A three-dimensional chip packaging method and system

Through thermal simulation and thermal perception technology, the chip heat distribution is identified, the frequency voltage and thread allocation are dynamically adjusted, and the thermal interface material is introduced, which solves the problem of insufficient heat dissipation in chip packaging and improves the heat dissipation efficiency and chip life.

CN118747438BActive Publication Date: 2024-11-19DONGGUAN HUAHUI ELECTRONICS SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410943534.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-15
Publication Date
2024-11-19
Estimated Expiration
2044-07-15

AI Technical Summary

Technical Problem

Chip stacking during chip packaging increases power density, resulting in more serious heat dissipation problems.

Method used

Identify the heat distribution data of the packaged stacked chip through thermal simulation tools, dynamically adjust the frequency voltage of the chip, apply thermal sensing thread scheduling algorithm and thermal sensing compiler, optimize the thread allocation and data storage location of the chip, introduce thermal interface materials, and adjust the internal of the chip Structural layout to improve heat dissipation efficiency.

Benefits of technology

The refined analysis of heat distribution is achieved, which avoids local overheating, improves overall heat dissipation efficiency, reduces heat accumulation caused by data access, and extends the service life of the chip.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118747438B_ABST
    Figure CN118747438B_ABST
Patent Text Reader

Abstract

The present invention provides a three-dimensional chip packaging method and system, which are applied to the field of chip data processing; the present invention uses a thermal simulation tool to identify the heat distribution data of packaged stacked chips under different load conditions, especially the hot spot location area and the insufficient heat dissipation area, to achieve refined analysis of heat distribution, and at the same time applies a thermal-aware thread scheduling algorithm to dynamically adjust the thread allocation of the chip according to the heat distribution to avoid local overheating and improve the overall heat dissipation efficiency, and dynamically adjust the storage location of data between different levels of chip cache to improve data access efficiency and reduce heat accumulation caused by data access.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of chip data processing, and in particular to a three-dimensional chip packaging method and system. Background Art

[0002] As the functions of electronic devices become increasingly complex, traditional chip packaging methods can no longer meet the requirements of high performance, low power consumption and small size. In order to integrate more functional modules in a limited space, improve system performance and reduce power consumption, three-dimensional packaging technology has emerged. However, chip stacking during chip packaging increases power consumption density, making the heat dissipation problem more serious. Summary of the invention

[0003] The present invention aims to solve the problem that chip stacking during chip packaging increases power consumption density, resulting in more severe heat dissipation, and provides a three-dimensional chip packaging method and system.

[0004] The present invention adopts the following technical means to solve the technical problem:

[0005] The present invention provides a three-dimensional chip packaging method, comprising:

[0006] Based on the thermal simulation content preset by the processing terminal for the packaged stacked chip, the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content is identified, and the thermal area of ​​the packaged stacked chip is collected according to the heat distribution data, wherein the thermal area specifically includes a hot spot location area and an insufficient heat dissipation area;

[0007] determining whether the heat area exceeds a preset upper limit of the processing terminal;

[0008] If so, the packaged stacked chips are unpacked, a thermal model of each chip is constructed from the processing terminal, the packaged thermal distribution data of the chip is identified based on the thermal model, the heat conduction path and the air circulation direction of the chip are obtained, the heat accumulation area of ​​the chip is identified through the heat conduction path and the air circulation direction, a preset thermal interface material is introduced into the heat accumulation area, the temperature data of the chip is monitored in real time, and the preset frequency voltage of the chip is dynamically adjusted based on the temperature data, wherein the packaged thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink;

[0009] Determine whether the frequency voltage can execute the calculation test task preset by the processing terminal;

[0010] If not, the computing load of the computing test task is evenly distributed, and the thread allocation of the chip is dynamically adjusted by applying a preset thermal-aware thread scheduling algorithm. The execution rate of the hot code block is identified based on the pre-developed thermal-aware compiler, and the execution order and execution position of the hot code block are balanced during the compilation process. According to the data access efficiency and access heat of the chip, the storage location of the data between different levels of cache of the chip is dynamically adjusted to complete the test process of the chip, and the chip is repackaged to obtain the packaged stacked chip.

[0011] Furthermore, the step of constructing a thermal model of each chip from the processing terminal and identifying the package thermal distribution data of the chip according to the thermal model includes:

[0012] Establishing a geometric model of the chip and its package based on a preset tool of the processing terminal, collecting parameter settings of the chip, performing a test according to the parameter settings to obtain power consumption distribution data of the chip, and marking the regional power density of the chip from the geometric model, wherein the parameter settings specifically include chip levels, thermophysical parameters, and heat dissipation boundary conditions;

[0013] Determining whether the regional power density matches the preset density of the processing terminal;

[0014] If so, the image data of the chip and its package are generated by the preset tool, the distribution information of the chip in different areas is displayed from the image data, and the thermal resistance between the chip and the packaging material is calculated based on the distribution information, wherein the image data specifically includes a temperature distribution map and a heat flux distribution map, and the distribution information specifically includes a temperature gradient and a heat flux density direction.

[0015] Furthermore, before the step of identifying the heat accumulation area of ​​the chip through the heat conduction path and the air circulation direction and introducing a preset thermal interface material on the heat accumulation area, the step further includes:

[0016] Based on the preset imaging parameters of the chip by the processing terminal, the preset thermal imaging is used to dynamically monitor the operation process of the chip to capture the real-time temperature change and heat accumulation trend of the chip, wherein the imaging parameters specifically include imaging angle and imaging distance, and the operation process specifically includes idle, low load, high load and peak load;

[0017] Determining whether the operating time of the chip can reach a preset time period;

[0018] If not, the hotspot position and range in the chip are identified through high-resolution temperature measurement and thermal imaging preset at the processing terminal, and the internal structural layout of the chip is readjusted according to the hotspot distribution of the chip. The thermal conductivity of each packaging material is evaluated on the processing terminal according to the structural layout, and heat dissipation paths with different efficiencies are constructed by combining the various packaging materials.

[0019] Furthermore, the step of identifying the execution rate of the hot code block based on the pre-developed thermal-aware compiler and balancing the execution order and execution position of the hot code block during the compilation process includes:

[0020] Constructing a program control flow graph of the chip through the processing terminal, and identifying a hot loop of the hot code block from the program control flow graph;

[0021] Determining whether the hotspot loop exceeds a preset loop control overhead;

[0022] If so, the position and execution frequency of the hot code block are collected, the code order is rearranged based on the position and the execution frequency, and an instruction set that supports thermal-aware optimization is used during the compilation process of the hot code block, and high-power instructions are replaced with low-power equivalent instructions according to the instruction set.

[0023] Furthermore, the step of determining whether the heat area exceeds a preset upper limit of the processing terminal further includes:

[0024] Based on the preset time segment of the processing terminal, the temperature data points corresponding to the initial time node are recorded, and the weighted average value of the temperature data points is calculated according to the preset weight distribution;

[0025] Determining whether the weighted average matches a preset moving window;

[0026] If so, move the moving window to the right by a single temperature data point, intercept the next time node from the preset time segment, remove a single temperature data point to the left in the initial time node, add the next time node to the initial time node to form iterative temperature data, apply the weight distribution to calculate the moving average of the iterative temperature data in sequence, until the next time node can no longer be intercepted in the preset time segment, and integrate to obtain the temperature trend data of the preset time segment.

[0027] Furthermore, the step of determining whether the frequency voltage can execute the calculation test task preset by the processing terminal further includes:

[0028] Based on the load type preset in the computing test task, the computing test task is run by the chip in the simulation environment preset in the processing terminal to collect simulation data, wherein the load type specifically includes CPU intensive, memory intensive, and I / O intensive;

[0029] Determining whether the simulation data matches a preset data format;

[0030] If so, the simulation data is input into a pre-built energy consumption model, the power consumption of the chip under different frequency and voltage configurations is predicted from the energy consumption model, and the energy consumption model and the thermal model are combined to simulate the chip temperature changes under different configurations.

[0031] Further, the step of identifying the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content and collecting the thermal area of ​​the packaged stacked chip according to the heat distribution data includes:

[0032] Using a visualization tool preset in the processing terminal to visualize the heat distribution data and generate a temperature field image;

[0033] Determining whether the temperature field image completely presents the packaged stacked chip;

[0034] If so, the internal temperature distribution of the packaged stacked chip is identified in the temperature field image, high temperature areas and potential hot spots are identified, and the structure of the packaged stacked chip is simulated from the temperature field image to simulate the heat dissipation effect corresponding to the preset heat dissipation structure.

[0035] The present invention also provides a three-dimensional chip packaging system, comprising:

[0036] A collection module, for identifying heat distribution data of the packaged stacked chip under a load condition corresponding to the thermal simulation content preset by the processing terminal for the packaged stacked chip, and collecting a heat area of ​​the packaged stacked chip according to the heat distribution data, wherein the heat area specifically includes a hot spot location area and an insufficient heat dissipation area;

[0037] A judging module, used for judging whether the heat area exceeds a preset upper limit of the processing terminal;

[0038] an execution module, for, if yes, unpacking the packaged stacked chips, constructing a thermal model of each chip from the processing terminal, identifying the packaged thermal distribution data of the chip according to the thermal model, obtaining the thermal conduction path and the air circulation direction of the chip, identifying the heat accumulation area of ​​the chip through the thermal conduction path and the air circulation direction, introducing a preset thermal interface material on the heat accumulation area, monitoring the temperature data of the chip in real time, and dynamically adjusting the preset frequency voltage of the chip based on the temperature data, wherein the packaged thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink;

[0039] A second judgment module is used to judge whether the frequency voltage can execute the calculation test task preset by the processing terminal;

[0040] The second execution module is used to evenly distribute the computing load of the computing test task if not, apply a preset thermal-aware thread scheduling algorithm to dynamically adjust the thread allocation of the chip, identify the execution rate of the hot code block based on the pre-developed thermal-aware compiler, balance the execution order and execution position of the hot code block during the compilation process, dynamically adjust the storage location of the data between different levels of cache on the chip according to the data access efficiency and access heat of the chip, complete the test process of the chip, and re-package the chip to obtain the packaged stacked chip.

