Thermal prediction and regulation during integrated circuit testing

Abstract

One embodiment of the present invention describes a technique for performing an inspection of an integrated circuit. This technique includes generating one or more temperature predictions of a region of interest within the integrated circuit based on sensor data related to the integrated circuit. This technique also includes determining one or more control parameters related to the integrated circuit test based on one or more temperature predictions and one or more control characteristics related to the integrated circuit. This technique further includes performing the integrated circuit test based on one or more control parameters.

Description

Technology Field

[0001] Cross-reference of related applications

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 755,975, filed on February 7, 2025, under the heading “Thermal Prediction and Adjustment in Testing of Integrated Circuits.” The subject matter of this application is incorporated herein by reference in its entirety.

[0003] The present invention generally relates to the inspection of electronic devices and integrated circuits, and more specifically, to heat prediction and adjustment during the inspection of integrated circuits. Background Technology

[0004] Inspection of an integrated circuit (IC) generally involves a series of tests to verify the functionality, performance, and reliability of the IC and / or semiconductor device. These tests are generally performed to detect various types of defects and / or problems related to the IC and / or various manufacturing stages involving the IC.

[0005] A typical sequence of IC testing may include sort testing to verify the conductivity, leakage, and / or other electrical characteristics of the dies on the wafer. Dies that pass the sort test are separated and packaged, and a burn-in test is performed on each packaged die to apply a load exceeding the corresponding rated operating temperature to identify and eliminate devices associated with early failure. Alternatively, each packaged die may undergo a final test to evaluate its power consumption, frequency, I / O functionality, and other characteristics, and to verify that the packaged die meets design specifications regarding integrity and reliability. Alternatively, multiple dies may be assembled into an integrated system (e.g., System on Chip (SoC), System in Package (SiP), Multi Chip Module (MCM), Printed Circuit Board (PCB), etc.), and system-level testing may be performed to verify the correct operation of the integrated system for the intended end use.

[0006] During IC testing, the IC may be exposed to a wide range of conditions that cause the temperature within the IC to rise and / or fluctuate. To prevent the temperature from reaching unsafe and / or damage levels, a thermal system is commonly used to monitor and / or adjust the temperature when the IC is tested. For example, a conventional thermal system may compare temperature measurements from individual temperature sensors, average temperature values ​​(e.g., from multiple temperature sensors), power consumption, and / or other values ​​measured or determined during a given IC test with one or more preset limits. If the corresponding temperature measurement or temperature value reaches or exceeds the preset limit, the thermal system may generate a control signal to heat or cool the IC. The problem to be solved

[0007] However, as the functionality and integration density of modern ICs increase over time, test requirements become more complex, making IC temperature control more difficult. More specifically, as IC power consumption increases over time, the risk of thermal damage to the IC also rises accordingly. At the same time, in specific areas within complex ICs, packages, and / or integrated IC systems, there is a possibility that temperatures significantly different from those measured by temperature sensors may occur. Furthermore, thermal systems may be unable to respond in a timely manner to sudden changes in heat generation during events such as rapid transitions between test phases and / or operating modes (though not limited to these). Due to the inability to monitor the temperature of the entire IC and / or to compensate for sudden changes in heat generation during IC testing, there is a possibility that the temperature may reach or / or exceed corresponding limits, potentially causing damage or destruction to the IC.

[0008] As mentioned above, what is needed in this technical field is a more effective technology for determining and adjusting the temperature within the IC during IC inspection. means of solving the problem

[0009] One embodiment of the present invention describes a technique for performing an inspection of an integrated circuit. The technique includes generating one or more temperature predictions of a region of interest within the integrated circuit based on sensor data related to the integrated circuit. The technique also includes determining one or more control parameters related to the integrated circuit test based on one or more temperature predictions and one or more control characteristics related to the integrated circuit. The technique further includes performing the integrated circuit test based on one or more control parameters.

[0010] One technical advantage of the disclosed technology over the prior art is the ability to accurately and timely monitor and control the temperature of various Areas of Interest (AOIs) within an IC. As a result, the disclosed technology can avoid and / or mitigate thermal damage and / or destruction caused by the limitations of temperature monitoring by temperature sensors and / or the inability to respond quickly to sudden changes in heat generation during IC inspection. These technical advantages result in one or more technical improvements over the approach of the prior art.

[0011] In order to enable a detailed understanding of the above-described features of the present invention, a more specific description of the present invention, which has been briefly summarized above, can be obtained by referring to embodiments, some of which are illustrated in the attached drawings. However, it should be noted that the attached drawings merely represent typical embodiments of the present invention and, therefore, do not limit the scope thereof, as the present invention may include other equally valid embodiments. Brief explanation of the drawing

[0012] FIG. 1 is a block diagram of a computing system configured to implement one or more aspects of various embodiments. FIG. 2 is a more detailed drawing of the training engine and execution engine of FIG. 1 according to various embodiments. FIG. 3 is a diagram showing exemplary training data of the machine learning model of FIG. 2 according to various embodiments. FIG. 4 is a drawing showing an exemplary system for performing inspection of an integrated circuit (IC) according to various embodiments. FIG. 5 is a flowchart of method steps for performing IC inspection according to various embodiments. Specific details for implementing the invention

[0013] In the following description, a number of specific details are described to provide a more complete understanding of the embodiments of the present invention. However, it will be obvious to those skilled in the art that the embodiments of the present invention may be practiced without one or more of these specific details.

[0014] <Overview>

[0015] As described above, a conventional thermal system can adjust the temperature within an IC during an IC test by comparing the temperature measurement from each temperature sensor, the average temperature value (e.g., from multiple temperature sensors), the power consumption, and / or other values ​​measured or determined during a specific IC test with one or more preset limits. If the preset limits are reached or exceeded during the test, the thermal system can generate a control signal to heat or cool the corresponding IC.

[0016] However, specific areas within complex ICs, packages, and / or integrated IC systems may experience temperatures significantly different from those measured by temperature sensors. Furthermore, thermal systems may be unable to respond in a timely manner to sudden changes in heat generation during events such as rapid transitions between test phases and / or operating modes (though not limited to these). Due to the inability to monitor the temperature of the entire IC and / or to compensate for sudden changes in heat generation during IC testing, there is a possibility that the temperature may reach or / or exceed corresponding limits, potentially causing damage or destruction to the IC.

[0017] To address the above limitations, the disclosed technology predicts and adjusts the temperature of various regions of interest (AOIs) within the IC during inspection of the IC. Each AOI may include two-dimensional (2D) and / or three-dimensional (3D) grid cells, functional blocks, regions of any shape, and / or other types of regions or zones within the IC (though not limited thereto).

[0018] More specifically, the disclosed technology uses machine learning models, physics-based models, simulations, and / or other technologies to generate a temperature prediction for each AOI within the IC based on sensor data, test parameters, and / or other inputs collected during the inspection of the IC. Each temperature prediction may represent the "real silicon temperature" at the corresponding AOI instead of the temperature measured by one or more temperature sensors placed outside the AOI and / or at various distances from the AOI.

[0019] Temperature prediction is also used in conjunction with machine learning models, physics-based models, simulations, and / or other techniques to generate control parameters used to adjust the temperature within the AOI during the inspection of the IC. These control parameters may include (but are not limited to) a target temperature, a thermal control signal of the thermal system, a control signal for adjusting the pace of the inspection, and / or a control signal for initiating or stopping the inspection operation. The control parameters may be determined based on temperature prediction and / or control characteristics related to the test, such as (but not limited to) temperature limits, temperature gradients, power gradients, safety offsets, and / or other values ​​characterizing the progress and / or conditions of the test.

