. Individual features of the drawings are not necessarily to scale, unless explicitly stated so. Some of the embodiments will be described using only circuits and functional blocks to maintain clarity. It will be apparent to those skilled in the art what these functional blocks can be implemented with.

[0050] The flowcharts shown in the figures are only exemplary and do not necessarily have to include all the described steps and do not necessarily have to be executed in the described order. For example, some steps can be split into multiple steps, some steps can be combined or partially combined, and the order of execution can be changed as appropriate. The terms "one", "a" and "the" are used to denote one or more elements / components / etc. The terms "comprise", "include" and "have" are used to mean that one or more elements / components / etc. are included, but do not exclude additional elements / components / etc.

[0051] Referring to Figure 1 An embodiment of the present disclosure first provides a system for performing trim test based on a classification model, comprising:

[0052] a network memory 102 for storing historical trim test data of integrated circuit chips;

[0053] The data processing device 104 is connected in communication with the network storage 102, configured to acquire historical tuning test data, and perform model training based on the historical tuning test data to generate a classification tuning model.

[0054] The plurality of test devices 106 are connected in communication with the data processing device 104, configured to acquire a tuning file generated based on the classification tuning model, and transmit tuning parameters generated based on the tuning file to corresponding target chips, so as to tune the tuning circuit of the target chips, and acquire test results of the target chips after tuning.

[0055] The network storage 102, the data processing device 104 and the plurality of test devices 106 are connected through a communication link to build a wafer factory network, configured to perform detection on wafer chips.

[0056] With reference to Figure 2 The embodiments of the present disclosure further provide a method for tuning test based on a classification model, applied to a data processing device, comprising:

[0057] In step S202, historical tuning test data of integrated circuit chips is acquired, and the historical tuning test data is configured as a classification training sample set.

[0058] The historical tuning test data includes but is not limited to position information of a test chip on an original wafer, an actual test value of the test chip, a tuning target value of the test chip and corresponding tuning parameters.

[0059] Specifically, the classification training sample set can be classified according to the area where the test chip is located in the wafer.

[0060] In step S204, a classification algorithm is used to perform model training on the classification training sample set to obtain a classification tuning model.

[0061] In an exemplary embodiment of the present disclosure, the classification algorithm includes at least one of a random forest, a support vector machine, a K-nearest neighbor algorithm and a decision tree classification algorithm.

[0062] Specifically, the classification tuning model in the present disclosure refers to a model obtained by classification training based on the position information of the test chip.

[0063] In step S206, a tuning file generated based on the classification tuning model is transmitted to a test device, so that the test device tunes the circuit parameters of the tuning circuit of a target chip based on the tuning file, and tests the target chip after tuning.

[0064] The trimming file can be a script file generated based on a classification trimming model, or a data file generated based on trimming parameters, the trimming parameters being parameter values output by the classification trimming model by inputting initial test values of the target chip.

[0065] In this embodiment, by collecting historical trimming test data of integrated circuit chips as a classification training sample set, a classification algorithm is used to train the classification training sample set to obtain a trained classification trimming model. Thus, when trimming test is performed on a target chip, the actual test value of the target chip is input into the classification trimming model to output corresponding trimming parameters based on the position information of the target chip. The circuit parameters of the trimming circuit of the target chip are trimmed based on the trimming parameters, so that the actual test value after trimming matches the trimming target value. Thus, the massive Trim real test data can be effectively and reliably utilized to generate a classification trimming model, and the classification trimming model is further used to perform trimming test on a target chip to be tested, which is beneficial to simplifying the test process, improving the test efficiency, and reducing the test cost.

[0066] Specifically, compared with obtaining the best trimming parameter TrimCode by traversal test (for example, 32 tests are required to trim the GlobalReference voltage), the trimming test scheme based on the classification model in the present application does not require traversal test, and thus the test time cost can be reduced.

[0067] As shown in Figure 3 In an exemplary embodiment of the present disclosure, configuring the historical trimming test data as a classification training sample set includes:

[0068] In step S302, the wafer corresponding to the historical trimming test data is divided into multiple regions.

[0069] In an exemplary embodiment of the present disclosure, one specific implementation of dividing the wafer corresponding to the historical trimming test data into multiple regions includes:

[0070] The wafer is divided into a plurality of regions based on the number of partitions and the partitioning manner of the wafer based on the process parameters, wherein the plurality of regions include a central region and a plurality of annular regions outside the central region.

[0071] As shown in Figure 4 The plurality of regions include a central region 402, and annular regions 404, 406, 408, and 410.

[0072] In step S304, the belonging region is determined based on the position information of the die of the integrated circuit chip on the wafer.