[0041] Furthermore, the execution module also includes:

[0042] a marking unit, used to establish a geometric model of the chip and its package based on a preset tool of the processing terminal, collect parameter settings of the chip, perform a test according to the parameter settings to obtain power consumption distribution data of the chip, and mark the regional power density of the chip from the geometric model, wherein the parameter settings specifically include chip levels, thermophysical parameters, and heat dissipation boundary conditions;

[0043] A determination unit, configured to determine whether the regional power density matches a preset density of the processing terminal;

[0044] An execution unit is used to generate image data of the chip and its package through the preset tool, display distribution information of the chip in different areas from the image data, and calculate the thermal resistance between the chip and the packaging material based on the distribution information, wherein the image data specifically includes a temperature distribution map and a heat flux distribution map, and the distribution information specifically includes a temperature gradient and a heat flux density direction.

[0045] Furthermore, it also includes:

[0046] A capture module, used for dynamically monitoring the operation process of the chip by applying preset thermal imaging based on the imaging parameters preset by the processing terminal for the chip, and capturing the real-time temperature change and heat accumulation trend of the chip, wherein the imaging parameters specifically include imaging angle and imaging distance, and the operation process specifically includes idle, low load, high load and peak load;

[0047] A third judgment module is used to judge whether the operation time of the chip can reach a preset time period;

[0048] The third execution module is used to, if not, identify the hotspot position and range in the chip through the high-resolution temperature measurement preset by the processing terminal and the thermal imaging, readjust the internal structural layout of the chip according to the hotspot distribution of the chip, evaluate the thermal conductivity of each packaging material on the processing terminal according to the structural layout, and construct heat dissipation paths with different efficiencies by combining the various packaging materials.

[0049] The present invention provides a three-dimensional chip packaging method and system, which has the following beneficial effects:

[0050] The present invention uses a thermal simulation tool to identify the heat distribution data of packaged stacked chips under different load conditions, especially the hot spot location area and the insufficient heat dissipation area, to achieve refined analysis of the heat distribution, and at the same time applies a thermal-aware thread scheduling algorithm to dynamically adjust the thread allocation of the chip according to the heat distribution to avoid local overheating and improve the overall heat dissipation efficiency; and dynamically adjust the storage location of data between different levels of chip cache to improve data access efficiency and reduce heat accumulation caused by data access. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a process flow of an embodiment of a three-dimensional chip packaging method of the present invention;

[0052] Figure 2 The structure block diagram of an embodiment of the three-dimensional packaging system of the chip of the present invention. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. The implementation of the objectives, functional features and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Reference Figure 1 , is a three-dimensional chip packaging method in an embodiment of the present invention, comprising:

[0056] S1: Based on the thermal simulation content preset by the processing terminal for the packaged stacked chip, identifying the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content, and collecting the thermal area of ​​the packaged stacked chip according to the heat distribution data, wherein the thermal area specifically includes a hot spot location area and an insufficient heat dissipation area;

[0057] S2: Determine whether the heat area exceeds a preset upper limit of the processing terminal;

[0058] S3: If so, unpack the packaged stacked chips, construct a thermal model of each chip from the processing terminal, identify the packaged thermal distribution data of the chip based on the thermal model, obtain the thermal conduction path and air circulation direction of the chip, identify the heat accumulation area of ​​the chip through the thermal conduction path and the air circulation direction, introduce a preset thermal interface material on the heat accumulation area, monitor the temperature data of the chip in real time, and dynamically adjust the preset frequency voltage of the chip based on the temperature data, wherein the packaged thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink;

[0059] S4: Determine whether the frequency voltage can execute the calculation test task preset by the processing terminal;

[0060] S5: If not, the computing load of the computing test task is evenly distributed, and the thread allocation of the chip is dynamically adjusted by applying a preset thermal-aware thread scheduling algorithm. The execution rate of the hot code block is identified based on a pre-developed thermal-aware compiler, and the execution order and execution position of the hot code block are balanced during the compilation process. According to the data access efficiency and access heat of the chip, the storage location of the data between different levels of cache of the chip is dynamically adjusted to complete the test process of the chip, and the chip is repackaged to obtain the packaged stacked chip.

[0061] In this embodiment, the system identifies the heat distribution data of the packaged stacked chips under the load conditions corresponding to the thermal simulation content based on the thermal simulation content pre-set by the processing terminal for the packaged stacked chips, and collects the thermal areas of the packaged stacked chips according to different heat distribution data. The thermal areas specifically include hot spot location areas and insufficient heat dissipation areas. Then the system determines whether these thermal areas exceed the area upper limit pre-set by the processing terminal to execute corresponding steps; for example, when the system determines that the thermal area of ​​the packaged stacked chips does not exceed the area upper limit pre-set by the processing terminal, the system will consider that the temperature of the chip is within a safe range and there is no overheating, and the system will continue to use the current Packaging and heat dissipation design saves the current thermal simulation and temperature monitoring data as baseline data for possible future analysis and optimization. At the same time, the normal operation and task execution of the packaged stacked chips continue, and the chip temperature is monitored in real time to ensure that the temperature is still within the safe range under different workload conditions. Although the current temperature does not exceed the upper limit, it is still necessary to regularly check and monitor the heat distribution to prevent future load increases or environmental changes from causing temperature to exceed the standard. For example, when the system determines that the thermal area of ​​the packaged stacked chip exceeds the upper limit of the area pre-set by the processing terminal, the system will consider that the chip temperature is not within the safe range and is prone to overheating. The system will Unpack the stacked chips, build different thermal models corresponding to different types of chips from the processing terminal, identify the chip packaging thermal distribution data based on the thermal model, the packaging thermal distribution data specifically includes the thermal interface between the chip layers, the thermal conductivity of the packaging material and the design of the heat sink, obtain the chip's thermal conduction path and air circulation direction, identify the chip's heat accumulation area through different thermal conduction paths and air circulation directions, introduce pre-set thermal interface materials in different heat accumulation areas, monitor the chip's temperature data in real time, and dynamically adjust the chip's pre-set frequency and voltage based on different temperature data; the system accurately identifies different heat accumulation areas through thermal models and simulation analysis Accumulation area, can take targeted measures to optimize the heat dissipation design, introduce preset thermal interface materials in the heat accumulation area, improve heat conduction efficiency, reduce thermal resistance, and improve overall heat dissipation performance. At the same time, based on real-time temperature data, dynamically adjust the frequency and voltage of the chip, reduce power consumption and heat generation, and maintain the system running within a safe temperature range. It can promptly detect and respond to abnormal temperature conditions to prevent overheating, and based on thermal models and simulation data, conduct preventive thermal design and optimization to improve heat dissipation capabilities and reserve margins for possible future load changes. The system then determines whether the frequency and voltage of the chip can execute the calculation and test tasks pre-set by the processing terminal to execute the corresponding steps.For example, when the system determines that the frequency and voltage of the chip can execute the calculation and test tasks pre-set by the processing terminal, the system will believe that the current frequency and voltage settings ensure that the chip can efficiently complete the calculation and test tasks while maintaining a safe temperature range to ensure the reliability and stability of the system. The system will continue to execute the preset calculation and test tasks to ensure the normal progress of the workflow, while monitoring the temperature and working status of the chip in real time to ensure that no abnormalities occur during the task execution, and record the temperature, frequency, voltage and task execution data under the current operating conditions in detail to provide a basis for subsequent analysis and optimization; for example, when the system determines that the frequency and voltage of the chip cannot execute the calculation and test tasks pre-set by the processing terminal, the system will believe that it is difficult for the chip to efficiently complete the calculation and test tasks. The system will evenly distribute the calculation load for the calculation and test tasks that the chip needs to process, apply a pre-set thermal-aware thread scheduling algorithm to dynamically adjust the chip's thread allocation, and identify the execution rate of hot code blocks based on the pre-developed thermal-aware compiler. During the compilation process, the execution order and execution position of hot code blocks are balanced. According to the data access efficiency and access heat of the chip, the storage position of data between different levels of chip cache is dynamically adjusted to complete the chip test process, and the chip is repackaged to obtain a packaged stacked chip. The system evenly distributes the computing load to avoid overload of certain areas or threads, reduce the heat accumulation in the hot area, improve the overall computing efficiency, eliminate the performance bottleneck caused by load imbalance, make all parts of the chip work evenly, and make full use of chip resources. At the same time, the thermal-aware thread scheduling algorithm is applied to dynamically adjust the thread allocation of the chip according to real-time temperature data to avoid overheating in the hot area and improve the heat dissipation efficiency. By dynamically adjusting the thread allocation, the heat distribution is more uniform, the risk of local overheating is reduced, and the service life of the chip is extended. The perceptual compiler identifies and optimizes the execution order and execution position of the hot code blocks, reduces the heat accumulation in the hot area, balances the execution order of the hot code blocks during the compilation process, optimizes the code execution path, reduces the temperature rise rate, and improves the system stability. ;

[0062] It should be noted that a specific example of constructing a thermal model of each chip from the processing terminal and identifying the package thermal distribution data of the chip based on the thermal model is as follows:

[0063] Assume that multi-core processor chips are prone to local overheating under high load, affecting performance and reliability;

[0064] The system first needs to build a thermal model. Since the processor chip is made of silicon material, its thermal conductivity is 148 W / m·K, specific heat capacity is 700 J / kg·K, density is 2330 kg / m³, and the chip size is 20 mm x 20 mm, thickness is 0.5 mm, and the package type is BGA, so to establish the internal heat conduction model of the chip, it is necessary to consider the thermal conductivity and interface thermal resistance of each layer of material; that is, the convection heat transfer coefficient between the chip surface and the air is 25 W / m²·K, and the surface emissivity is 0.7; by using ANSYS for thermal simulation, setting the processor to run at 100% load, simulating the temperature distribution inside the chip, according to the simulation results, the temperature in the central area of ​​the chip is the highest, reaching 95°C, while the temperature in the edge area is lower, about 75°C. The center of the chip is identified as a hot spot; therefore, the system needs to introduce high-conductivity thermal grease between the central area of ​​the chip and the heat sink to reduce thermal resistance, and by increasing the area of ​​the heat sink and optimizing the fan layout, improve air circulation efficiency and accelerate heat dissipation.