[0020] Temperature prediction and / or control parameters can be generated by an independent computing unit (e.g., Computer On Module (COM), single-board computer, etc.) that operates separately and simultaneously from the thermal system that executes thermal control of the IC using the control parameters and the test system that executes testing of the IC (e.g., Automated Test Equipment (ATE), test program, etc.). This independent computing unit functions as an intermediate control point separating the test system and the thermal system from each other, while enabling the operation of each system to influence the operation of the other system. Through the independent computing unit, the thermal system, the test system, and / or other components can also access the temperature related to the IC on an "on-demand" basis (e.g., without a "blind period" when the temperature sensor is unavailable).

[0021] Temperature prediction and control may be performed across other stages of IC testing, such as wafer sort inspection, die-level inspection, burn-in inspection, final inspection, and / or system-level inspection (though not limited thereto). Data collected during previous inspection stages may be tracked and used to perform thermal prediction and control in subsequent stages of the same IC. Alternatively, this data may also be used to design, evaluate, and / or modify the die layout and / or package containing the IC, the test configuration for testing, and / or other parameters related to the inspection and / or use of the IC.

[0022] One technical advantage of the disclosed technology over the prior art is the ability to accurately and timely monitor and control the temperatures of various AOIs within the IC. As a result, the disclosed technology can avoid and / or mitigate thermal damage and / or destruction caused by limited temperature monitoring by temperature sensors and / or the inability to respond quickly to sudden changes in heat generation during IC inspection. Furthermore, by incorporating thermal data collected during a previous inspection step into thermal prediction and adjustment during a subsequent inspection step involving the same IC, the disclosed technology can improve the design and / or thermal management of integrated systems of multiple components with complex ICs, packages, and / or different thermal responses. These technical advantages result in one or more technical improvements over the approach of the prior art.

[0023] System Overview

[0024] FIG. 1 is a block diagram of a computing system (100) configured to implement one or more aspects of various embodiments. In at least one embodiment, the computing system (100) may include any type of computing device, including, but not limited to, a server machine, a server platform, a desktop machine, a laptop machine, a handheld / mobile device, a digital kiosk, an infotainment system for vehicles, a smart speaker or display, a TV, and / or a wearable device. In at least one embodiment, the computing system (100) is a server machine operating in a data center or cloud computing environment that provides scalable computing resources as a service over a network.

[0025] In various embodiments, the computing system (100) includes, but is not limited to, one or more processors (102) and one or more memories (104) coupled to a parallel processing subsystem (112) via a memory bridge (105) and a communication path (113). The memory bridge (105) is further coupled to an I / O (input / output) bridge (107) via a communication path (106), and then the I / O bridge (107) is coupled to a switch (116).

[0026] In one embodiment, the I / O bridge (107) is configured to receive user input information from any input device (108), such as a keyboard, mouse, touch screen, sensor data analysis (e.g., evaluating gestures, voice, or other information regarding the use of one or more of the field of view or sensory field of view of one or more sensors), VR / MR / AR headset, gesture recognition system, steering wheel, mechanical, digital, or touch-sensitive button or input component, and / or microphone, and to transmit the input information to a processor (102) for processing. In at least one embodiment, the computing system (100) may be a server machine in a cloud computing environment. In this embodiment, the computing system (100) may receive equivalent input information as a command (e.g., in response to one or more inputs from a remote computing device) and / or a message, which is transmitted over a network and received through a network adapter (118), without the input device (108). In at least one embodiment, the switch (116) is configured to provide a connection between the I / O bridge (107) and other components of the computing system (100), such as a network adapter (118) and various ADD-IN CARDs (120 and 121).

[0027] In at least one embodiment, the I / O bridge (107) is coupled to a system disk (114) which may be configured to store content, applications, and data used by the processor (102) and the parallel processing subsystem (112). In one embodiment, the system disk (114) may include a fixed or removable hard disk drive, a flash memory device, and a CD-ROM (Compact Disc Read-Only-Memory), DVD-ROM (digital versatile disc-ROM), Blu-ray (trademark), HD-DVD (high-definition DVD), or other magnetic, optical, or solid-state storage device, which provides non-volatile storage for applications and data. In various embodiments, other components such as a USB (Universal Serial Bus) or other port connection, a compact disc drive, a digital versatile disc drive, or a film recording device may also be connected to the I / O bridge (107).

[0028] In various embodiments, the memory bridge (105) may be a Northbridge chip, and the I / O bridge (107) may be a Southbridge chip. Additionally, communication paths (106 and 113) and other communication paths within the computing system (100) may be implemented using any technically suitable protocol, including but not limited to Accelerated Graphics Port (AGP), Hyper Transport, or any other bus or point-to-point communication protocol known in the art.

[0029] In at least one embodiment, the parallel processing subsystem (112) includes a graphics subsystem that supplies pixels to any display device (110), which may be any conventional cathode ray tube, liquid crystal display, light-emitting diode display, etc. In such an embodiment, the parallel processing subsystem (112) may include a circuit optimized for graphics and video processing, for example, including a video output circuit. Such a circuit may be integrated into one or more parallel processing units (PPUs), which are also referred to herein as parallel processors, included within the parallel processing subsystem (112).

[0030] In at least one embodiment, the parallel processing subsystem (112) includes a circuit optimized (e.g., optimized) for general-purpose and / or computational processing. That is, such a circuit may be integrated into one or more PPUs included within the parallel processing subsystem (112) configured to perform general-purpose and / or computational operations. In another embodiment, one or more PPUs included within the parallel processing subsystem (112) may be configured to perform graphics processing, general-purpose processing, and / or computational processing operations. The memory (104) includes at least one device driver configured to manage the processing operations of one or more PPUs within the parallel processing subsystem (112). Additionally, the memory (104) includes instructions for realizing a training engine (122) and an execution engine (124) that can be executed by the processor and / or the parallel processing subsystem (112).

[0031] In various embodiments, the parallel processing subsystem (112) may be integrated with one or more of the other elements of FIG. 1 to form a single system. For example, the parallel processing subsystem (112) may be integrated with the processor (102) and other connection circuits on a single chip to form a system-on-chip (SoC).

[0032] The processor (102) may include any suitable processor implemented as a combination of other processing units, such as a CPU (Central Processing Unit), GPU (Graphics Processing Unit), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array), AI (Artificial Intelligence) accelerator, DLA (Deep Learning Accelerator), Parallel Processing Unit (PPU), Data Processing Unit (DPU), Vision Processing Unit (VPU), Programmable Vision Accelerator (PVA) (may include one or more VPUs, PPE (Pixel Processing Engine), and / or DMA (Direct Memory Access) system), any other type of processing unit, or a CPU configured to operate in conjunction with a GPU. Generally, the processor (102) may include any technically feasible hardware unit capable of processing data and / or executing software applications. Additionally, in the context of the present disclosure, the computing element illustrated in the computing system (100) may correspond to a physical computing system (e.g., a system within a data center or machine) and / or may correspond to a virtual computing instance running within a computing cloud.

[0033] In at least one embodiment, the processor (102) issues a command to control the operation of the PPU. In at least one embodiment, the communication path (113) is a PCIe (Peripheral Component Interconnect Express) link in which a dedicated lane is allocated to each PPU. Other communication paths may be used. The PPU may advantageously implement a high-parallel processing architecture, and the PPU may be equipped with local parallel processing memory (PP memory) of any capacity.