[0073] Step S306, classifying the historical trimming test data based on the home area, and determining the classified set as the classified training sample set.

[0074] In this embodiment, since the dies in different annular areas have certain differences in the corresponding process-related parameters, the trimming parameters of the trimming circuit also have certain differences. The wafer is divided into a central area and a plurality of annular areas outside the central area. The historical trimming test data is classified based on the plurality of areas to obtain a classified training sample set. After receiving an input actual test value, the classified trimming model can determine the area to which the actual test value belongs, output the adaptive trimming parameter based on the area, and ensure the reliability of the application of the classified trimming model.

[0075] As shown in FIG. 1, in an exemplary embodiment of the present disclosure, the classification algorithm includes a K-nearest neighbor algorithm. Figure 5 The model training of the classified training sample set using the classification algorithm includes:

[0076] Step S502, generating a first feature vector based on the first position information, the first test value, the first target value, and the first parameter of the to-be-tested sample.

[0077] The classified training sample set is divided into the to-be-tested sample and the training sample.

[0078] The first position information refers to the coordinate information of the to-be-tested sample on the wafer, the first test value refers to the actual test value of the to-be-tested sample, the first target value refers to the trimming target value of the to-be-tested sample, and the first parameter refers to the trimming parameter of the to-be-tested sample.

[0079] Step S504, generating a plurality of second feature vectors based on the second position information, the second test value, the second target value, and the second parameter of the plurality of training samples in the classified training sample set.

[0080] The second position information refers to the coordinate information of the training sample on the wafer, the second test value refers to the actual test value of the training sample, the second target value refers to the trimming target value of the training sample, and the second parameter refers to the trimming parameter of the training sample.

[0081] Step S506, calculating the distance between the first feature vector and the plurality of second feature vectors.

[0082] Step S508, selecting K training samples closest to the to-be-tested sample based on the calculation result, determining the home area of the first position information based on the home areas of the K second position information, and K is the number of the plurality of areas.

[0083] In step S510, based on the relationship between the home area of the first position information and the actual home area of the test sample, model optimization is performed to generate a classification tuning model based on the model optimization result.

[0084] In this embodiment, by adopting the K-nearest neighbor algorithm as the classification algorithm for training the classification model to obtain the classification tuning model, by reasonably determining the K value, the model output accurate tuning parameters while ensuring that the model has a small model error.

[0085] In an exemplary embodiment of the present disclosure, sending the tuning file generated based on the classification tuning model to the test equipment includes: obtaining a test initial value of the target chip; taking the test initial value as a test sample of the classification tuning model to test the test sample based on the classification tuning model to obtain corresponding tuning parameters; and sending the tuning parameters to the test equipment as a tuning file.

[0086] In another implementation manner of sending the tuning file generated based on the classification tuning model to the test equipment in an exemplary embodiment of the present disclosure, the tuning file is directly generated based on the classification tuning model, and the tuning file is sent to the test equipment.

[0087] In this embodiment, the test initial value of the target chip can be generated on the data processing equipment or on the test equipment, and by reasonably allocating the operation process, the task amount on the data processing equipment and the test equipment can be reasonably allocated.

[0088] In an exemplary embodiment of the present disclosure, the feedback information sent by the test equipment based on the tuning file is also received.

[0089] In one implementation manner, the feedback information includes indication information of model updating, and the classification tuning model is updated.

[0090] In another implementation manner, the number of periods of tuning test in the feedback information is extracted, and when it is detected that the number of periods is greater than or equal to a period threshold, the classification tuning model is updated.

[0091] In this embodiment, when the classification tuning model needs to be updated and optimized, the model update and optimization operation can be performed by receiving the indication information of the updated model directly fed back by the test equipment, or the classification tuning model can be updated and optimized by detecting the number of execution periods of the feedback model, so that different generation ends of model update information can be selected based on different working conditions.

[0092] In one exemplary embodiment of this disclosure, updating the classification adjustment model includes: acquiring test results sent by a test device; updating the classification training sample set based on the test results; and updating the classification adjustment model based on the updated classification training sample set.

[0093] In this embodiment, the classification training sample set is updated by collecting test results sent by the test device, and the classification tuning model is updated and optimized based on the updated classification training sample set to improve the output performance of the classification tuning model.

[0094] In one exemplary embodiment of this disclosure, the method further includes: adjusting the classification adjustment model based on the acquired model adjustment parameters in response to the model adjustment parameters, wherein the model adjustment parameters include at least one of process adjustment parameters, material adjustment parameters, and equipment adjustment parameters.

[0095] In this embodiment, when process adjustments, material adjustments, or testing equipment adjustments occur, the aforementioned adjustment parameters are obtained, and the classification and adjustment model is adjusted based on these parameters.