[0065] A specific example of applying a preset thermal-aware thread scheduling algorithm to dynamically adjust thread allocation of the chip is as follows:

[0066] Assume that a multi-core processor has 8 cores. Under high load, some cores are prone to overheating, affecting overall performance and reliability.

[0067] First, the system collects real-time temperature data. Since each core has a built-in temperature sensor, the core temperature can be monitored in real time. The temperature monitoring system collects the temperature data of each core once a second, and then calculates the thermal distribution state. The core temperature threshold is set to 80°C. If the temperature exceeds this, it is a hot spot. By generating a real-time thermal distribution map, it is known that the current thermal distribution map shows that the temperature of cores 1 and 3 is close to 80°C, and the temperature of other cores is below 60°C. Then, the thread allocation is dynamically adjusted. By reducing the priority of threads on cores 1 and 3, new tasks are assigned to cores with lower temperatures first, and some cores 1 and 3 are assigned to hot spots. The threads are allocated to cores 2, 4, and 5 to ensure load balancing of each core. Then, tasks are migrated to migrate high-load threads on cores 1 and 3 to cores 2 and 4 to reduce the temperature of cores 1 and 3. The load change within the next minute is predicted based on the current load and historical data, and some low-priority tasks are allocated to cores 6 and 7 with lower temperatures in advance. Finally, the adjusted system performance and temperature changes of each core are monitored to ensure load balancing and thermal management effects. Based on real-time data and feedback, the thermal-aware thread scheduling algorithm is optimized to improve the response speed of thread migration and load balancing.

[0068] In summary, by applying the thermal-aware thread scheduling algorithm, the system can dynamically adjust the thread allocation of the chip to balance the computing load and heat distribution. The specific steps include collecting real-time temperature data, calculating the thermal distribution status, dynamically adjusting thread allocation, task migration, and optimization and feedback. These measures can effectively prevent overheating of hot spots and improve overall system performance and reliability.

[0069] It should be added that the thermal simulation content specifically includes geometric structure, material properties and power consumption distribution; the geometric structure includes the specific size, shape, number of layers and packaging type of the chip; material properties: material properties include thermal conductivity, specific heat capacity, density and thermal expansion coefficient of each layer of the chip; power consumption distribution includes the power consumption distribution of each functional unit (such as processor core, cache, I / O unit, etc.) under different workloads.

[0070] Thermal interface materials specifically include thermal paste, thermal adhesive, thermal gasket, phase change material, thermal tape and liquid metal.

[0071] In this embodiment, the step S3 of constructing a thermal model of each chip from the processing terminal and identifying the package thermal distribution data of the chip according to the thermal model includes:

[0072] S31: establishing a geometric model of the chip and its package based on a preset tool of the processing terminal, collecting parameter settings of the chip, performing a test according to the parameter settings to obtain power consumption distribution data of the chip, and marking the regional power density of the chip from the geometric model, wherein the parameter settings specifically include chip levels, thermophysical parameters, and heat dissipation boundary conditions;

[0073] S32: Determine whether the regional power density matches the preset density of the processing terminal;

[0074] S33: If so, the image data of the chip and its package are generated by the preset tool, the distribution information of the chip in different areas is displayed from the image data, and the thermal resistance between the chip and the packaging material is calculated based on the distribution information, wherein the image data specifically includes a temperature distribution map and a heat flux distribution map, and the distribution information specifically includes a temperature gradient and a heat flux density direction.

[0075] In this embodiment, the system establishes a geometric model of the chip and its package based on a drawing tool pre-set in the processing terminal, collects parameter settings of the chip, and the parameter settings specifically include chip hierarchy, thermal physical parameters and heat dissipation boundary conditions. According to these parameter settings, tests are performed to obtain the power consumption distribution data of the chip, and the regional power density of the chip is marked from the geometric model. Then, the system determines whether these regional power densities match the density pre-set by the processing terminal to execute the corresponding steps; for example, when the system determines that the regional power density of the chip cannot match the density pre-set by the processing terminal, the system will consider that the power consumption distribution of the chip in actual work is inconsistent with the design expectation, which may cause local overheating or uneven heat dissipation. The system will use thermal simulation tools and power consumption distribution data to determine the power consumption concentration and hot spot areas, adjust the position of the chip functional units, reduce the concentration of high power consumption units, and evenly distribute the power consumption. At the same time, high-quality, high-thermal conductivity materials are selected to reduce manufacturing deviations, optimize manufacturing process parameters, ensure the consistency and uniformity of materials, and monitor the various chip components in real time. The temperature of the area is measured by lowering the frequency and voltage in the hot spot area to reduce power consumption and heat generation. For example, when the system determines that the regional power density of the chip can match the density pre-set by the processing terminal, the system will consider that the power consumption distribution of the chip in actual work is consistent with the design expectation, and the system will generate image data of the chip and its package through a pre-set drawing tool. The image data specifically includes a temperature distribution map and a heat flux distribution map. The distribution information of the chip in different areas is displayed from the image data. The distribution information specifically includes the temperature gradient and the direction of the heat flux density. The thermal resistance between the chip and the packaging material is calculated based on this distribution information. The system provides an intuitive display of the temperature gradient and the direction of the heat flux density in each area of ​​the chip through the generated temperature distribution map and heat flux distribution map, which helps to understand the current status of thermal management. At the same time, the thermal resistance between the chip and the packaging material is calculated based on the distribution information, which helps to optimize the selection and use of thermal interface materials and improve the heat conduction efficiency. In addition, the information on the temperature gradient and the direction of the heat flux density can guide the optimization of the radiator design, such as the arrangement of the heat sink and the installation position of the fan.

[0076] It should be noted that a specific example of calculating the thermal resistance between the chip and the packaging material based on the distribution information is as follows:

[0077] Assume that the system needs to calculate the thermal resistance between the smartphone processor and the heat sink. The system first needs to collect data, obtain the temperature distribution map of the smartphone processor and the heat sink from the thermal simulation software, obtain the heat flux distribution map of the interface between the processor and the heat sink, identify the central area of ​​the processor as the hot spot, and then select multiple measurement points on the contact surface between the processor surface and the heat sink to calculate the temperature gradient. Since the temperature difference between the processor surface and the heat sink surface is 10°C and the distance is 1mm, the temperature gradient = 10°C / 0.001m=10000°C / m, then calculate the heat flux density. From the heat flux distribution map, the heat flux density is 5000 W / m². Finally, calculate the thermal resistance. Thermal resistance = temperature gradient / heat flux density, that is, thermal resistance = 10000°C / m / 5000 W / m²=2°C / m / W.

[0078] In summary, the system calculates the thermal resistance between the chip and the packaging material in detail. This process not only helps identify weak links in thermal management, but also provides a scientific basis for optimizing thermal design. For example, in the thermal resistance calculation between the smartphone processor and the heat sink, the temperature gradient and heat flux density are determined to obtain an accurate thermal resistance value, which is crucial for designing a more efficient thermal solution.

[0079] In this embodiment, the heat accumulation area of ​​the chip is identified by the heat conduction path and the air flow direction, and before the step S3 of introducing a preset thermal interface material on the heat accumulation area, the method further includes:

[0080] S301: Based on the preset imaging parameters of the chip by the processing terminal, the preset thermal imaging is used to dynamically monitor the operation process of the chip to capture the real-time temperature change and heat accumulation trend of the chip, wherein the imaging parameters specifically include an imaging angle and an imaging distance, and the operation process specifically includes idle, low load, high load and peak load;

[0081] S302: Determine whether the operating time of the chip reaches a preset time period;

[0082] S303: If not, then identify the hotspot position and range in the chip through the high-resolution temperature measurement preset by the processing terminal and the thermal imaging, readjust the internal structural layout of the chip according to the hotspot distribution of the chip, evaluate the thermal conductivity of each packaging material on the processing terminal according to the structural layout, and construct heat dissipation paths with different efficiencies by combining the various packaging materials.