[0034] It will be understood that the system presented in this specification is exemplary and that various variations and modifications are possible. The connection topology, including the number and arrangement of bridges, the number of processors (102), and the number of parallel processing subsystems (112), may be changed as needed. For example, in at least one embodiment, the memory (104) may be connected directly to the processor (102) without passing through the memory bridge (105), and other devices may communicate with the memory (104) through the memory bridge (105) and the processor (102). In another embodiment, the parallel processing subsystem (112) may be connected directly to the processor (102) rather than to the I / O bridge (107) or the memory bridge (105). In yet another embodiment, the I / O bridge (107) and the memory bridge (105) may be integrated into a single chip instead of existing as one or more separate devices. In certain embodiments, one or more components shown in FIG. 1 may not be present. For example, the switch (116) may be omitted, and the network adapter (118) and ADD-IN CARD (120, 121) are directly connected to the I / O bridge (107). Additionally, in certain embodiments, one or more components shown in FIG. 1 may be implemented as virtual resources within a virtual computing environment, such as a cloud computing environment.

[0035] In one or more embodiments, the computer system (100) is configured to execute a training engine (122), an execution engine (124), and a management engine (126) present in system memory (104). The training engine (122), the execution engine (124), and the management engine (126) may be stored in a system disk (114) and / or other storage device and loaded into system memory (104) at the time of execution.

[0036] In some embodiments, the training engine (122), the execution engine (124), and the management engine (126) include the functionality to perform thermal prediction and adjustment during the inspection of an integrated circuit (IC). More specifically, the training engine (122) and the execution engine (124) may use machine learning models, physics-based models, simulations, experiments, and / or other techniques to generate temperature predictions for various regions of interest (AOIs) within a given IC. The training engine (122) and the execution engine (124) may also use the same techniques and / or other techniques to generate control parameters and / or control signals used to adjust the temperature within the AOI during the inspection of the IC. The temperature predictions, control parameters, and / or control signals may be generated and / or updated within a control loop, and as a result, the temperature received by the AOI may be adjusted before the temperature reaches a level that could damage the IC.

[0037] The management engine (126) can generate, store, and / or update records of temperature predictions, control parameters, control signals, test results, die layouts, test configurations, and / or other data used in various stages of IC inspection, including (but not limited to) wafer sorting, die-level processing, burn-in inspection, final inspection, and / or system-level testing. The management engine (126) may further use data collected in previous inspection stages to perform thermal prediction and / or adjustment of the same IC in subsequent inspection stages. The operation of the training engine (122), the execution engine (124), and the management engine (126) is described in more detail below.

[0038] <Thermal Prediction and Adjustment During Integrated Circuit Testing>

[0039] FIG. 2 is a more detailed drawing of the training engine (122), execution engine (124), and management engine (126) of FIG. 1 according to various embodiments. As described above, the training engine (122), execution engine (124), and management engine (126) include the functionality to perform thermal prediction and adjustment during IC inspection. Each of these components is described in more detail below.

[0040] In one or more embodiments, thermal adjustment and control are performed during a given IC test using one or more machine learning models (208). More specifically, one or more machine learning models (208) can generate temperature predictions (242) for various AOIs on the IC based on inputs (240) related to the test and / or IC. Each temperature prediction represents the "actual silicon temperature" at the corresponding AOI instead of the temperature measured by one or more temperature sensors placed outside the AOI and / or at various distances from the AOI. As a result, thermal responses and / or characteristics at locations within the IC that cannot be directly measured using temperature sensors can be evaluated using the temperature predictions (242).

[0041] The same machine learning model (208) and / or one or more additional machine learning models (208) may further generate control parameters (246) used to adjust the temperature within the AOI during testing based on temperature prediction (242) and / or other inputs (240). For example, using one or more machine learning models (208), a target temperature, a temperature limit, a thermal control signal, a signal to adapt the test phase distribution and / or a signal to stop the test, and / or other values ​​that can be used to control and / or adjust the temperature within the IC and / or each AOI within the IC.

[0042] In one or more embodiments, an AOI corresponds to a region within the IC. For example, the IC may be divided into two-dimensional (2D) and / or three-dimensional (3D) grid cells corresponding to different AOIs. In another example, each AOI may be defined as a region within the IC having any shape and / or size. In a third example, a given AOI may correspond to a functional block, layer, and / or other individual part of the IC.

[0043] The machine learning model (208) may include a multilayer perceptron, a Recurrent Neural Network (RNN), a Convolutional Neural Network (CNN), a transformer neural network, a regression model, a tree-based model, a time series model, a Support Vector Regression (SVR) model, and / or other types of models capable of processing numerical data and / or time series data (but not limited thereto). Predictions generated by a given machine learning model may be supplemented and / or replaced by predictions generated through one or more other techniques, such as physics-based models, control loops, simulations, and / or experiments (but not limited thereto). For example, multiple temperature predictions generated by a machine learning model (208), a physics-based model, a control loop, a simulation, and / or experiments for a given AOI may be combined into an overall temperature prediction of the AOI through averaging, weighted combination, additional machine learning models, and / or other intensive techniques.

[0044] The training engine (122) trains a machine learning model (208) to generate temperature predictions (242) and / or control parameters (246) related to AOIs. As shown in FIG. 2, the training engine (122) updates the model parameters (206) of the machine learning model (208) using training data (200) including training inputs (202) and training Ground Truth (GT) values ​​(204). For example, the training engine (122) may generate and / or acquire training data (200) through one or more simulations, experiments, physics-based models, measurements, past tests of ICs, and / or other techniques. The training engine (122) may input training inputs (202) related to one or more AOIs into one or more machine learning models (208). The training engine (122) can process the training input (202) using the model parameters (206) of the machine learning model to obtain a training output (210) including temperature predictions (242) and / or control parameters (246) of the AOI from one or more layers, blocks, and / or components of the machine learning model. The training engine can calculate one or more losses (254) (e.g., mean squared error, mean absolute error, Huber loss, etc.) between the training output (210) and the corresponding training GT (Ground Truth) values ​​(204) in the training data. Subsequently, the training engine (122) can iteratively update the model parameters (206) of the machine learning model to reduce the loss (254) using a training technique (e.g., gradient descent and backpropagation).

[0045] FIG. 3 illustrates exemplary training data (200) of the machine learning model (208) of FIG. 2 according to various embodiments. As shown in FIG. 3, the training data (200) may include a test configuration (234) related to a predetermined test and / or IC, sensor data (302), control characteristics (304), a Ground Truth (GT) temperature (306), and / or Ground Truth (GT) control parameters (308) (but are not limited thereto).

[0046] Each of the test configurations (234) includes information regarding a specific type of test of a given IC. For example, the test configuration may include one or more target temperatures, one or more temperature limits (e.g., a temperature limit of the test base, a damage limit, a damage "pre-limit" that pauses the test program until a safe temperature is reached, etc.), an AOI related to the test, a measurement performed during the test, a test step performed during the test, a device used in the test, and / or other values ​​characterizing the test (but not limited thereto).

[0047] Sensor data (302) includes actual and / or simulated measurements of attributes related to a given IC. These attributes include temperature (e.g., at the location of one or more sensors on and / or around the IC), power (e.g. via a power monitor), aging degradation, current, voltage, voltage drop, and / or other values ​​characterizing the state of the IC (though not limited thereto).

[0048] The control characteristic (304) includes information regarding the progress of a predetermined test and / or the thermal behavior of the IC during the test. For example, the control characteristic (304) may include the sum of temperature values ​​exceeding and falling below a target temperature and / or limit, a temperature gradient, a power gradient, an increase or decrease in the temperature or power gradient, a minimum deviation and / or maximum deviation of the temperature (or other attribute) from the corresponding target and / or limit, a warning and / or alarm regarding a damage limit (or other type of event), a test continuation signal, one or more safety offsets regarding the target temperature, a distribution of power values ​​per AOI, and / or anomalous / outlier values ​​for temperature, power, and / or other attributes (but is not limited thereto).