[0096] Specifically, those skilled in the art will understand that during model training, the adjustment parameters are transformed into corresponding actual test values, adjustment target values, and adjustment parameters to update the classification training sample set. This results in an adjusted classification adjustment model, thereby improving the applicability of the classification adjustment model.

[0097] like Figure 6 As shown, a method for training a classification adjustment model according to an embodiment of the present disclosure includes:

[0098] Step S602: Obtain historical adjustment test data imported from the network storage.

[0099] Step S604: Based on the position of the test chip on the wafer, classify the historical adjustment test data to obtain a classification training sample set.

[0100] Step S606: Use the K-nearest neighbor algorithm to train the model on the classification training sample set to obtain the classification adjustment model.

[0101] Step S608: Output the preferred adjustment parameters based on the classification adjustment model, verify the model error of the classification adjustment model based on the adjustment parameters, and optimize the classification adjustment model based on the verification results.

[0102] Reference Figure 7 The embodiments of this disclosure also provide a method for tuning tests based on a classification model, applied to a test device, including:

[0103] In step S702, the tuning parameter is generated based on the obtained tuning file, and the tuning file is generated based on the classification tuning model.

[0104] In step S704, the tuning parameter is sent to the target chip to tune the tuning circuit of the target chip, and a test result of the target chip after tuning is obtained.

[0105] In this embodiment, by obtaining the tuning file generated based on the classification tuning model, the corresponding tuning parameter is obtained when the target chip is tested, and the circuit parameters of the tuning circuit of the target chip are tuned based on the tuning parameter, so that the actual test value after tuning matches the tuning target value, which is beneficial to simplify the test process, improve the test efficiency and reduce the test cost.

[0106] Referring to Figure 8 According to another method for tuning test based on a classification model according to an embodiment of the present disclosure, applied to a test device, comprising:

[0107] In step S802, the tuning parameter is generated based on the obtained tuning file, and the tuning file is generated based on the classification tuning model.

[0108] In step S804, the tuning parameter is sent to the target chip to tune the tuning circuit of the target chip, and a test result of the target chip after tuning is obtained.

[0109] In step S806, it is determined that the tuning parameter does not meet the tuning requirement based on the test result, a revised tuning parameter is obtained, and the target chip is re-tuned based on the revised tuning parameter, and the revised tuning parameter is generated by revising the tuning parameter based on the classification tuning model.

[0110] In step S808, feedback information is generated based on the test result of the tuning test, and the feedback information is sent to a data processing device to adjust the model based on the feedback information by the data processing device, wherein the data processing device is used to send the tuning file to the test device.

[0111] In step S810, it is determined that the tuning parameter meets the tuning requirement based on the test result, and a corresponding tuning log is generated.

[0112] In this embodiment, by acquiring the test result of the target chip based on the tuning parameter, it is detected whether the tuning parameter meets the tuning requirement. If the tuning parameter meets the tuning requirement, the tuning log is directly output. If the tuning parameter does not meet the tuning requirement, the tuning parameter is revised based on the classification tuning model in a loop, the revised tuning parameter is generated, and the tuning test is performed again based on the revised tuning parameter. If the number of tuning tests performed based on the revised tuning parameter generated in the loop reaches a preset value, feedback information is sent to the data processing device, so that the data processing device updates and optimizes the model based on the feedback information. Real-time optimization and update of the classification tuning model in the actual test process are realized, so as to continuously optimize the classification tuning model and improve the tuning effect of the tuning parameter output by the classification tuning model.

[0113] As shown in Figure 9 In an example embodiment of the present disclosure, in step S806, it is determined that the tuning parameter does not meet the tuning requirement based on the test result, the revised tuning parameter is acquired, and the target chip is re-tuned based on the revised tuning parameter. A specific implementation manner includes:

[0114] In step S902, the test result includes an actual test value of the target chip. It is detected that the deviation between the actual test value and the reference target value is greater than a deviation threshold. It is determined that the tuning parameter does not meet the tuning requirement, and the revised tuning parameter is acquired.

[0115] In step S904, the target chip is tuned based on the revised tuning parameter, and the number of execution times of the tuning test is counted when the revised tuning parameter still does not meet the tuning requirement.

[0116] In step S906, it is detected that the execution times are less than the allowed times. The revised tuning parameter is re-acquired to tune the target chip, and the revised actual test value is obtained.

[0117] In step S908, it is detected that the execution times are greater than or equal to the allowed times. It is determined that one test period is completed, the number of test periods is incremented by 1, and step S902 is returned. The execution times are re-counted when the tuning test is performed based on the revised tuning parameter next time.