[0083] In this embodiment, the system uses pre-set imaging parameters for the chip based on the processing terminal, and the imaging parameters specifically include imaging angle and imaging distance. The system uses pre-set thermal imaging to dynamically monitor the operation process of the chip. The operation process specifically includes idle, low load, high load and peak load, and captures the real-time temperature change and heat accumulation trend of the chip. Then the system determines whether the operation time of the chip can reach the pre-set time period to execute the corresponding steps; for example, when the system determines that the operation time of the chip can reach the pre-set time period, the system will consider that the temperature change and heat accumulation of the chip during the entire preset operation cycle are acceptable and controllable. The system will select the best imaging angle and distance to obtain a clear thermal imaging image, record the temperature distribution of the chip in the idle state, observe whether there is basic heat accumulation, and record the temperature change data and heat accumulation trend of the chip in each operating state, generate corresponding thermal imaging images and data logs, and when the operation time of the chip reaches or exceeds the pre-set time period, it means that the temperature and heat distribution of the chip are good during the entire operation cycle, the thermal management design is effective, and the next stage of testing or mass production can be continued; Example For example, when the system determines that the operating time of the chip cannot reach the preset period, the system will consider that the temperature change and heat accumulation of the chip cannot be controlled during the entire preset operating cycle. The system will identify the hotspot position and hotspot range in the chip through the high-resolution temperature measurement and thermal imaging preset by the processing terminal, readjust the internal structure layout of the chip according to the hotspot distribution of the chip, evaluate the thermal conductivity of various packaging materials on the processing terminal according to the new structure layout, and construct heat dissipation paths with different efficiencies by combining various packaging materials; the system reduces the concentration of hotspots by redesigning the chip layout and optimizing the packaging materials, thereby improving the overall heat dissipation efficiency, dispersing high-power modules, evenly distributing heat, and avoiding local overheating problems. At the same time, through effective thermal management, the temperature peak of the chip during operation is reduced to avoid the impact of excessive temperature on chip performance and life. Good heat dissipation management reduces the damage of thermal stress to chip materials and structures, thereby extending the service life of the chip, and ensuring that the chip remains stable under high load and long-term operation, improving overall performance and user experience, optimizing the heat dissipation path, reducing the increase in power consumption caused by overheating, and improving energy utilization efficiency.

[0084] In this embodiment, based on the pre-developed thermal-aware compiler, the execution rate of the hot code block is identified, and the step S5 of balancing the execution order and execution position of the hot code block during the compilation process includes:

[0085] S51: constructing a program control flow graph of the chip through the processing terminal, and identifying a hot loop of the hot code block from the program control flow graph;

[0086] S52: Determine whether the hotspot loop exceeds a preset loop control overhead;

[0087] S53: If so, collect the position and execution frequency of the hot code block, rearrange the code order based on the position and the execution frequency, use an instruction set that supports thermal-aware optimization during the compilation process of the hot code block, and replace high-power instructions with low-power equivalent instructions according to the instruction set.

[0088] In this embodiment, the system pre-builds the program control flow graph of the chip through the processing terminal, identifies the hot loops of the hot code blocks from the program control flow graph, and then determines whether these hot loops exceed the preset loop control overhead to execute the corresponding steps; for example, when the system determines that the hot loop does not exceed the preset loop control overhead, the system will consider that the execution efficiency and resource consumption of the hot code block are within the controllable range, and the design of the hot loop is reasonable. The system will use the compiler tool chain to generate the control flow graph of the program, display the call relationship and execution path of each function, code block and loop, analyze the control flow graph to find the critical path and potential performance bottlenecks, and use the performance analysis tool to perform performance analysis on the program to find the code blocks and loops with the longest and most frequent execution time, mark these hot code blocks and loops in the control flow graph for further analysis and optimization, and pre-set the control overhead standards of the hot loops, such as CPU cycles, memory access times and cache hit rate, and evaluate whether the overhead of the identified hot loops is within an acceptable range according to the above standards; for example When the system determines that the hotspot loop exceeds the pre-set loop control overhead, the system will consider that the execution efficiency and resource consumption of the hotspot code block are difficult to control. The system will collect the location and execution frequency of the hotspot code block, rearrange the code order based on different locations and different execution frequencies, and use the instruction set that supports thermal-aware optimization during the compilation of the hotspot code block. According to different instruction sets, high-power instructions are replaced with equivalent low-power instructions; the system rearranges the code order, optimizes the memory layout, reduces the miss rate of the instruction cache and data cache, and improves the code execution efficiency. The optimized code order and low-power instruction set make the hotspot code block execute faster and improve the overall performance. At the same time, high-power instructions are replaced with low-power instructions, which effectively reduces the energy consumption of the chip, extends the battery life and reduces the heat dissipation requirements. The optimized instruction set and code layout make the system more energy-efficient, which helps to reduce the carbon footprint and operating cost of the system, and reduces the operating temperature of the chip by reducing power consumption and optimizing thermal management, improving the stability and reliability of the system. The lowered temperature reduces the impact of thermal stress on chip materials and extends the service life of the chip.

[0089] In this embodiment, the step S2 of determining whether the hot area exceeds the preset upper limit of the processing terminal further includes:

[0090] S21: Based on the preset time segment of the processing terminal, record the temperature data points corresponding to the initial time node, and calculate the weighted average value of the temperature data points according to the preset weight distribution;

[0091] S22: Determine whether the weighted average value matches a preset moving window;

[0092] S23: If so, move the moving window to the right by a single temperature data point, intercept the next time node from the preset time segment, remove a single temperature data point to the left in the initial time node, add the next time node to the initial time node to form iterative temperature data, apply the weight distribution to calculate the moving average of the iterative temperature data in sequence, until the next time node can no longer be intercepted in the preset time segment, and integrate to obtain the temperature trend data of the preset time segment.

[0093] In this embodiment, the system records the temperature data points corresponding to the initial time nodes based on the time segments preset by the processing terminal, calculates the weighted average of the temperature data points according to the preset weight distribution, and then the system determines whether the weighted average matches the preset moving window to execute the corresponding steps; for example, when the system determines that the weighted average cannot match the preset moving window, the system will consider that the temperature change of the chip exceeds the preset range, indicating that the current thermal management strategy may not be able to effectively control the temperature of the chip. The system will use data analysis tools and thermal simulation software to check the pattern and trend of temperature changes, analyze the temperature data points and weighted averages , find out the specific cause of the temperature anomaly, and use dynamic voltage and frequency adjustment technology to adjust the chip's operating parameters according to real-time temperature data, dynamically adjust the chip's operating frequency and voltage, reduce power consumption and heat generation, and use a thermal-aware compiler and the operating system's scheduler to dynamically adjust thread allocation, apply a preset thermal-aware thread scheduling algorithm, and reallocate threads to avoid overheating in hot spots; for example, when the system determines that the weighted average can match the pre-set moving window, the system will consider that the chip's temperature change is within the preset range, and the chip temperature can be effectively controlled. The system will move the moving window to the right by a single temperature data point, from the pre-set time period The next time node is intercepted from the falling edge, a single temperature data point is removed to the left from the initial time node, and the next time node is added to the initial time node to form iterative temperature data. The moving average of the iterative temperature data is calculated by applying weight distribution in sequence until the next time node can no longer be intercepted from the preset time segment, and the temperature trend data corresponding to the preset time segment is obtained by integration. The system can monitor the change of chip temperature in real time through the moving window method, dynamically adjust the calculation process, ensure the continuity and timeliness of temperature data, iteratively calculate the moving average of temperature data, adapt to changes in different time periods, and flexibly respond to fluctuations in temperature changes. At the same time, through the calculation of weighted average values ​​and iterative updates of moving windows, the system can accurately capture the trend of chip temperature changes and avoid misjudgment caused by fluctuations in a single data point. The integrated temperature trend data can help the system conduct in-depth temperature trend analysis and provide accurate data support for subsequent temperature management strategies. Real-time monitoring and dynamic adjustment ensure that the chip temperature operates within a safe range, improve the stability and reliability of the system, and avoid performance degradation and failures caused by overheating. Accurate temperature trend data helps to optimize resource allocation and thermal management strategies, improve system performance, and ensure that good performance can be maintained under high load and long-term operation.

[0094] It should be noted that the specific examples are as follows:

[0095] Assume that during operation, the system needs to monitor the temperature change of the chip in real time to ensure that the temperature of the chip is within a safe range. Then the system will use the moving window and weighted average method to monitor and analyze the temperature.

[0096] First, the initial temperature data is collected. The system collects one temperature data point per second within a preset time period (such as 10 seconds). The initial collected temperature data points are: [45, 46, 47, 45, 44, 46, 47, 48, 46, 45] (unit: ℃). Then the initial weighted average is calculated. The system assigns weights to each data point (such as the most recent time point has a greater weight, and the more distant the weight is, the smaller the weight). Assuming the weight distribution is: [0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1] (i.e. uniform weight), the calculated weighted average is:

[0097] =(45*0.1+46*0.1+47*0.1+45*0.1+44*0.1+46*0.1+47*0.1+48*0.1+46*0.1+45*0.1)

[0098] =45.9;

[0099] Then determine whether the weighted average matches the moving window. Assuming the preset moving window range is 45°C-50°C, then 45.9°C is within the preset range. The system determines that the chip temperature change is within the preset range. The system moves the moving window to the right by a single temperature data point, extracts the next time node data from the time segment, removes the leftmost temperature data point 45, and adds the newly collected temperature data point of the next time node (assuming it is 47). Then the new temperature data point sequence is: [46,47,45,44,46,47,48,46,45,47]. The weighted average needs to be recalculated by reapplying the same weight distribution: [0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.1,0.1], then the new weighted average is calculated as:

[0100] =(46*0.1+47*0.1+45*0.1+44*0.1+46*0.1+47*0.1+48*0.1+46*0.1+45*0.1+47*0.1)

[0101] =46.1;

[0102] The system repeats the above steps to determine whether the new weighted average value of 46.1°C is within the preset moving window range (45°C-50°C). If it matches, the system continues to move the moving window to the right, updates the temperature data point sequence, and recalculates the weighted average value. If the next time node cannot be intercepted, the temperature trend data corresponding to the preset time segment is integrated. Through the above method, the system can integrate the temperature data within the entire preset time segment to obtain the trend data of the chip temperature.