[0049] The GT (Ground Truth) temperature (306) includes the temperature value of each AOI within the IC. For example, each GT (Ground Truth) temperature may be generated by simulation, experiment, physics-based model, measurement, and / or other techniques. Each GT (Ground Truth) temperature may also be associated with a corresponding test configuration, a set of sensor data (302), and / or control characteristics (304), and as a result, the test configuration, sensor data (302), and / or control characteristics (304) function as features that can be used by a machine learning model to predict the GT (Ground Truth) temperature.

[0050] The GT (Ground Truth) control parameter (308) includes values ​​related to the control and / or adjustment of the temperature within the IC and / or each AOI within the IC. For example, the GT (Ground Truth) control parameter (308) may include a target temperature related to a temperature sensor within and / or around the IC, a target temperature related to a specific AOI within the IC and / or other types of thermal control signals used to control a thermal system that adjusts the temperature of one or more parts of the IC, and / or any other types of inputs that can be used to control the thermal state of the IC (but are not limited thereto). Alternatively, the GT (Ground Truth) control parameter (308) may also include control inputs related to the execution of the test, such as signals related to adapting the pace of the test (e.g., enabling cooling and / or temperature stabilization) and / or stopping the test (e.g., to prevent damage to the IC) (but are not limited thereto). Alternatively, the Ground Truth (GT) control parameters (308) may also include other information and / or test results that can be recorded for quality improvement, statistics, analysis of the root cause of failure, and / or evaluation of the quality of the IC, and / or for subsequent use. A set of Ground Truth (GT) control parameters (308) may be associated with a set of test configurations, sensor data (302), control characteristics (304), and / or a set of Ground Truth (GT) temperatures (306), and as a result, the test configurations, sensor data (302), control characteristics (304), and / or Ground Truth (GT) temperatures (306) function as features that can be used by a machine learning model to predict their Ground Truth (GT) control parameters (308).

[0051] In some embodiments, a test configuration (234), sensor data (302), and / or control characteristics (304) are used as training inputs (202) processed by a predetermined machine learning model to generate a corresponding training output (210). These training inputs (202) may also be associated with a training GT (Ground Truth) value (204) that includes a GT (Ground Truth) temperature (306), a GT (Ground Truth) control parameter (308), and / or other values ​​to be predicted from the training inputs (202). For example, a predetermined test configuration, a set of sensor data (302), and / or a set of control characteristics (304) may be input to a machine learning model trained to predict one or more corresponding GT (Ground Truth) temperatures (306), and / or selected using it. In another example, a predetermined test configuration, a set of sensor data (302), a set of control characteristics (304), and / or a Ground Truth (GT) temperature (306) may be input into a machine learning model trained to predict one or more corresponding Ground Truth (GT) control parameters (308), and / or selected using it.

[0052] In one or more embodiments, the training engine (122) trains the machine learning model (208) over one or more training steps to optimize the machine learning model (208) based on various losses (254), types of training outputs (210), performance characteristics, and / or other purposes or factors. For example, the training engine (122) may train one or more machine learning models (208) of a first set to predict the Ground Truth (GT) temperature (306) based on the corresponding test configuration (234), sensor data (302), and / or control characteristics (304). The training engine (122) may separately train one or more machine learning models (208) of a second set to predict the Ground Truth (GT) control parameters (308) based on the corresponding test configuration (234), sensor data (302), control characteristics (304), and / or Ground Truth (GT) temperature (306). Or instead, the training engine (122) may also train two sets of machine learning models (208) in an end-to-end manner, and as a result, the control parameters (246) generated by the second set of machine learning models (208) are optimized based on the temperature predictions (242) generated by the first set of machine learning models (208). Or instead, the training engine (122) may also train each machine learning model (208) to generate both the temperature predictions (242) and the control parameters (246) for different ICs and / or AOIs within the ICs.

[0053] Returning to the discussion of FIG. 2, after the training of the machine learning model (208) is completed, the training engine (122) may store the parameters and / or other representations of the trained machine learning model (208) in a database, cloud storage, distributed file system, and / or other type of data store (212) for subsequent retrieval and use.

[0054] The execution engine (124) generates temperature predictions (242) and control parameters (246) used to control the temperature and / or other parameters of various IC tests using trained machine learning models (208) from the training engine (122) and / or data store (212), physics-based models, simulations, experiments, control loops, and / or other techniques. More specifically, the execution engine (124) may collect inputs (240) including (but not limited to) temperature, power, aging degradation, current, voltage, voltage drop, and / or other types of sensor data characterizing the state of the IC. The inputs (240) may further include attributes of the target temperature, temperature limits, AOI, measurements to be performed, test steps, equipment used, and / or other test configurations characterizing the specific test of the IC. The input (240) may also include the sum of upper and lower temperature values ​​of the target temperature and / or limit, the temperature gradient, the power gradient, the change in the temperature or power gradient, the minimum and / or maximum deviation of the temperature from the corresponding target and / or limit, the warning and / or alarm regarding the damage limit, the test continuation signal, one or more safety offsets regarding the target temperature, the distribution of power values ​​per AOI, the attribute values ​​of anomalous / outlier values, and / or other control characteristics characterizing the progress of the test.

[0055] The execution engine (124) may generate a temperature prediction (242) of the AOI based on an input (240) using a machine learning model (208) and / or other techniques. The execution engine (124) may further generate control parameters (246) used to adjust the temperature within the AOI and / or IC based on the input (240) and / or temperature prediction (242) using a machine learning model (208) and / or other techniques. The execution engine (124) may further generate a test output (248) containing sensor data, control characteristics, test steps, and / or other information related to the temperature prediction (242) and / or control parameters (246) collected before, during, and / or after applying a predetermined set of the temperature prediction (242), control parameters (246), and control parameters (246) to the IC. The test outputs (248) of these may be stored in a data store (212) and / or provided as additional inputs (240) used by a machine learning model (208) and / or other techniques to generate new temperature predictions (242) and / or control parameters (246) for subsequent time steps of the test.

[0056] In one or more embodiments, the execution engine (124) operates on a separate, independent computing unit, distinct from the test system used to perform IC testing and the thermal system used to adjust the temperature of the IC being tested. As described in more detail below with respect to FIG. 4, the independent computing unit can function as an intermediate control point that separates the inspection of the IC by the test system from the thermal control of the IC by the thermal system, while enabling the test system and the thermal system to influence each other during testing.

[0057] FIG. 4 illustrates an exemplary system (400) for performing IC testing according to various embodiments. As shown in FIG. 4, the system (400) includes a thermal system (402), a test system (404), and an execution engine (124).

[0058] The test system (404) includes an automatic test device (ATE) that performs an inspection of a Device Under Test (DUT) (406) (e.g., IC) according to a series of test steps defined in a test program (408). For example, the test program (408) may include commands, control signals, and / or instructions executed by the test system (404) to execute the test steps. During the execution of the test steps, sensor data (412), including temperature measurements, power consumption values, voltage readings, current measurements, and / or other attributes (though not limited to these) characterizing the operating state of the IC, may be collected by a sensor integrated into the test system (404).

[0059] The execution engine (124) receives sensor data (412) from the DUT (406), details related to the current test step from the test program (408), and / or other inputs (240) related to the DUT (406) and / or the test. The execution engine (124) processes these inputs (240) using a trained machine learning model (208), a physics-based model, a simulation, and / or other techniques to generate temperature predictions (242) for various AOIs within the DUT (406).