[0118] In this embodiment, by reasonably setting the allowed times and the number of periods, the revised tuning parameter is continuously updated based on the classification tuning model in the process of repeatedly performing the tuning test, so that the actual test value approaches the reference target value. After multiple tuning, it is detected that the actual test value still does not meet the requirement, indicating that the model needs to be adjusted, and the data processing device is triggered to update the classification tuning model, so as to ensure the reliability of the model.

[0119] In an example embodiment of the present disclosure, as a specific implementation of obtaining the revised tuning parameter, the method specifically comprises:

[0120] The difference between the tuning parameter and the revision threshold is determined as a third parameter.

[0121] The third parameter refers to a parameter value obtained by subtracting the revision threshold from the initial tuning parameter.

[0122] The sum of the tuning parameter and the revision threshold is determined as a fourth parameter.

[0123] The fourth parameter refers to a parameter value obtained by adding the revision threshold to the initial tuning parameter.

[0124] The classification tuning model is obtained to determine one of the third parameter and the fourth parameter as the revised tuning parameter based on the classification tuning model.

[0125] In an example embodiment of the present disclosure, determining one of the third parameter and the fourth parameter as the revised tuning parameter based on the classification tuning model comprises:

[0126] The classification tuning model outputs a first model target value corresponding to the third parameter.

[0127] The classification tuning model outputs a second model target value corresponding to the fourth parameter.

[0128] The first deviation value between the first model target value and the reference target value, and the second deviation value between the second model target value and the reference target value are determined.

[0129] If the first deviation value is smaller than the second deviation value, the third parameter is determined as the revised tuning parameter.

[0130] If the second deviation value is smaller than the first deviation value, the fourth parameter is determined as the revised tuning parameter.

[0131] In this embodiment, the third parameter and the fourth parameter obtained by respectively modifying the tuning parameter are obtained, and it is detected based on the classification tuning model whether the target value corresponding to the third parameter is closer to the reference target value or the target value corresponding to the fourth parameter is closer to the reference target value, so that the parameter corresponding to the target value closer to the reference target value is determined as the revised tuning parameter based on the detection result, to ensure the reliability of revising the tuning parameter and improve the tuning accuracy when the revised tuning parameter is used for tuning test.

[0132] In one exemplary embodiment of this disclosure, generating feedback information based on the results of the tuning test includes: determining that the number of tuning test cycles is greater than or equal to a cycle threshold based on the results of the tuning test, and generating feedback information including indication information for model updates; or sending the number of tuning test cycles to a data processing device based on the results of the tuning test, and updating the classification tuning model when the data processing device detects that the number of cycles is greater than or equal to the cycle threshold.

[0133] In this embodiment, when it is necessary to update and optimize the classification adjustment model, the test equipment can perform the model update and optimization operation by receiving the update model instruction information directly fed back by the test equipment, or by detecting the number of execution cycles of the model fed back, thereby enabling the selection of different generation ends of model update information based on different working conditions.

[0134] like Figure 10 As shown, a method for adjusting tests based on a classification model according to another embodiment of this disclosure specifically includes:

[0135] Step S1002: Obtain the initial test value for the target information and record the initial test value as Value_default[X,Y].

[0136] Where [X,Y] are the coordinates of the target chip on its respective wafer.

[0137] Step S1004: Input the initial test value into the classification trimming model and output the trimming parameter, and record the trimming parameter as TrimCode[X,Y].

[0138] Step S1006: Send the adjustment parameters to the target chip to perform adjustment test and obtain the actual test value, which is recorded as Value_m[X,Y].

[0139] Step S1008: Determine whether the adjustment parameter corresponding to the actual test value is the optimal solution. If the determination result is "yes", proceed to step S1010. If the determination result is "no", proceed to step S1012.

[0140] Step S1010: Output the trim log.

[0141] Step S1012: Based on the classification adjustment model, detect the deviation between the target value corresponding to the third parameter and the reference target value, and the deviation between the target value corresponding to the fourth parameter and the reference target value.

[0142] The third parameter is denoted as TrimCode[X,Y]-1, and the fourth parameter is denoted as TrimCode[X,Y]+1, where 1 is the revision threshold.

[0143] Step S1014: Determine whether the number of adjustment tests exceeds the allowed number M. If the determination result is "yes", proceed to step S1018; if the determination result is "no", proceed to step S1016.

[0144] Step S1016: Take the smaller deviation value between the third parameter and the fourth parameter as the revised adjustment parameter for testing, and return to step S1006 as the revised adjustment parameter.

[0145] Step S1018: When the number of adjustment tests is greater than or equal to M, increment the number of adjustment test cycles by 1.

[0146] Step S1020: Determine whether the number of adjustment test cycles is greater than or equal to the cycle threshold N. If the determination result is "no", return to step S1006. If the determination result is "yes", proceed to step S1022.