[0103] In summary, the system can monitor the temperature changes of the chip in real time and make dynamic adjustments to ensure the continuity and timeliness of the temperature data. At the same time, through the calculation of weighted average values ​​and iterative updates of moving windows, the system can accurately capture the trend of chip temperature changes and avoid misjudgment caused by fluctuations in a single data point. Based on accurate temperature trend data, the system can optimize thermal management strategies, including dynamically adjusting frequency and voltage, optimizing thread scheduling, and strengthening heat dissipation measures, etc., to enhance the overall thermal management effect of the chip. Real-time monitoring and dynamic adjustments ensure that the chip temperature operates within a safe range, improve the stability and reliability of the system, and avoid performance degradation and failures caused by overheating.

[0104] In this embodiment, the step S4 of judging whether the frequency voltage can execute the calculation test task preset by the processing terminal further includes:

[0105] S41: Based on the load type preset in the computing test task, the computing test task is run by the chip in the simulation environment preset in the processing terminal to collect simulation data, wherein the load type specifically includes CPU intensive, memory intensive, and I / O intensive;

[0106] S42: Determine whether the simulation data matches a preset data format;

[0107] S43: If yes, the simulation data is input into a pre-built energy consumption model, the power consumption of the chip under different frequency and voltage configurations is predicted from the energy consumption model, and the temperature change of the chip under different configurations is simulated by combining the energy consumption model and the thermal model.

[0108] In this embodiment, the system runs the computing test task through the chip in the simulation environment pre-set by the processing terminal based on the load type pre-set by the computing test task, and the load types specifically include CPU intensive, memory intensive, and I / O intensive, to collect simulation data, and then the system determines whether the simulation data matches the pre-set data format to execute the corresponding steps; for example, when the system determines that the simulation data cannot match the pre-set data format, the system will consider that an abnormality has occurred during the execution of the computing test task, resulting in incorrect data format, and the system will search and record abnormal information in the log, such as data loss and data format error, and analyze whether there are error records in the data collection process by checking the log file, and at the same time confirm whether the resource allocation and configuration parameters in the simulation environment are correct, adjust the mismatched parameters, check and reset the simulation environment parameters, ensure that the environment settings are consistent with the computing test task requirements, and perform single-step debugging to ensure that the execution results of each step are consistent with expectations, re-run the computing test task, and verify whether the computing task can be executed correctly; for example, when the system determines When the simulation data can match the pre-set data format, the system will consider that the calculation test task is normal, and the system will input the simulation data into the pre-built energy consumption model, predict the power consumption of the chip under different frequency and voltage configurations from the energy consumption model, and simulate the chip temperature changes under different configurations by combining the energy consumption model and the thermal model. By inputting the simulation data into the pre-built energy consumption model, the system can accurately predict the power consumption of the chip under different frequency and voltage configurations, which helps to understand the power consumption characteristics of the chip under different configurations, and provide a reliable basis for optimizing the power management strategy, avoiding excessive energy consumption, and improving energy efficiency. At the same time, by combining the energy consumption model and the thermal model, the system can simulate the temperature changes of the chip under different configurations, help identify possible overheating areas and high temperature points, thereby optimizing the heat dissipation design and layout, ensuring that the chip runs within a safe temperature range, and improving system stability and reliability. In addition, by simulating the chip temperature changes under different frequency and voltage configurations, the chip frequency and voltage settings can be dynamically adjusted to achieve on-demand adjustment of frequency and voltage, balance performance and power consumption, extend chip life, reduce heat loss, and improve the energy efficiency of the overall system.

[0109] In this embodiment, the step S1 of identifying the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content and collecting the thermal area of ​​the packaged stacked chip according to the heat distribution data includes:

[0110] S11: Using a visualization tool preset in the processing terminal to visualize the heat distribution data to generate a temperature field image;

[0111] S12: Determine whether the temperature field image completely presents the packaged stacked chip;

[0112] S13: If yes, then identify the internal temperature distribution of the packaged stacked chip in the temperature field image, identify the high temperature area and potential hot spots, perform structural simulation on the packaged stacked chip from the temperature field image, and simulate the heat dissipation effect corresponding to the preset heat dissipation structure.

[0113] In this embodiment, the system uses a visualization tool pre-set in the processing terminal to visualize the heat distribution data and generate a temperature field image. The system then determines whether the temperature field image fully presents the packaged stacked chips to execute the corresponding steps. For example, when the system determines that the temperature field image cannot fully present the packaged stacked chips, the system will consider that the thermal imaging method or image processing technology of the processing terminal cannot fully capture the temperature distribution inside the chip. The system will adjust the imaging angle and distance to ensure that all areas of the chip are within the field of view of the thermal imaging device. At the same time, before thermal imaging, the system will appropriately adjust or remove obstructions to ensure that all heat sources can be clearly captured, and rearrange the temperature sensors to ensure that all key areas of the chip are covered to provide comprehensive temperature data ... When the image can fully present the packaged stacked chips, the system will assume that the processing terminal can capture the temperature distribution inside the chip. The system will identify the internal temperature distribution of the packaged stacked chips in the temperature field image, identify high-temperature areas and potential hotspots, and perform structural simulation of the packaged stacked chips from the temperature field image to simulate the heat dissipation effect of the pre-set heat dissipation structure. The system optimizes the heat dissipation design by identifying high-temperature areas and potential hotspots, improves the heat dissipation efficiency of the chip, and ensures that the chip operates within a safe temperature range, thereby improving the stability and reliability of the system. It also verifies the effectiveness of the heat dissipation solution through structural simulation, reduces trial and error costs and time, and speeds up the product development process. It can monitor and dynamically adjust the heat dissipation strategy in real time to ensure that the chip temperature remains within a controllable range, thereby ensuring the long-term stable operation of the system.

[0114] Reference Figure 2 , is a three-dimensional chip packaging system in one embodiment of the present invention, comprising:

[0115] The acquisition module 10 is used to identify the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content preset by the processing terminal for the packaged stacked chip, and to acquire the heat area of ​​the packaged stacked chip according to the heat distribution data, wherein the heat area specifically includes a hot spot location area and an insufficient heat dissipation area;

[0116] A judging module 20, used to judge whether the heat area exceeds a preset upper limit of the processing terminal;

[0117] The execution module 30 is used for, if yes, unpacking the packaged stacked chips, constructing a thermal model of each chip from the processing terminal, identifying the packaged thermal distribution data of the chip according to the thermal model, obtaining the thermal conduction path and the air circulation direction of the chip, identifying the heat accumulation area of ​​the chip through the thermal conduction path and the air circulation direction, introducing a preset thermal interface material on the heat accumulation area, monitoring the temperature data of the chip in real time, and dynamically adjusting the preset frequency voltage of the chip based on the temperature data, wherein the packaged thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink;

[0118] A second judgment module 40 is used to judge whether the frequency voltage can perform the calculation test task preset by the processing terminal;

[0119] The second execution module 50 is used to evenly distribute the computing load of the computing test task if not, apply a preset thermal-aware thread scheduling algorithm to dynamically adjust the thread allocation of the chip, identify the execution rate of the hot code block based on the pre-developed thermal-aware compiler, balance the execution order and execution position of the hot code block during the compilation process, dynamically adjust the storage location of the data between different levels of cache of the chip according to the data access efficiency and access heat of the chip, complete the test process of the chip, and re-package the chip to obtain the packaged stacked chip.