[0060] The execution engine (124) also generates control parameters (246) based on temperature prediction (242), control characteristics related to the test, and / or other information related to the test. The control parameters (246) may include a target temperature for thermal adjustment, a thermal control signal, adjustment of a test step or test action, and / or other parameters affecting the thermal state of the DUT (406).

[0061] The thermal system (402) receives control parameters (246) from the execution engine (124) and uses these control parameters (246) to control the temperature of the DUT (406). For example, the thermal system (402) may include a heating element, a cooling system, a temperature controller, and / or other thermal control unit (410) that can be used to adjust the thermal environment of the DUT (406).

[0062] As shown in FIG. 4, the thermal system (402) receives sensor data (412) directly from the test system (404) and can use the sensor data (412) to adjust the thermal control unit (410) separately from the control parameter (246) received from the execution engine (124). Likewise, the test system (404) can receive the values ​​of the temperature prediction (242) and / or control parameter (246) from the execution engine (124), and / or the values ​​of the thermal control unit (410) from the thermal system (402). The test system (404) can use the received values ​​to adjust the execution of the test step specified in the test program (408) (e.g., by adjusting the pace distribution of the test, stopping the test, executing and / or omitting a specific test step, etc.).

[0063] As described above, the execution engine (124) can be executed on an independent computing unit (e.g., computer-on-module (COM), single-board computer, etc.) separate from the test system (404) and the thermal system (402). This independent computing unit functions as an intermediate control point that separates the inspection of the DUT (406) by the test system (404) from the thermal control of the DUT (406) by the thermal system (402). At the same time, the temperature prediction (242) and / or control parameters (246) generated by the execution engine (124) are generated based on test content related to (i) the test program (408) and / or the test system (404), and (ii) thermal sensor data (412) that reflects the influence of the thermal control unit (410) on the temperature within the DUT (406). By this, the test system (404) and the thermal system (402) can influence each other during the test. Additionally, since the execution engine (124) generates temperature predictions (242) and / or control parameters (246) at a rate independent of the rate at which sensor data (412) is collected, the thermal system (402), test system (404), and / or other components can access thermal data (e.g., temperature predictions (242), control parameters (246), etc.) related to the DUT (406) on an "on-demand" basis (e.g., in a way that avoids "blind periods" during which the temperature and / or other sensors are unavailable to generate sensor data (412)).

[0064] Returning to the discussion of FIG. 2, the management engine (126) generates and updates records (214) used to track test outputs (248) and / or other types of data or entities involved in the inspection of a given IC. These records (214) may be stored in a data store (212) (e.g., one or more database tables, files, etc.) and / or shared with the training engine (122), the execution engine (124), and / or other components involved in thermal prediction and adjustment during IC testing.

[0065] As shown in FIG. 2, the record (214) includes sort data (220), die-level data (222), package-level data (224), burn-in data (226), final test data (228), and / or system-level data (230) collected during the corresponding steps of IC inspection. The sort data (220) may include conductivity, leakage current, voltage, frequency, thermal data, and / or data collected during wafer sort inspection of the wafer. As a result, the sort data (220) can acquire the thermal behavior of each die on the wafer before packaging.

[0066] Die-level data (222) may include data generated during die-level inspection of each die after separation from the wafer and before packaging. For example, the die-level data (222) may include a temperature profile, a power dissipation pattern, a thermal resistance value, and other thermal characteristics unique to each die. Alternatively, the die-level data (222) may also include information regarding the placement, orientation, physical characteristics, and / or other properties of the die that affect thermal behavior during inspection.

[0067] Package level data (224) may include information collected after the die is assembled into the package. For example, package level data (224) may include packaging materials, thermal interfaces, package geometry, and / or other attributes that affect the thermal response of the IC. Or instead, package level data (224) may also include case temperature, thermal resistance between the junction and the case, heat dissipation pattern, thermal coupling between multiple dies within a multi-die package, and / or other thermal characteristics of a given package.

[0068] Burn-in data (226) may include information collected during the burn-in inspection of the packaged die. For example, the burn-in data (226) may include changes in temperature drift, changes in power consumption, thermal stability indicators, and / or other thermal characteristics under time-dependent and / or changing stress conditions. Alternatively, the burn-in data (226) may also include information related to thermal cycling effects and temperature-dependent failure modes in the packaged die.

[0069] The final test data (228) may include information collected during the final inspection of the packaged die. For example, the final test data (228) may include various operating modes, power states, and / or temperature profiles related to test conditions.

[0070] System-level data (230) may include information collected during system-level inspection of multiple ICs integrated into a larger system. For example, system-level data (230) may obtain thermal interactions between other components, system-level thermal management requirements, and / or thermal performance under various operating conditions and / or end uses. Alternatively, system-level data (230) may also include information related to thermal coupling effects, thermal diffusion, and system-level thermal control strategies.

[0071] In some embodiments, sort data (220), die level data (222), package level data (224), burn-in data (226), final test data (228), and / or system level data (230) include an identifier associated with a corresponding inspection step (e.g., test step identifier, IC identifier, etc.), an input (240), a temperature prediction (242), a control parameter (246), and / or a test output (248). Additionally, sort data (220), die level data (222), package level data (224), burn-in data (226), final test data (228), and / or system level data (230) of a given IC may be stored in a data store (212) and used by a management engine (126) to improve inspection of the IC and / or thermal adjustment across inspection steps.

[0072] More specifically, the management engine (126) may provide data collected during a previous inspection step as input (240) to a machine learning model (208) and / or other components to generate temperature predictions (242), control parameters (246), and / or test outputs (248) in a subsequent inspection step including the same IC and / or components. For example, the management engine (126) may use sort data (220) and / or die-level data (222) for a given die as a set of baseline thermal characteristics that can be used to generate temperature predictions (242), control parameters (246), and / or thermal responses for the same die in a subsequent inspection step. In another example, the management engine (126) may use package-level data (224), burn-in data (226), and / or final test data (228) of a given package to notify system-level inspection and / or thermal adjustment of an integrated system including the package.

[0073] Or instead, the management engine (126) may also generate and / or modify a die layout (232), test configuration (234), and / or test result (236) associated with a given IC by using sort data (220), die level data (222), package level data (224), burn-in data (226), final test data (228), and / or system level data (230) within the record (214). As shown in FIG. 2, the management engine (126) may provide a user interface (216) that can be used to view, search, modify, and / or access the record (214). Or instead, the user interface (216) may also be used to create, display, and / or modify the die layout (232) within the package and / or integrated system based on the thermal response of the corresponding die (e.g., control parameters (246), as represented by the temperature prediction (242), and / or other thermal information within the sort data (220) and / or die level data (222) associated with the die). Or instead, the user interface (216) may also be used to set and / or adjust the thermal limits and / or other attributes of the test configuration (234) for one or more test steps based on the thermal response determined during a prior test step involving the same IC. Or instead, the user interface (216) may also be used to view and / or compare test results (236) for the same IC and / or different ICs across test steps. As a result, users and / or other entities involved in the design and / or inspection of wafers, dies, packages, integrated systems, and / or other IC-based systems or components can optimize various steps in the design and inspection of IC-based systems or components by means of the management engine (126) and / or user interface (216).

[0074] FIG. 5 is a flowchart of method steps for performing an inspection of an integrated circuit according to various embodiments. Although the method steps are described in relation to the system of FIG. 1, FIG. 2 and FIG. 4, those skilled in the art will understand that any system configured to perform some or all of the method steps in any order is included within the scope of this disclosure.

[0075] As described, in step (502), the training engine (122) collects training data including sensor data, temperature, test parameters, control characteristics, and / or control parameters related to the testing of one or more ICs. For example, the training engine (122) may generate and / or retrieve training data through one or more simulations, experiments, physics-based models, measurements, past tests of ICs, and / or other techniques.