[0147] Step S1022: Feed back model update information to the data processing device so that the data processing device can update the classification and adjustment model according to the collected test results and send it to the test device.

[0148] In one exemplary embodiment of this disclosure, generating tuning parameters based on the obtained tuning file includes: extracting the tuning parameters included in the tuning file; or generating the tuning file based on a classification tuning model, testing the initial test values ​​of the target chip based on the classification tuning model, and obtaining the tuning parameters.

[0149] In this embodiment, the initial test values ​​of the target chip can be generated on the data processing device or on the test device. By reasonably allocating the calculation process, it is beneficial to reasonably allocate the workload on the data processing device and the test device.

[0150] like Figure 11 As shown, an interactive method for tuning tests based on a classification model according to an embodiment of this disclosure specifically includes:

[0151] In step S1102, the network memory sends the historical adjustment test data of the integrated circuit chip to the data processing device.

[0152] In step S1104, the data processing device trains the classification model based on the historical adjustment test data of the integrated circuit chip to obtain the classification adjustment model.

[0153] Step S1106: The data processing device generates a trimming file based on the classification trimming model.

[0154] In step S1108, the data processing device sends a modification file to the testing device.

[0155] Step S1110, the test device obtains the trimming parameter based on the trimming file.

[0156] Step S1112, the test device sends the trimming parameter to the target chip.

[0157] Step S1114, the target chip feeds back the actual test value to the test device.

[0158] Step S1116, the test device detects that the actual test value does not meet the test requirement, revises the trimming parameter based on the classification trimming model, and obtains the revised trimming parameter.

[0159] Step S1118, the test device detects that the number of trimming test cycles reaches the cycle threshold, and generates feedback information.

[0160] Step S1120, the test device sends the feedback information to the data processing device.

[0161] Step S1122, the data processing device updates and optimizes the classification trimming model based on the feedback information.

[0162] Step S1124, the data processing device sends the optimized trimming file to the test device.

[0163] As shown in the figure, the interactive method for trimming test based on the classification model according to another embodiment of the present disclosure specifically includes: Figure 12

[0164] Step S1202, the test device stores the wafer test or package test data on the data network storage, and is configured as a classification training sample set.

[0165] Specifically, the dies of the wafer are defined by coordinates X and Y, and the wafer is divided into five regions 402, 404, 406, 408 and 410 as shown in the figure. Figure 4 The five regions are divided according to the actual process related parameters, so that each training sample in the generated classification training sample set includes the coordinate information of the chip die, the actual test value of the chip die, the trimming target value and the corresponding trimming parameter.

[0166] Step S1204, the data processing device obtains the above-mentioned classification training sample set, and adopts the K-nearest neighbor algorithm to train the model, and obtains the classification trimming model.

[0167] ​Specifically, the test data on the network memory is divided into multiple groups according to the wafer die coordinates. The multiple groups refer to the wafer being divided into 5 regions. Wafer dies in different regions will have corresponding test data. Thus, the test data can be grouped into one data module according to the die coordinates, and the process of importing data is the process of extracting data from the network memory so that it can be classified in the processor later.

[0168] Then, each module is classified according to its coordinate information and trimming parameter Trim code and used as a training set.

[0169] The initial test value Value_default[X,Y] is used as the test set, and the KNN algorithm is used to obtain the best TrimCode[X,Y].

[0170] In step S1206, the data processing device configures the adjustment parameters into a query file and sends it to the test device so that the test program can directly call it.

[0171] Step S1208: Determine whether the generated classification adjustment model needs to be adjusted based on the test results.

[0172] Specifically, during the entire testing process, the number of cycles (count) and the cycle threshold are set. Within a cycle, when the number of adjustment tests exceeds the specified allowable number, the count value is incremented by 1. When the count exceeds the cycle threshold, the process returns to step S1204.

[0173] The following reference Figure 13 This describes an apparatus 1300 for performing adjustment tests based on a classification model according to this embodiment of the invention. Figure 13 The apparatus 1300 shown for adjusting tests based on a classification model is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0174] The device 1300 for tuning tests based on a classification model is represented in the form of a hardware module. The components of the device may include, but are not limited to: a configuration module 1302, used to acquire historical tuning test data of the integrated circuit chip and configure the historical tuning test data as a classification training sample set; a training module 1304, used to train the model on the classification training sample set using a classification algorithm to obtain a classification tuning model; and a sending module 1306, used to send the tuning file generated based on the classification tuning model to the test equipment, so that the test equipment can tune the circuit parameters of the tuning circuit of the target chip based on the tuning file and test the tuned target chip.