[0120] In this embodiment, the acquisition module 10 identifies the heat distribution data of the packaged stacked chips under the load conditions corresponding to the thermal simulation content based on the thermal simulation content pre-set by the processing terminal for the packaged stacked chips, and acquires the thermal areas of the packaged stacked chips according to different heat distribution data. The thermal areas specifically include hot spot location areas and insufficient heat dissipation areas. Then the judgment module 20 judges whether these thermal areas exceed the area upper limit pre-set by the processing terminal to execute the corresponding steps; for example, when the system determines that the thermal area of ​​the packaged stacked chips does not exceed the area upper limit pre-set by the processing terminal, the system will consider that the temperature of the chip is within a safe range and there is no overheating, and the system will continue to use The current packaging and heat dissipation design saves the current thermal simulation and temperature monitoring data as the baseline data for possible future analysis and optimization, while continuing the normal operation and task execution of the packaged stacked chips, and monitoring the temperature of the chips in real time to ensure that the temperature is still within the safe range under different workload conditions. Although the current temperature does not exceed the upper limit, it is still necessary to regularly check and monitor the heat distribution to prevent the temperature from exceeding the standard due to future load increases or environmental changes. For example, when the system determines that the thermal area of ​​the packaged stacked chips exceeds the upper limit of the area pre-set by the processing terminal, the execution module 30 will consider that the temperature of the chip is not within the safe range and is prone to overheating. The system The system will unpack the packaged stacked chips, build different thermal models corresponding to different types of chips from the processing terminal, and identify the chip's package thermal distribution data based on the thermal model. The package thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink. The thermal conduction path and air circulation direction of the chip are obtained, and the heat accumulation area of ​​the chip is identified through different thermal conduction paths and air circulation directions. Pre-set thermal interface materials are introduced in different heat accumulation areas, and the temperature data of the chip is monitored in real time. The pre-set frequency and voltage of the chip are dynamically adjusted based on different temperature data. The system accurately identifies different heat accumulation areas through thermal models and simulation analysis. domain, can take targeted measures to optimize the heat dissipation design, introduce preset thermal interface materials in the heat accumulation area, improve heat conduction efficiency, reduce thermal resistance, and improve overall heat dissipation performance. At the same time, based on real-time temperature data, dynamically adjust the frequency and voltage of the chip to reduce power consumption and heat generation, maintain the system running within a safe temperature range, and be able to promptly detect and respond to temperature anomalies to prevent overheating. Based on thermal models and simulation data, preventive thermal design and optimization are performed to improve heat dissipation capabilities and reserve margins for possible future load changes. Then the second judgment module 40 determines whether the frequency and voltage of the chip can execute the calculation and test tasks pre-set by the processing terminal to execute the corresponding steps.For example, when the system determines that the frequency and voltage of the chip can execute the calculation and test tasks pre-set by the processing terminal, the system will believe that the current frequency and voltage settings ensure that the chip can efficiently complete the calculation and test tasks while maintaining a safe temperature range to ensure the reliability and stability of the system. The system will continue to execute the preset calculation and test tasks to ensure the normal progress of the workflow, while monitoring the temperature and working status of the chip in real time to ensure that no abnormalities occur during the task execution, and record the temperature, frequency, voltage and task execution status data under the current operating conditions in detail to provide a basis for subsequent analysis and optimization; for example, when the system determines that the frequency and voltage of the chip cannot execute the calculation and test tasks pre-set by the processing terminal, the second execution module 50 will believe that the chip is difficult to efficiently complete the calculation and test tasks. The system will evenly distribute the calculation load for the calculation and test tasks that the chip needs to process, apply a pre-set thermal-aware thread scheduling algorithm to dynamically adjust the chip's thread allocation, and identify the execution of hot code blocks based on a pre-developed thermal-aware compiler. rate, balance the execution order and execution position of hot code blocks during the compilation process, dynamically adjust the storage position of data between different levels of chip cache according to the chip's data access efficiency and access heat, complete the chip test process, and repackage the chip to obtain a packaged stacked chip; the system evenly distributes the computing load to avoid overload of certain areas or threads, reduce heat accumulation in hot areas, improve overall computing efficiency, eliminate performance bottlenecks caused by load imbalance, make all parts of the chip work evenly, and make full use of chip resources. At the same time, the thermal-aware thread scheduling algorithm is applied to dynamically adjust the chip's thread allocation according to real-time temperature data to avoid overheating in hot areas and improve heat dissipation efficiency. By dynamically adjusting thread allocation, heat distribution is made more uniform, the risk of local overheating is reduced, and the service life of the chip is extended. The perceptual compiler identifies and optimizes the execution order and execution position of hot code blocks, reduces heat accumulation in hot areas, balances the execution order of hot code blocks during the compilation process, optimizes the code execution path, reduces the temperature rise rate, and improves system stability. ;

[0121] In this embodiment, the execution module further includes:

[0122] a marking unit, used to establish a geometric model of the chip and its package based on a preset tool of the processing terminal, collect parameter settings of the chip, perform a test according to the parameter settings to obtain power consumption distribution data of the chip, and mark the regional power density of the chip from the geometric model, wherein the parameter settings specifically include chip levels, thermophysical parameters, and heat dissipation boundary conditions;

[0123] A determination unit, configured to determine whether the regional power density matches a preset density of the processing terminal;

[0124] An execution unit is used to generate image data of the chip and its package through the preset tool, display distribution information of the chip in different areas from the image data, and calculate the thermal resistance between the chip and the packaging material based on the distribution information, wherein the image data specifically includes a temperature distribution map and a heat flux distribution map, and the distribution information specifically includes a temperature gradient and a heat flux density direction.

[0125] In this embodiment, the system establishes a geometric model of the chip and its package based on a drawing tool pre-set in the processing terminal, collects parameter settings of the chip, and the parameter settings specifically include chip hierarchy, thermal physical parameters and heat dissipation boundary conditions. According to these parameter settings, tests are performed to obtain the power consumption distribution data of the chip, and the regional power density of the chip is marked from the geometric model. Then, the system determines whether these regional power densities match the density pre-set by the processing terminal to execute the corresponding steps; for example, when the system determines that the regional power density of the chip cannot match the density pre-set by the processing terminal, the system will consider that the power consumption distribution of the chip in actual work is inconsistent with the design expectation, which may cause local overheating or uneven heat dissipation. The system will use thermal simulation tools and power consumption distribution data to determine the power consumption concentration and hot spot areas, adjust the position of the chip functional units, reduce the concentration of high power consumption units, and evenly distribute the power consumption. At the same time, high-quality, high-thermal conductivity materials are selected to reduce manufacturing deviations, optimize manufacturing process parameters, ensure the consistency and uniformity of materials, and monitor the various chip components in real time. The temperature of the area is measured by lowering the frequency and voltage in the hot spot area to reduce power consumption and heat generation. For example, when the system determines that the regional power density of the chip can match the density pre-set by the processing terminal, the system will consider that the power consumption distribution of the chip in actual work is consistent with the design expectation, and the system will generate image data of the chip and its package through a pre-set drawing tool. The image data specifically includes a temperature distribution map and a heat flux distribution map. The distribution information of the chip in different areas is displayed from the image data. The distribution information specifically includes the temperature gradient and the direction of the heat flux density. The thermal resistance between the chip and the packaging material is calculated based on this distribution information. The system provides an intuitive display of the temperature gradient and the direction of the heat flux density in each area of ​​the chip through the generated temperature distribution map and heat flux distribution map, which helps to understand the current status of thermal management. At the same time, the thermal resistance between the chip and the packaging material is calculated based on the distribution information, which helps to optimize the selection and use of thermal interface materials and improve the heat conduction efficiency. In addition, the information on the temperature gradient and the direction of the heat flux density can guide the optimization of the radiator design, such as the arrangement of the heat sink and the installation position of the fan.

[0126] In this embodiment, it also includes:

[0127] A capture module, used for dynamically monitoring the operation process of the chip by applying preset thermal imaging based on the imaging parameters preset by the processing terminal for the chip, and capturing the real-time temperature change and heat accumulation trend of the chip, wherein the imaging parameters specifically include imaging angle and imaging distance, and the operation process specifically includes idle, low load, high load and peak load;

[0128] A third judgment module is used to judge whether the operation time of the chip can reach a preset time period;

[0129] The third execution module is used to, if not, identify the hotspot position and range in the chip through the high-resolution temperature measurement preset by the processing terminal and the thermal imaging, readjust the internal structural layout of the chip according to the hotspot distribution of the chip, evaluate the thermal conductivity of each packaging material on the processing terminal according to the structural layout, and construct heat dissipation paths with different efficiencies by combining the various packaging materials.

[0130] In this embodiment, the system uses pre-set imaging parameters for the chip based on the processing terminal, and the imaging parameters specifically include imaging angle and imaging distance. The system uses pre-set thermal imaging to dynamically monitor the operation process of the chip. The operation process specifically includes idle, low load, high load and peak load, and captures the real-time temperature change and heat accumulation trend of the chip. Then the system determines whether the operation time of the chip can reach the pre-set time period to execute the corresponding steps; for example, when the system determines that the operation time of the chip can reach the pre-set time period, the system will consider that the temperature change and heat accumulation of the chip during the entire preset operation cycle are acceptable and controllable. The system will select the best imaging angle and distance to obtain a clear thermal imaging image, record the temperature distribution of the chip in the idle state, observe whether there is basic heat accumulation, and record the temperature change data and heat accumulation trend of the chip in each operating state, generate corresponding thermal imaging images and data logs, and when the operation time of the chip reaches or exceeds the pre-set time period, it means that the temperature and heat distribution of the chip are good during the entire operation cycle, the thermal management design is effective, and the next stage of testing or mass production can be continued; Example For example, when the system determines that the operating time of the chip cannot reach the preset period, the system will consider that the temperature change and heat accumulation of the chip cannot be controlled during the entire preset operating cycle. The system will identify the hotspot position and hotspot range in the chip through the high-resolution temperature measurement and thermal imaging preset by the processing terminal, readjust the internal structure layout of the chip according to the hotspot distribution of the chip, evaluate the thermal conductivity of various packaging materials on the processing terminal according to the new structure layout, and construct heat dissipation paths with different efficiencies by combining various packaging materials; the system reduces the concentration of hotspots by redesigning the chip layout and optimizing the packaging materials, thereby improving the overall heat dissipation efficiency, dispersing high-power modules, evenly distributing heat, and avoiding local overheating problems. At the same time, through effective thermal management, the temperature peak of the chip during operation is reduced to avoid the impact of excessive temperature on chip performance and life. Good heat dissipation management reduces the damage of thermal stress to chip materials and structures, thereby extending the service life of the chip, and ensuring that the chip remains stable under high load and long-term operation, improving overall performance and user experience, optimizing the heat dissipation path, reducing the increase in power consumption caused by overheating, and improving energy utilization efficiency.

[0131] In this embodiment, the second execution module further includes:

[0132] A construction unit, used to construct a program control flow graph of the chip through the processing terminal, and identify hot loops of the hot code blocks from the program control flow graph;

[0133] A second judgment unit, used to judge whether the hotspot loop exceeds a preset loop control overhead;

[0134] The second execution unit is used to collect the position and execution frequency of the hot code block, rearrange the code order based on the position and the execution frequency, use an instruction set that supports thermal-aware optimization during the compilation process of the hot code block, and replace high-power instructions with low-power equivalent instructions according to the instruction set.