[0076] In step (504), the training engine (122) trains one or more machine learning models to predict temperature and / or control parameters related to one or more AOIs within the IC using training data. For example, the training engine (122) may train a first machine learning model to predict the temperature of an AOI within a given IC based on an input including sensor data collected during IC testing. Alternatively, the training engine (122) may also train a second machine learning model to predict a target temperature, a thermal control signal, and / or other types of thermal control parameters related to the AOI and / or IC based on the predicted temperature output by the first machine learning model (or through other techniques) and / or control characteristics calculated during IC testing.

[0077] In step (506), the execution engine (124) generates one or more temperature predictions for each AOI in the IC based on sensor data collected during a test instance corresponding to the test. For example, the execution engine (124) may be executed on an independent computing unit during the execution of the test instance. Once sensor data is collected, the execution engine (124) may generate temperature predictions for each AOI periodically and / or continuously using one or more machine learning models trained using steps (502) and (504) and / or other techniques. If temperature predictions need to be generated simultaneously for multiple AOIs on the IC and / or multiple ICs, the execution engine (124) may generate temperature predictions in parallel (e.g., multiple processors, processor cores, etc.). Each temperature prediction may represent the "actual silicon temperature" in the corresponding AOI instead of the temperature measured by a sensor potentially located outside the AOI.

[0078] In some embodiments, the execution engine (124) generates a temperature prediction using additional information related to the IC. For example, the execution engine (124) can predict the temperature of each AOI within the IC based on an input including sensor data, already known and / or simulated physical properties of the IC, statistical information related to the IC and / or IC inspection, existing test results related to the IC (e.g., from past IC tests and / or previous inspection steps of the IC), test parameters related to the test, manufacturing data related to the IC, and / or other applicable information related to the IC and / or test, using additional machine learning models, physics-based models, simulations, and / or other techniques trained during steps (502) and (504).

[0079] In step (508), the execution engine (124) determines one or more control parameters based on one or more control characteristics related to temperature prediction and test instances. For example, the execution engine (124) may generate a target temperature, thermal control signal, command, and / or other type of value that can be used to control the operation of a thermal system controlling the temperature of an IC during a test instance and / or during the operation of a test system executing a test instance, using one or more machine learning models trained using steps (502) and (504) and / or other techniques.

[0080] In step (510), the execution engine (124) executes a test instance based on control parameters. For example, the execution engine (124) may transmit control parameters to a thermal system and / or a test system and / or generate an output that distributes the control parameters to the thermal system and / or the test system. The thermal system may use the control parameters to control the temperature of the IC and / or each part of the IC (e.g., so that the temperature limit is avoided, so that the target temperature is subsequently reached and / or maintained). The test system may use the control parameters to change the phase distribution of the test instance, pause the test instance, modify the test steps executed in the test instance, and / or adjust the execution of the test instance. Alternatively, the execution engine (124) and / or the management engine (126) may also generate and / or record test results and / or other data related to the test instance based on temperature predictions, control parameters, and / or other information collected and / or generated during the test instance.

[0081] In step (512), the execution engine (124) determines whether to continue the execution of the test instance. For example, the execution engine (124) may determine that the test instance must continue to be executed until the step of the test is executed, the result of the test is determined, a certain amount of time has elapsed, and / or other conditions are met.

[0082] While the execution engine (124) determines that it should continue the execution of the test instance, the execution engine (124) repeats steps (506), (508), and (510) to generate other outputs that can be used to predict and / or adjust the temperature, control parameters, and / or the temperature in the IC. The execution engine (124) also executes step (512) periodically and / or continuously to determine whether to continue the execution of the test instance.

[0083] In step (512), when the execution engine (124) determines that the execution of the test instance should no longer continue, the management engine (126) executes step (514) to store temperature predictions, control parameters, and / or other thermal data generated during the test instance in one or more records related to the test instance. For example, the management engine (126) may store a mapping between the thermal data and one or more identifiers of the test instance and / or IC in the records. The records are then stored in a data store and linked to additional data collected during other inspection steps involving the IC, and may be used to coordinate subsequent thermal control and / or inspection of the IC, to create and / or modify the design of a package and / or integrated system containing the IC, and / or to perform other tasks related to thermal coordination, inspection, and / or IC design.

[0084] In summary, the disclosed technology predicts and adjusts the temperature of various regions of interest (AOIs) within an IC during inspection of the IC. Each AOI may include two-dimensional (2D) and / or three-dimensional (3D) grid cells, functional blocks, regions of any shape, and / or other types of regions or zones within the IC (though not limited thereto).

[0085] More specifically, the disclosed technology uses machine learning models, physics-based models, simulations, and / or other technologies to generate a temperature prediction for each AOI within the IC based on sensor data, test parameters, and / or other inputs collected during the inspection of the IC. Each temperature prediction may represent the "actual silicon temperature" at the corresponding AOI instead of the temperature measured by one or more temperature sensors placed outside the AOI and / or at various distances from the AOI.

[0086] Temperature prediction is also used in conjunction with machine learning models, physics-based models, simulations, and / or other techniques to generate control parameters used to adjust the temperature within the AOI during the inspection of the IC. These control parameters may include (but are not limited to) a target temperature, a thermal control signal of the thermal system, a control signal for adjusting the pace of the inspection, and / or a control signal for initiating or stopping the inspection operation. The control parameters may be determined based on temperature prediction and / or control characteristics related to the test, such as (but not limited to) temperature limits, temperature gradients, power gradients, safety offsets, and / or other values ​​characterizing the progress and / or conditions of the test.

[0087] Temperature prediction and / or control parameters can be generated by an independent computing unit (e.g., Computer-on-Module (COM), single-board computer, etc.) that operates separately from and simultaneously with the thermal system executing thermal control of the IC using the control parameters and the test system executing the IC inspection (e.g., Automated Test Device (ATE), test program, etc.). This independent computing unit functions as an intermediate control point separating the test system and the thermal system, while enabling the operation of each system to influence the operation of the other system. Through the independent computing unit, the thermal system, the test system, and / or other components can also access the temperature related to the IC on an "on-demand" basis (e.g., without a "blind period" during which the temperature sensor is unavailable).

[0088] Temperature prediction and control may be performed across other stages of IC testing, such as wafer sort inspection, die-level inspection, burn-in inspection, final inspection, and / or system-level inspection (though not limited thereto). Data collected during previous inspection stages may be tracked and used to perform thermal prediction and control in subsequent stages of the same IC. Alternatively, this data may also be used to design, evaluate, and / or modify the die layout and / or package containing the IC, the test configuration for testing, and / or other parameters related to the inspection and / or use of the IC.

[0089] One technical advantage of the disclosed technology over the prior art is the ability to accurately and timely monitor and control the temperatures of various AOIs within the IC. As a result, the disclosed technology can avoid and / or mitigate thermal damage and / or destruction caused by the limitations of temperature monitoring by temperature sensors and / or the inability to respond quickly to sudden changes in heat generation during IC inspection. Furthermore, by incorporating thermal data collected during a previous inspection step into thermal prediction and adjustment during a subsequent inspection step involving the same IC, the disclosed technology can improve the design and / or thermal management of integrated systems of multiple components with complex ICs, packages, and / or different thermal responses. These technical advantages result in one or more technical improvements over the approach of the prior art.

[0090] The description of various embodiments is provided for illustrative purposes only, but is not intended to be comprehensive or limited to the disclosed embodiments. It will be obvious to those skilled in the art that various modifications and variations are possible within the scope and spirit of the described embodiments.