[0175] The following reference Figure 14An apparatus 1400 for performing a tuning test based on a classification model according to this embodiment of the present application is described. Figure 14 The apparatus 1400 for performing a tuning test based on a classification model shown is merely an example and should not bring any limitation to the function and applicability of the embodiments of the present application.

[0176] The apparatus 1400 for performing a tuning test based on a classification model is in the form of a hardware module. The components of the apparatus for performing a tuning test based on a classification model can include, but are not limited to: a generating module 1402 for generating tuning parameters based on an acquired tuning file, the tuning file being generated based on a classification tuning model; a testing module 1404 for sending the tuning parameters to a target chip to tune the tuning circuit of the target chip and acquiring a test result after tuning the target chip.

[0177] Reference is made below to Figure 15 which shows a structural schematic diagram of a computer system 1500 of an electronic device suitable for implementing the embodiments of the present disclosure. Figure 15 The computer system 1500 of the electronic device shown is merely an example and should not bring any limitation to the function and applicability of the embodiments of the present disclosure.

[0178] As shown in Figure 15 , the computer system 1500 includes a central processing unit (CPU) 1501, which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 1502 or programs loaded from a storage portion 1508 into a random access memory (RAM) 1503. Various programs and data required for system operation are also stored in the RAM 1503. The CPU 1501, the ROM 1502, and the RAM 1503 are connected to each other through a bus 1504. An input / output (I / O) interface 1505 is also connected to the bus 1504.

[0179] The following components are connected to the I / O interface 1505: an input portion 1506 including a keyboard, a mouse, and the like; an output portion 1507 including a cathode ray tube (CRT), a liquid crystal display (LCD), and the like, and a speaker, and the like; a storage portion 1508 including a hard disk, and the like; and a communication portion 1509 including a network interface card such as a LAN card, a modem, and the like. The communication portion 1509 performs communication processing via a network such as the Internet. A drive 1510 is also connected to the I / O interface 1505 as necessary. A removable medium 1511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 1510 as necessary, so that a computer program read therefrom is installed in the storage portion 1508 as necessary.

[0180] In particular, the processes described above with reference to the flowcharts can be implemented as a computer software program in accordance with embodiments of the present disclosure. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer readable medium, the computer program comprising program code for executing the methods illustrated by the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication section 1209, and / or installed from the detachable medium 1211. When the computer program is executed by the central processing unit (CPU) 1201, the above-described functions defined in the system of the present application are executed.

[0181] It should be noted that the computer readable medium shown in the present disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium may, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or apparatus, or any combination of the above. More specific examples of the computer readable storage medium can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, the computer readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device or apparatus. In the present disclosure, the computer readable signal medium can include a data signal carried in a baseband or as a part of a carrier wave, which carries computer readable program code. Such a propagated data signal can take many forms, including but not limited to an electromagnetic signal, an optical signal or any suitable combination of the above. The computer readable signal medium can also be any computer readable medium that can send, propagate or transmit a program for use by or in connection with an instruction execution system, device or apparatus. The program code contained on the computer readable medium can be transmitted by any suitable medium, including but not limited to wireless, wire, optical cable, RF or the like, or any suitable combination of the above.

[0182] The flowcharts and block diagrams in the drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowcharts or block diagrams can represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or the blocks may also be executed in reverse order, depending on the functionality involved. It will also be noted that each block of the block diagrams or flowcharts, and combinations thereof, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0183] The units described in the embodiments of the present disclosure can be implemented by software, or by hardware, or by a combination of software and hardware. The names of the units described in the embodiments of the present disclosure do not constitute a limitation on the units themselves in some cases.

[0184] As another aspect, the present disclosure also provides a computer readable medium, which can be included in the electronic device described in the above embodiments, or can exist separately and not be assembled into the electronic device. The computer readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device implements the method for performing trimming test based on the classification model as described in the above embodiments.

[0185] For example, the electronic device can implement the following as shown in Figure 2 Step S202, obtaining historical trimming test data of an integrated circuit chip, and configuring the historical trimming test data as a classification training sample set; Step S204, performing model training on the classification training sample set using a classification algorithm, to obtain a classification trimming model; and Step S206, sending a trimming file generated based on the classification trimming model to a test device, so that the test device performs trimming on circuit parameters of a trimming circuit of a target chip based on the trimming file, and tests the trimmed target chip.

[0186] It should be noted that, although several modules or units of the devices for action execution are mentioned in the above detailed description, the division into such modules or units is not mandatory. Indeed, according to an embodiment of the disclosure, the features and functionalities of two or more of the above-described modules or units can be embodied in one module or unit. Conversely, the features and functionalities of one of the above-described modules or units can be further divided into several modules or units.