[0135] In this embodiment, the system pre-builds the program control flow graph of the chip through the processing terminal, identifies the hot loops of the hot code blocks from the program control flow graph, and then determines whether these hot loops exceed the preset loop control overhead to execute the corresponding steps; for example, when the system determines that the hot loop does not exceed the preset loop control overhead, the system will consider that the execution efficiency and resource consumption of the hot code block are within the controllable range, and the design of the hot loop is reasonable. The system will use the compiler tool chain to generate the control flow graph of the program, display the call relationship and execution path of each function, code block and loop, analyze the control flow graph to find the critical path and potential performance bottlenecks, and use the performance analysis tool to perform performance analysis on the program to find the code blocks and loops with the longest and most frequent execution time, mark these hot code blocks and loops in the control flow graph for further analysis and optimization, and pre-set the control overhead standards of the hot loops, such as CPU cycles, memory access times and cache hit rate, and evaluate whether the overhead of the identified hot loops is within an acceptable range according to the above standards; for example When the system determines that the hotspot loop exceeds the pre-set loop control overhead, the system will consider that the execution efficiency and resource consumption of the hotspot code block are difficult to control. The system will collect the location and execution frequency of the hotspot code block, rearrange the code order based on different locations and different execution frequencies, and use the instruction set that supports thermal-aware optimization during the compilation of the hotspot code block. According to different instruction sets, high-power instructions are replaced with equivalent low-power instructions; the system rearranges the code order, optimizes the memory layout, reduces the miss rate of the instruction cache and data cache, and improves the code execution efficiency. The optimized code order and low-power instruction set make the hotspot code block execute faster and improve the overall performance. At the same time, high-power instructions are replaced with low-power instructions, which effectively reduces the energy consumption of the chip, extends the battery life and reduces the heat dissipation requirements. The optimized instruction set and code layout make the system more energy-efficient, which helps to reduce the carbon footprint and operating cost of the system, and reduces the operating temperature of the chip by reducing power consumption and optimizing thermal management, improving the stability and reliability of the system. The lowered temperature reduces the impact of thermal stress on chip materials and extends the service life of the chip.

[0136] In this embodiment, the judgment module further includes:

[0137] A recording unit, configured to record the temperature data points corresponding to the initial time nodes based on the preset time segments of the processing terminal, and calculate the weighted average value of the temperature data points according to the preset weight distribution;

[0138] A third judging unit, configured to judge whether the weighted average value matches a preset moving window;

[0139] The third execution unit is used to move the moving window to the right by a single temperature data point, intercept the next time node from the preset time segment, remove a single temperature data point to the left in the initial time node, add the next time node to the initial time node to form iterative temperature data, apply the weight distribution to calculate the moving average of the iterative temperature data in sequence, until the next time node can no longer be intercepted in the preset time segment, and integrate to obtain the temperature trend data of the preset time segment.

[0140] In this embodiment, the system records the temperature data points corresponding to the initial time nodes based on the time segments preset by the processing terminal, calculates the weighted average of the temperature data points according to the preset weight distribution, and then the system determines whether the weighted average matches the preset moving window to execute the corresponding steps; for example, when the system determines that the weighted average cannot match the preset moving window, the system will consider that the temperature change of the chip exceeds the preset range, indicating that the current thermal management strategy may not be able to effectively control the temperature of the chip. The system will use data analysis tools and thermal simulation software to check the pattern and trend of temperature changes, analyze the temperature data points and weighted averages , find out the specific cause of the temperature anomaly, and use dynamic voltage and frequency adjustment technology to adjust the chip's operating parameters according to real-time temperature data, dynamically adjust the chip's operating frequency and voltage, reduce power consumption and heat generation, and use a thermal-aware compiler and the operating system's scheduler to dynamically adjust thread allocation, apply a preset thermal-aware thread scheduling algorithm, and reallocate threads to avoid overheating in hot spots; for example, when the system determines that the weighted average can match the pre-set moving window, the system will consider that the chip's temperature change is within the preset range, and the chip temperature can be effectively controlled. The system will move the moving window to the right by a single temperature data point, from the pre-set time period The next time node is intercepted from the falling edge, a single temperature data point is removed to the left from the initial time node, and the next time node is added to the initial time node to form iterative temperature data. The moving average of the iterative temperature data is calculated by applying weight distribution in sequence until the next time node can no longer be intercepted from the preset time segment, and the temperature trend data corresponding to the preset time segment is obtained by integration. The system can monitor the change of chip temperature in real time through the moving window method, dynamically adjust the calculation process, ensure the continuity and timeliness of temperature data, iteratively calculate the moving average of temperature data, adapt to changes in different time periods, and flexibly respond to fluctuations in temperature changes. At the same time, through the calculation of weighted average values ​​and iterative updates of moving windows, the system can accurately capture the trend of chip temperature changes and avoid misjudgment caused by fluctuations in a single data point. The integrated temperature trend data can help the system conduct in-depth temperature trend analysis and provide accurate data support for subsequent temperature management strategies. Real-time monitoring and dynamic adjustment ensure that the chip temperature operates within a safe range, improve the stability and reliability of the system, and avoid performance degradation and failures caused by overheating. Accurate temperature trend data helps to optimize resource allocation and thermal management strategies, improve system performance, and ensure that good performance can be maintained under high load and long-term operation.

[0141] In this embodiment, the second determination module further includes:

[0142] A collecting unit, configured to run the computing test task through the chip in a simulation environment preset by the processing terminal based on a load type preset by the computing test task, and collect simulation data, wherein the load type specifically includes CPU intensive, memory intensive, and I / O intensive;

[0143] A fourth determination unit, used to determine whether the simulation data matches a preset data format;

[0144] The fourth execution unit is used to input the simulation data into a pre-built energy consumption model, predict the power consumption of the chip under different frequency and voltage configurations from the energy consumption model, and combine the energy consumption model and the thermal model to simulate the chip temperature changes under different configurations.

[0145] In this embodiment, the system runs the computing test task through the chip in the simulation environment pre-set by the processing terminal based on the load type pre-set by the computing test task, and the load types specifically include CPU intensive, memory intensive, and I / O intensive, to collect simulation data, and then the system determines whether the simulation data matches the pre-set data format to execute the corresponding steps; for example, when the system determines that the simulation data cannot match the pre-set data format, the system will consider that an abnormality has occurred during the execution of the computing test task, resulting in incorrect data format, and the system will search and record abnormal information in the log, such as data loss and data format error, and analyze whether there are error records in the data collection process by checking the log file, and at the same time confirm whether the resource allocation and configuration parameters in the simulation environment are correct, adjust the mismatched parameters, check and reset the simulation environment parameters, ensure that the environment settings are consistent with the computing test task requirements, and perform single-step debugging to ensure that the execution results of each step are consistent with expectations, re-run the computing test task, and verify whether the computing task can be executed correctly; for example, when the system determines When the simulation data can match the pre-set data format, the system will consider that the calculation test task is normal, and the system will input the simulation data into the pre-built energy consumption model, predict the power consumption of the chip under different frequency and voltage configurations from the energy consumption model, and simulate the chip temperature changes under different configurations by combining the energy consumption model and the thermal model. By inputting the simulation data into the pre-built energy consumption model, the system can accurately predict the power consumption of the chip under different frequency and voltage configurations, which helps to understand the power consumption characteristics of the chip under different configurations, and provide a reliable basis for optimizing the power management strategy, avoiding excessive energy consumption, and improving energy efficiency. At the same time, by combining the energy consumption model and the thermal model, the system can simulate the temperature changes of the chip under different configurations, help identify possible overheating areas and high temperature points, thereby optimizing the heat dissipation design and layout, ensuring that the chip runs within a safe temperature range, and improving system stability and reliability. In addition, by simulating the chip temperature changes under different frequency and voltage configurations, the chip frequency and voltage settings can be dynamically adjusted to achieve on-demand adjustment of frequency and voltage, balance performance and power consumption, extend chip life, reduce heat loss, and improve the energy efficiency of the overall system.

[0146] In this embodiment, the acquisition module further includes:

[0147] A generating unit, configured to visualize the heat distribution data using a visualization tool preset by the processing terminal to generate a temperature field image;

[0148] A fifth judging unit, configured to judge whether the temperature field image completely presents the packaged stacked chip;

[0149] The fifth execution unit is used to identify the internal temperature distribution of the packaged stacked chip in the temperature field image, identify the high temperature area and potential hot spots, perform structural simulation on the packaged stacked chip from the temperature field image, and simulate the heat dissipation effect corresponding to the preset heat dissipation structure.

[0150] In this embodiment, the system uses a visualization tool pre-set in the processing terminal to visualize the heat distribution data and generate a temperature field image. The system then determines whether the temperature field image fully presents the packaged stacked chips to execute the corresponding steps. For example, when the system determines that the temperature field image cannot fully present the packaged stacked chips, the system will consider that the thermal imaging method or image processing technology of the processing terminal cannot fully capture the temperature distribution inside the chip. The system will adjust the imaging angle and distance to ensure that all areas of the chip are within the field of view of the thermal imaging device. At the same time, before thermal imaging, the system will appropriately adjust or remove obstructions to ensure that all heat sources can be clearly captured, and rearrange the temperature sensors to ensure that all key areas of the chip are covered to provide comprehensive temperature data ... When the image can fully present the packaged stacked chips, the system will assume that the processing terminal can capture the temperature distribution inside the chip. The system will identify the internal temperature distribution of the packaged stacked chips in the temperature field image, identify high-temperature areas and potential hotspots, and perform structural simulation of the packaged stacked chips from the temperature field image to simulate the heat dissipation effect of the pre-set heat dissipation structure. The system optimizes the heat dissipation design by identifying high-temperature areas and potential hotspots, improves the heat dissipation efficiency of the chip, and ensures that the chip operates within a safe temperature range, thereby improving the stability and reliability of the system. It also verifies the effectiveness of the heat dissipation solution through structural simulation, reduces trial and error costs and time, and speeds up the product development process. It can monitor and dynamically adjust the heat dissipation strategy in real time to ensure that the chip temperature remains within a controllable range, thereby ensuring the long-term stable operation of the system.