[0091] Clause 1. In various embodiments, a method executed by a computer for performing an inspection of an integrated circuit comprises generating one or more temperature predictions of a region of interest within the integrated circuit based on sensor data related to the integrated circuit, determining one or more control parameters related to the integrated circuit test based on one or more temperature predictions and one or more control characteristics related to the integrated circuit, and performing the integrated circuit test based on one or more control parameters.

[0092] Clause 2. A method executed by a computer as described in Clause 1, further comprising generating one or more additional temperature predictions and one or more additional control parameters based on the results of a test, and performing additional integrated circuit tests based on one or more additional control parameters.

[0093] Clause 3. A method performed by a computer as described in either Clause 1 or Clause 2, wherein the test is associated with a wafer or die associated with the integrated circuit, and additional testing is associated with a package associated with the integrated circuit.

[0094] Clause 4. A method of performing an integrated circuit test based on one or more control parameters, comprising controlling the temperature associated with the integrated circuit using one or more control parameters, performed by a computer as described in any one of Clauses 1 to 3.

[0095] Clause 5. A method executed by a computer as described in any one of Clauses 1 to 4, wherein one or more temperature predictions are generated by a first machine learning model and one or more control parameters are generated by a second machine learning model.

[0096] Clause 6. A method executed by a computer as described in any one of Clauses 1 to 5, wherein one or more temperature predictions and one or more control parameters are generated based on at least one of a machine learning model, a physics-based model, a control loop, a simulation, or an experiment.

[0097] Clause 7. A method executed by a computer as described in any one of Clauses 1 to 6, wherein one or more control parameters include at least one of a target temperature, a control signal of a thermal system related to an integrated circuit test, a control signal related to the pace of the test, or a control signal related to the start or stop of the test.

[0098] Clause 8. A method executed by a computer as described in any one of Clauses 1 to 7, wherein one or more control characteristics include at least one of a temperature limit, a maximum deviation from a temperature limit, a minimum deviation from a temperature limit, a temperature gradient, a temperature integral, a sum of temperature deviations, a power gradient, a safety offset, a distribution of power values, or one or more outliers.

[0099] Clause 9. A method executed by a computer as described in any one of Clauses 1 to 8, wherein sensor data includes at least one of temperature, power, aging degradation, current, voltage, or voltage drop.

[0100] Clause 10. A method executed by a computer as described in any one of Clauses 1 to 9, wherein one or more temperature predictions are further generated based on at least one of the physical characteristics of the integrated circuit, statistical information related to the integrated circuit, test results related to the integrated circuit, test parameters related to the test, or manufacturing data related to the integrated circuit.

[0101] Clause 11. In one or more embodiments, one or more non-transient computer-readable media store instructions that, when executed by one or more processors, instruct the one or more processors to execute the steps of: generating one or more temperature predictions of regions of interest within an integrated circuit based on sensor data related to the integrated circuit; determining one or more control parameters related to an integrated circuit test based on one or more temperature predictions and one or more control characteristics related to the integrated circuit; and adjusting one or more thermal control units related to the test based on one or more control parameters.

[0102] Clause 12. One or more non-transient computer-readable media described in Clause 11, wherein the instruction comprises: a step of determining (i) a set of Ground Truth (GT) temperatures associated with a region of interest and (ii) a set of Ground Truth (GT) control parameters associated with an integrated circuit; a step of training one or more machine learning models using the set of Ground Truth (GT) temperatures and the set of Ground Truth (GT) control parameters, wherein the one or more machine learning models further execute a step subsequently used to generate one or more temperature predictions or one or more control parameters.

[0103] Clause 13. One or more non-transient computer-readable media described in any one of Clauses 11 to 12, wherein training one or more machine learning models comprises generating a predicted temperature of a region of interest based on an input including training sensor data and training test program data related to an integrated circuit through the execution of a machine learning model included in one or more machine learning models, and training the machine learning model based on one or more losses calculated using a corresponding Ground Truth (GT) temperature included in a set of predicted temperatures and Ground Truth (GT) temperatures.

[0104] Clause 14. One or more non-transient computer-readable media described in any one of Clauses 11 to 13, wherein training one or more machine learning models comprises generating predictive control parameters based on inputs including predicted temperatures of a region of interest through the execution of a machine learning model included in one or more machine learning models, and training the machine learning models based on one or more losses calculated using the predicted control parameters and corresponding Ground Truth (GT) control parameters included in a set of Ground Truth (GT) control parameters.

[0105] Clause 15. One or more non-transient computer-readable media described in any one of Clauses 11 to 14, wherein the instruction further executes the step of generating one or more additional temperature predictions and one or more additional control parameters based on one or more temperature predictions and one or more control parameters, and the step of adjusting one or more thermal control units based on one or more additional control parameters, in one or more processors.

[0106] Clause 16. One or more thermal control units, one or more non-transient computer-readable media described in any one of Clauses 11 through 15, which are adjusted during integrated circuit testing or during subsequent testing.

[0107] Clause 17. One or more temperature predictions and one or more control parameters are generated based on at least one of a machine learning model, a physics-based model, a control loop, a simulation, or an experiment, one or more non-transient computer-readable media described in any one of Clauses 11 to 16.

[0108] Clause 18. One or more non-transient computer-readable media described in any one of Clauses 11 through 17, wherein sensor data is collected by one or more sensors positioned outside the region of interest.

[0109] Clause 19. One or more non-transient computer-readable media described in any one of Clauses 11 through 18, wherein the region of interest comprises a two-dimensional or three-dimensional region within an integrated circuit.

[0110] Clause 20. In one or more embodiments, the system comprises one or more memories for storing instructions, and one or more processors coupled to the one or more memories and configured to execute steps for generating one or more temperature predictions of a region of interest within an integrated circuit based on sensor data related to one or more locations within the integrated circuit outside the region of interest when executing instructions, steps for determining one or more control parameters related to an integrated circuit test based on one or more temperature predictions and one or more control characteristics related to the integrated circuit, and steps for executing an integrated circuit test based on one or more control parameters.

[0111] Any combination of the elements of the claim described in any one of the claims and / or any one of the elements described in the present application is included in the embodiments and the scope of protection in any form.

[0112] The description of various embodiments is provided for illustrative purposes only and is not intended to be comprehensive or to limit the scope to the disclosed embodiments. Various modifications and variations will be obvious to those skilled in the art without departing from the scope and spirit of the described embodiments.

[0113] Aspects of the present embodiment may be implemented as a system, method, or computer program product. Accordingly, aspects of the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware embodiments, all of which are generally referred to herein as "module," "system," or "computer." Additionally, any hardware and / or software technology, process, function, component, engine, module, or system described in the present disclosure may be implemented as a circuit or set of circuits. Additionally, aspects of the present disclosure may take the form of a computer program product implemented on one or more computer-readable media containing computer-readable program code.

[0114] Any combination of one or more computer-readable media may be used. A computer-readable medium may be a computer-readable signal medium or a computer-readable recording medium. A computer-readable recording medium may be, for example, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or any suitable combination thereof, but is not limited thereto. More specific examples (a non-exclusive list) of computer-readable recording media include electrical connections having one or more wires, portable computer diskettes, hard disks, RAM (Random Access Memory), Read Only Memory (Read Only Memory), Erasable Programmable Read Only Memory (EEPROM), or flash memory), optical fibers, portable CD-ROM (Compact Disc Read Only Memory), optical memory devices, magnetic memory devices, or any suitable combination thereof. In the context of this specification, a computer-readable recording medium may be any type of medium capable of containing or storing a program for use by or in connection with an instruction execution system, device, or device.