[0187] Moreover, although the various steps of the methods of the disclosure are described in a particular order in the figures, this is not required or implied as to the order of the steps or the necessity of performing all of the steps to achieve the desired result. Additionally or alternatively, certain steps can be omitted, multiple steps can be combined into one step, one step can be broken into multiple steps, etc.

[0188] From the above description of the embodiments, those skilled in the art will readily perceive that the example embodiments described herein can be implemented by software and / or by software in combination with the necessary hardware. Thus, the technical solution according to the embodiments of the disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, U disk, mobile hard disk, etc.) or network, and includes several instructions to make a computing device (which can be a personal computer, server, mobile terminal, or network device, etc.) execute the method according to the embodiments of the disclosure.

[0189] Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the features of the disclosure as disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the disclosure being indicated by the following claims.

Claims

1. A method of trim test based on a classification model, characterized in that, Applied to a data processing device, comprising: Obtaining historical trim test data of an integrated circuit chip, and configuring the historical trim test data as a classification training sample set; Using a classification algorithm to train a model on the classification training sample set to obtain a classification trim model; Sending a trim file generated based on the classification trim model to a test device, so that the test device adjusts the circuit parameters of the trim circuit of a target chip based on the trim file, and tests the target chip after trim.

2. The method for adjusting tests based on a classification model according to claim 1, characterized in that, The historical trim test data is configured as a classification training sample set, comprising: Dividing a wafer corresponding to the historical trim test data into multiple regions; Determine the home region based on the location information of the die of the integrated circuit chip on the wafer; Classify the historical trim test data based on the home region, and determine the classified set as the classification training sample set.

3. The method of claim 2, wherein the classification model is a decision tree model. The wafer corresponding to the historical trim test data is divided into multiple regions, comprising: Based on the wafer process parameter configuration, the number of partitions and the partitioning method for partitioning the wafer are determined; Based on the number of partitions and the partitioning method, the multiple regions are divided, Wherein, the multiple regions include a center region and multiple annular regions outside the center region.

4. The classification model-based trim test method of claim 2, wherein: The classification algorithm includes at least one of random forest, support vector machine, K-nearest neighbor algorithm, and decision tree classification algorithm.

5. The method of claim 2, wherein the classification model is a decision tree model. The classification algorithm includes K-nearest neighbor algorithm, and the model training of the classification training sample set using the classification algorithm to obtain the classification trim model comprises: Generating a first feature vector based on the first location information, the first test value, the first target value, and the first parameter of the test sample; Generating a plurality of second feature vectors based on the second location information, the second test value, the second target value, and the second parameter of the plurality of training samples in the classification training sample set; Respectively calculating the distance between the first feature vector and the plurality of second feature vectors; Selecting the K training samples closest to the test sample based on the calculation result, to determine the home region of the first location information based on the home region of the K second location information, K being the number of the plurality of regions; Based on the relationship between the home region of the first location information and the actual home region of the test sample, the model is optimized to generate the classification trim model based on the model optimization result.

6. The method of claim 1, wherein, The trim file generated based on the classification trim model is sent to the test device, comprising: Obtaining a test initial value of the target chip; Taking the test initial value as a test sample of the classification trim model to test the test sample based on the classification trim model to obtain corresponding trim parameters, and sending the trim parameters to the test device as the trim file; or Directly generating the trim file based on the classification trim model, and sending the trim file to the test device.

7. The method of trim test based on a classification model according to any one of claims 1 to 6, characterized in that, Further comprising: Receiving feedback information sent by the test device based on the trim file; detecting that the feedback information comprises indication information of model updating, and updating the classification tuning model; or extracting a number of tuning test cycles in the feedback information, and updating the classification tuning model when it is detected that the number of cycles is greater than or equal to a cycle threshold.

8. The method of claim 7, wherein the classification model is a decision tree model. The updating of the classification tuning model comprises: obtaining a test result sent by the test equipment; updating the classification training sample set based on the test result; updating the classification tuning model based on the updated classification training sample set.

9. The method of claim 1 to 6, wherein, Further comprising: adjusting the classification tuning model based on the obtained model adjustment parameter, wherein the model adjustment parameter comprises at least one of a process adjustment parameter, a material adjustment parameter and a device adjustment parameter.

10. A method for tuning tests based on a classification model, characterized in that, Applied to test equipment, comprising: generating a tuning parameter based on an obtained tuning file, the tuning file being generated based on a classification tuning model; sending the tuning parameter to a target chip to tune a tuning circuit of the target chip, and obtaining a test result of the target chip after tuning; determining that the tuning parameter does not meet the tuning requirement based on the test result, obtaining a revised tuning parameter, and re-tuning the target chip based on the revised tuning parameter, the revised tuning parameter being generated by revising the tuning parameter based on the classification tuning model; generating feedback information based on the result of the tuning test, and sending the feedback information to a data processing device to adjust the model based on the feedback information by the data processing device, wherein the data processing device is configured to send the tuning file to the test equipment.