[0151] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A three-dimensional chip packaging method, characterized in that: The following steps are involved: Based on the thermal simulation content preset by the processing terminal for the packaged stacked chip, the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content is identified, and the thermal area of ​​the packaged stacked chip is collected according to the heat distribution data, wherein the thermal area specifically includes a hot spot location area and an insufficient heat dissipation area; determining whether the heat area exceeds a preset upper limit of the processing terminal; If so, the packaged stacked chips are unpacked, a thermal model of each chip is constructed from the processing terminal, the packaged thermal distribution data of the chip is identified based on the thermal model, the heat conduction path and the air circulation direction of the chip are obtained, the heat accumulation area of ​​the chip is identified through the heat conduction path and the air circulation direction, a preset thermal interface material is introduced into the heat accumulation area, the temperature data of the chip is monitored in real time, and the preset frequency voltage of the chip is dynamically adjusted based on the temperature data, wherein the packaged thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink; Determine whether the frequency voltage can execute the calculation test task preset by the processing terminal; If not, the computing load of the computing test task is evenly distributed, and the thread allocation of the chip is dynamically adjusted by applying a preset thermal-aware thread scheduling algorithm. The execution rate of the hot code block is identified based on the pre-developed thermal-aware compiler, and the execution order and execution position of the hot code block are balanced during the compilation process. According to the data access efficiency and access heat of the chip, the storage location of the data between different levels of cache of the chip is dynamically adjusted to complete the test process of the chip, and the chip is repackaged to obtain the packaged stacked chip.

2. The three-dimensional chip packaging method according to claim 1, characterized in that: The step of constructing a thermal model of each chip from the processing terminal and identifying the package thermal distribution data of the chip according to the thermal model includes: Establishing a geometric model of the chip and its package based on a preset tool of the processing terminal, collecting parameter settings of the chip, performing a test according to the parameter settings to obtain power consumption distribution data of the chip, and marking the regional power density of the chip from the geometric model, wherein the parameter settings specifically include chip levels, thermophysical parameters, and heat dissipation boundary conditions; Determining whether the regional power density matches the preset density of the processing terminal; If so, the image data of the chip and its package are generated by the preset tool, the distribution information of the chip in different areas is displayed from the image data, and the thermal resistance between the chip and the packaging material is calculated based on the distribution information, wherein the image data specifically includes a temperature distribution map and a heat flux distribution map, and the distribution information specifically includes a temperature gradient and a heat flux density direction.

3. The three-dimensional chip packaging method according to claim 1, characterized in that: Before the step of identifying the heat accumulation area of ​​the chip through the heat conduction path and the air circulation direction and introducing a preset thermal interface material on the heat accumulation area, the method further includes: Based on the preset imaging parameters of the chip by the processing terminal, the preset thermal imaging is used to dynamically monitor the operation process of the chip to capture the real-time temperature change and heat accumulation trend of the chip, wherein the imaging parameters specifically include imaging angle and imaging distance, and the operation process specifically includes idle, low load, high load and peak load; Determining whether the operating time of the chip can reach a preset time period; If not, the hotspot position and range in the chip are identified through high-resolution temperature measurement and thermal imaging preset at the processing terminal, and the internal structural layout of the chip is readjusted according to the hotspot distribution of the chip. The thermal conductivity of each packaging material is evaluated on the processing terminal according to the structural layout, and heat dissipation paths with different efficiencies are constructed by combining the various packaging materials.

4. The three-dimensional chip packaging method according to claim 1, characterized in that: The step of identifying the execution rate of the hot code block based on the pre-developed thermal-aware compiler and balancing the execution order and execution position of the hot code block during the compilation process includes: Constructing a program control flow graph of the chip through the processing terminal, and identifying a hot loop of the hot code block from the program control flow graph; Determining whether the hotspot loop exceeds a preset loop control overhead; If so, the position and execution frequency of the hot code block are collected, the code order is rearranged based on the position and the execution frequency, and an instruction set that supports thermal-aware optimization is used during the compilation process of the hot code block, and high-power instructions are replaced with low-power equivalent instructions according to the instruction set.

5. The three-dimensional chip packaging method according to claim 1, characterized in that: The step of determining whether the heat area exceeds the preset upper limit of the processing terminal further includes: Based on the preset time segment of the processing terminal, the temperature data points corresponding to the initial time node are recorded, and the weighted average value of the temperature data points is calculated according to the preset weight distribution; Determining whether the weighted average matches a preset moving window; If so, move the moving window to the right by a single temperature data point, intercept the next time node from the preset time segment, remove a single temperature data point to the left in the initial time node, add the next time node to the initial time node to form iterative temperature data, apply the weight distribution to calculate the moving average of the iterative temperature data in sequence, until the next time node can no longer be intercepted in the preset time segment, and integrate to obtain the temperature trend data of the preset time segment.

6. The three-dimensional chip packaging method according to claim 1, characterized in that: The step of determining whether the frequency voltage can execute the calculation test task preset by the processing terminal further includes: Based on the load type preset in the computing test task, the computing test task is run by the chip in the simulation environment preset in the processing terminal to collect simulation data, wherein the load type specifically includes CPU intensive, memory intensive, and I / O intensive; Determining whether the simulation data matches a preset data format; If so, the simulation data is input into a pre-built energy consumption model, the power consumption of the chip under different frequency and voltage configurations is predicted from the energy consumption model, and the energy consumption model and the thermal model are combined to simulate the chip temperature changes under different configurations.

7. The three-dimensional chip packaging method according to claim 1, characterized in that: The step of identifying the heat distribution data of the packaged stacked chip under the load condition corresponding to the thermal simulation content, and collecting the thermal area of ​​the packaged stacked chip according to the heat distribution data includes: Using a visualization tool preset in the processing terminal to visualize the heat distribution data and generate a temperature field image; Determining whether the temperature field image completely presents the packaged stacked chip; If so, the internal temperature distribution of the packaged stacked chip is identified in the temperature field image, high temperature areas and potential hot spots are identified, and the structure of the packaged stacked chip is simulated from the temperature field image to simulate the heat dissipation effect corresponding to the preset heat dissipation structure.

8. A three-dimensional chip packaging system, characterized in that: include: A collection module, for identifying heat distribution data of the packaged stacked chip under a load condition corresponding to the thermal simulation content preset by the processing terminal for the packaged stacked chip, and collecting a heat area of ​​the packaged stacked chip according to the heat distribution data, wherein the heat area specifically includes a hot spot location area and an insufficient heat dissipation area; A judging module, used for judging whether the heat area exceeds a preset upper limit of the processing terminal; an execution module, for, if yes, unpacking the packaged stacked chips, constructing a thermal model of each chip from the processing terminal, identifying the packaged thermal distribution data of the chip according to the thermal model, obtaining the thermal conduction path and the air circulation direction of the chip, identifying the heat accumulation area of ​​the chip through the thermal conduction path and the air circulation direction, introducing a preset thermal interface material on the heat accumulation area, monitoring the temperature data of the chip in real time, and dynamically adjusting the preset frequency voltage of the chip based on the temperature data, wherein the packaged thermal distribution data specifically includes the thermal interface between chip layers, the thermal conductivity of the packaging material, and the design of the heat sink; A second judgment module is used to judge whether the frequency voltage can execute the calculation test task preset by the processing terminal; The second execution module is used to evenly distribute the computing load of the computing test task if not, apply a preset thermal-aware thread scheduling algorithm to dynamically adjust the thread allocation of the chip, identify the execution rate of the hot code block based on the pre-developed thermal-aware compiler, balance the execution order and execution position of the hot code block during the compilation process, dynamically adjust the storage location of the data between different levels of cache on the chip according to the data access efficiency and access heat of the chip, complete the test process of the chip, and re-package the chip to obtain the packaged stacked chip.

9. The three-dimensional chip packaging system according to claim 8, characterized in that: The execution module also includes: a marking unit, used to establish a geometric model of the chip and its package based on a preset tool of the processing terminal, collect parameter settings of the chip, perform a test according to the parameter settings to obtain power consumption distribution data of the chip, and mark the regional power density of the chip from the geometric model, wherein the parameter settings specifically include chip levels, thermophysical parameters, and heat dissipation boundary conditions; A judging unit, configured to judge whether the regional power density matches a preset density of the processing terminal; An execution unit is used to generate image data of the chip and its package through the preset tool, display distribution information of the chip in different areas from the image data, and calculate the thermal resistance between the chip and the packaging material based on the distribution information, wherein the image data specifically includes a temperature distribution map and a heat flux distribution map, and the distribution information specifically includes a temperature gradient and a heat flux density direction.

10. The three-dimensional chip packaging system according to claim 8, characterized in that: Also includes: A capture module, used for dynamically monitoring the operation process of the chip by applying preset thermal imaging based on the imaging parameters preset by the processing terminal for the chip, and capturing the real-time temperature change and heat accumulation trend of the chip, wherein the imaging parameters specifically include imaging angle and imaging distance, and the operation process specifically includes idle, low load, high load and peak load; A third judgment module is used to judge whether the operation time of the chip can reach a preset time period; The third execution module is used to, if not, identify the hotspot position and range in the chip through the high-resolution temperature measurement preset by the processing terminal and the thermal imaging, readjust the internal structural layout of the chip according to the hotspot distribution of the chip, evaluate the thermal conductivity of each packaging material on the processing terminal according to the structural layout, and construct heat dissipation paths with different efficiencies by combining the various packaging materials.