[0115] Aspects of the present disclosure have been described with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It will be understood that each block of the flowchart and / or block diagram, and combinations of blocks of the flowchart and / or block diagram, may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a dedicated computer, or other programmable data processing device to manufacture a machine. When the instructions are executed through a processor of a computer or other programmable data processing device, they enable the implementation of a specified function / operation in one or more blocks of the flowchart and / or block diagram. Such a processor may be, but is not limited to, a general-purpose processor, a dedicated processor, a specific-purpose processor, or a Field Programmable Gate Array (FPGA).

[0116] The flowcharts and block diagrams in the drawings represent the architecture, function, 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 of a flowchart or block diagram may represent a module, segment, or part of code containing one or more executable instructions for implementing a specified logical function. It should be noted that in some alternative embodiments, the functions described in the blocks may proceed in an order different from the order described in the drawings. For example, two blocks depicted consecutively may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order depending on the related functions. It should also be noted that each block of the block diagram and / or flowchart, and combinations of blocks of the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that executes the specified function or operation, or by a combination of dedicated hardware and computer instructions.

[0117] Although the above description focuses on the embodiments of the present disclosure, other embodiments and additional embodiments of the present disclosure may be devised without departing from the basic scope thereof, and the scope thereof is determined by the appended claims. Explanation of the symbols

[0118] 102: Processor 104: System Memory 105: Memory Bridge 106: Communication path 107: I / O Bridge 108: Input Devices 110: Display device 112: Parallel Processing Subsystem 113: Communication path 114: System Disk 116: Switch 118: Network adapter 120: ADD-IN CARD 121: ADD-IN CARD 122: Training Engine 124: Execution Engine 126: Management Engine 200: Training data 202: Training Input 204: Training GT (Ground Truth) values 206: Model Parameters 208: Machine learning models 210: Training Output 212: Data Store 214: Record 216: User Interface 220: Sort data 222: Die-level data 224: package-level data 226: Burn-in data 228: Final Test Data 230: System Level Data 232: Die layout 234: Test Configuration 236: Test Results 240: Input 242: Temperature Prediction 246: Control Parameters 248: Test output 254: Loss 302: Sensor data 304: Control characteristics 306: GT (Ground Truth) temperature 308: GT (Ground Truth) Control Parameters 402: Thermal System 404: Test System 406: DUT 408: Test Program 410: Thermal control unit 412: Sensor data

Claims

Claim 1 A method executed by a computer for conducting an inspection of an integrated circuit, comprising generating one or more temperature predictions of a region of interest within the integrated circuit based on sensor data related to the integrated circuit, determining one or more control parameters related to testing the integrated circuit based on the one or more temperature predictions and one or more control characteristics related to the integrated circuit, and conducting the test of the integrated circuit based on the one or more control parameters. Claim 2 A method according to claim 1, further comprising generating one or more additional temperature predictions and one or more additional control parameters based on the results of the test, and conducting additional tests of the integrated circuit based on the one or more additional control parameters. Claim 3 A method according to paragraph 2, wherein the test is associated with a wafer or die associated with the integrated circuit, and the additional test is associated with a package associated with the integrated circuit. Claim 4 A method according to claim 1, wherein conducting the test of the integrated circuit based on the one or a plurality of control parameters includes controlling the temperature associated with the integrated circuit using the one or a plurality of control parameters. Claim 5 A method according to claim 1, wherein the one or more temperature predictions are generated by a first machine learning model, and the one or more control parameters are generated by a second machine learning model. Claim 6 A method according to claim 1, wherein the one or more temperature predictions and the one or more control parameters are generated based on at least one of a machine learning model, a physics-based model, a control loop, a simulation, or an experiment. Claim 7 A method according to claim 1, wherein the one or more control parameters include at least one of a target temperature, a control signal of a thermal system related to the test of the integrated circuit, a control signal related to the pace of the test, or a control signal related to the start or stop of the test. Claim 8 A method according to claim 1, wherein the one or more control characteristics include at least one of a temperature limit, a maximum deviation from a temperature limit, a minimum deviation from a temperature limit, a temperature gradient, a temperature integral, a sum of temperature deviations, a power gradient, a safety offset, a distribution of power values, or one or more outliers. Claim 9 A method according to claim 1, wherein the sensor data comprises at least one of temperature, power, aging degradation, current, voltage, or voltage drop. Claim 10 A method according to claim 1, wherein the one or more temperature predictions are further generated based on at least one of the physical characteristics of the integrated circuit, statistical information related to the integrated circuit, test results related to the integrated circuit, test parameters related to the test, or manufacturing data related to the integrated circuit. Claim 11 One or more non-transient computer-readable media storing instructions for executing, when executed by one or more processors, a step of generating one or more temperature predictions of regions of interest within an integrated circuit based on sensor data related to the integrated circuit, said one or more processors; a step of determining one or more control parameters related to testing the integrated circuit based on said one or more temperature predictions and one or more control characteristics related to the integrated circuit; and a step of adjusting one or more thermal control units related to the test based on said one or more control parameters. Claim 12 In paragraph 11, the above instruction further executes the step of determining (i) a set of Ground Truth (GT) temperatures associated with the region of interest and (ii) a set of Ground Truth (GT) control parameters associated with the integrated circuit; and the step of training one or more machine learning models using the set of Ground Truth (GT) temperatures and the set of Ground Truth (GT) control parameters, wherein the one or more machine learning models are subsequently used to generate the one or more temperature predictions or the one or more control parameters, one or more non-transient computer-readable media. Claim 13 In claim 12, one or more non-transient computer-readable media comprising training the one or more machine learning models, generating a predicted temperature of the region of interest based on an input including training sensor data and training test program data related to the integrated circuit through the execution of the machine learning model included in the one or more machine learning models, and training the machine learning model based on one or more losses calculated using the predicted temperature and a corresponding Ground Truth (GT) temperature included in the set of Ground Truth (GT) temperatures. Claim 14 In claim 12, one or more non-transient computer-readable media comprising training the one or more machine learning models, generating predictive control parameters based on an input including the predicted temperature of the region of interest through the execution of the machine learning models included in the one or more machine learning models, and training the machine learning models based on one or more losses calculated using the predicted control parameters and corresponding Ground Truth (GT) control parameters included in the set of Ground Truth (GT) control parameters. Claim 15 In paragraph 11, the above command further executes the step of generating one or more additional temperature predictions and one or more additional control parameters based on the one or more temperature predictions and one or more control parameters in the one or more processors; and the step of adjusting the one or more thermal control units based on the one or more additional control parameters, in one or more non-transient computer-readable media. Claim 16 In paragraph 15, the one or more thermal control units are one or more non-transient computer-readable media that are adjusted during the test of the integrated circuit or during a subsequent test. Claim 17 In claim 11, the one or more temperature predictions and the one or more control parameters are one or more non-transient computer-readable media generated based on at least one of a machine learning model, a physics-based model, a control loop, a simulation, or an experiment. Claim 18 In paragraph 11, the sensor data is collected by one or more non-transient computer-readable media placed outside the region of interest. Claim 19 In claim 11, the region of interest comprises one or more non-transient computer-readable media, the region of interest comprising a two-dimensional or three-dimensional region within the integrated circuit. Claim 20 A system comprising: one or more memories for storing instructions; and one or more processors coupled to the one or more memories, wherein the one or more processors are configured to execute the following steps when executing the instructions: a step of generating one or more temperature predictions of a region of interest within an integrated circuit based on sensor data related to one or more locations within the integrated circuit outside the region of interest; a step of determining one or more control parameters related to testing the integrated circuit based on the one or more temperature predictions and one or more control characteristics related to the integrated circuit; and a step of performing the test of the integrated circuit based on the one or more control parameters.

Images