11. The method of claim 10, wherein the classification model is a decision tree model. The determination that the tuning parameter does not meet the tuning requirement based on the test result, the obtaining of the revised tuning parameter, and the re-tuning of the target chip based on the revised tuning parameter comprise: The test result comprises an actual test value of the target chip, and it is detected that the deviation between the actual test value and a reference target value is greater than a deviation threshold, then it is determined that the tuning parameter does not meet the tuning requirement, and the revised tuning parameter is obtained; tuning the target chip based on the revised tuning parameter, and counting the number of tuning test executions when the revised tuning parameter still does not meet the tuning requirement; detecting that the number of executions is less than an allowed number, re-obtaining the revised tuning parameter to tune the target chip; detecting that the number of executions is greater than or equal to the allowed number, determining that a test cycle is completed, and increasing the number of tuning test cycles by 1, and re-counting the number of executions when the next tuning test is performed based on the revised tuning parameter.

12. The method of claim 10, wherein the classification model is a decision tree model. The obtaining of the revised tuning parameter comprises: determining a difference between the tuning parameter and a revision threshold as a third parameter; determining the sum of the tuning parameter and the revision threshold as a fourth parameter; obtaining the classification tuning model to determine one of the third parameter and the fourth parameter as the revised tuning parameter based on the classification tuning model.

13. The method of claim 12, wherein the method further comprises: The determination of one of the third parameter and the fourth parameter as the revised tuning parameter based on the classification tuning model comprises: outputting a first model target value corresponding to the third parameter based on the classification tuning model; outputting a second model target value corresponding to the fourth parameter based on the classification tuning model; determining a first deviation value between the first model target value and a reference target value, and a second deviation value between the second model target value and the reference target value; determining the third parameter as the revised tuning parameter if the first deviation value is less than the second deviation value; determining the fourth parameter as the revised tuning parameter if the second deviation value is less than the first deviation value.

14. The method of claim 10, wherein, The generation of the feedback information based on the result of the tuning test comprises: determining, based on the result of the tuning test, that the number of cycles of the tuning test is greater than or equal to a cycle threshold, generating the feedback information comprising indication information of model updating; or sending, based on the result of the tuning test, the number of cycles of the tuning test to the data processing device, and updating the classification tuning model when the data processing device detects that the number of cycles is greater than or equal to the cycle threshold.

15. The method of trim test based on a classification model according to any one of claims 10 to 14, characterized in that, The generation of the tuning parameter based on the obtained tuning file comprises: extracting the tuning parameter included in the tuning file; or The tuning file is generated based on the classification tuning model, and the tuning parameter is obtained by testing the initial value of the target chip based on the classification tuning model.

16. The method of claim 10 to 14, wherein, Further comprising: determining, based on the test result, that the tuning parameter meets the tuning requirement, and generating a corresponding tuning log.

17. A device for adjustment testing based on a classification model, characterized in that, Applied to a data processing device, comprising: a configuration module configured to obtain historical tuning test data of an integrated circuit chip and configure the historical tuning test data as a classification training sample set; a training module configured to use a classification algorithm to perform model training on the classification training sample set to obtain a classification tuning model; a sending module configured to send a tuning file generated based on the classification tuning model to a test device, so that the test device tunes the circuit parameters of a tuning circuit of a target chip based on the tuning file, and tests the target chip after tuning.

18. A device for adjustment testing based on a classification model, characterized in that, Applied to a test device, comprising: a generation module configured to generate a tuning parameter based on an obtained tuning file, the tuning file being generated based on a classification tuning model; a test module configured to send the tuning parameter to a target chip to tune a tuning circuit of the target chip, and obtain a test result after tuning the target chip.

19. A system for trim test based on a classification model, characterized in that, Comprising: a network storage configured to store historical tuning test data of an integrated circuit chip; a data processing device in communication connection with the network storage, configured to obtain the historical tuning test data and perform model training based on the historical tuning test data to generate a classification tuning model; A plurality of test devices are connected in communication with the data processing device, configured to acquire a trimming file generated based on the classification trimming model, send a trimming parameter generated based on the trimming file to a corresponding target chip, trim a trimming circuit of the target chip, and acquire a test result after trimming the target chip.

20. An electronic device, comprising: Comprise: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the executable instructions to perform the method for trimming test based on the classification model according to any one of claims 1-9 or claims 10-16.

21. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method for trimming test based on the classification model according to any one of claims 1-16.

Images