Cooling performance evaluation method and device of cold plate, storage medium and electronic equipment

Abstract

The embodiment of the invention provides a cooling performance evaluation method and device of a cold plate, a storage medium and electronic equipment, and relates to the field of computers, and the method comprises the steps: obtaining a plurality of first temperature values of a central processing unit and a plurality of second temperature values of a graphic processor when the cold plate is in operation; according to the plurality of first temperature values and the plurality of second temperature values, performing first temperature difference judgment on a central processing unit and a graphic processing unit; under the condition that the result of the first temperature difference value judgment indicates that the first temperature difference value judgment is passed, first temperature standard deviation judgment is conducted on a central processing unit and a graphic processing unit according to the multiple first temperature numerical values and the multiple second temperature numerical values; under the condition that the result of the first temperature standard deviation judgment indicates that the first temperature standard deviation judgment is passed, according to the result of the first temperature difference value judgment and the result of the first temperature standard deviation judgment, first weighting judgment is carried out on a central processing unit and a graphic processing unit; and determining that the cold plate meets the first heat dissipation performance condition under the condition that the result of the first weighting judgment indicates that the cold plate passes the first weighting judgment.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computers, and in particular, to a method and device for evaluating the heat dissipation performance of a cold plate, a storage medium, and an electronic device. Background Art

[0002] In the process of server production and quality control, it is crucial to ensure that the heat dissipation performance of the liquid cooling system, especially the cold plate, reaches high standards. Traditional methods often rely on tests with a single temperature threshold, which are difficult to comprehensively and accurately evaluate the heat dissipation ability of the cold plate, thus resulting in a low accuracy of the heat dissipation performance evaluation of the cold plate.

[0003] Therefore, there is a technical problem of low accuracy in the heat dissipation performance evaluation of the cold plate in the related art. Summary of the Invention

[0004] The embodiments of the present application provide a method and device for evaluating the heat dissipation performance of a cold plate, a storage medium, and an electronic device, so as to at least solve the technical problem of low accuracy in the heat dissipation performance evaluation of the cold plate in the related art.

[0005] According to an embodiment of the present application, a method for evaluating the heat dissipation performance of a cold plate is provided, including: obtaining a plurality of first temperature values of a central processing unit and a plurality of second temperature values of a graphics processing unit when the cold plate is operating, where the cold plate is used to dissipate heat for the central processing unit and the graphics processing unit; when each temperature value in the plurality of first temperature values and the plurality of second temperature values is less than a first temperature threshold, making a first temperature difference judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values; when the result of the first temperature difference judgment indicates passing, making a first temperature standard deviation judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values; when the result of the first temperature standard deviation judgment indicates passing, making a first weighted judgment on the central processing unit and the graphics processing unit according to the result of the first temperature difference judgment and the result of the first temperature standard deviation judgment; when the result of the first weighted judgment indicates passing, determining that the cold plate meets the first heat dissipation performance condition.

[0006] According to another embodiment of the present application, a heat dissipation performance evaluation device for a cold plate is provided, including: an acquisition unit configured to acquire a plurality of first temperature values of a central processing unit and a plurality of second temperature values of a graphics processing unit when the cold plate is operating, wherein the cold plate is used to dissipate heat for the central processing unit and the graphics processing unit; a first judgment unit configured to perform a first temperature difference judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values when each of the plurality of first temperature values and the plurality of second temperature values is less than a first temperature threshold; a second judgment unit configured to perform a first temperature standard deviation judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values when the result of the first temperature difference judgment indicates passing; a third judgment unit configured to perform a first weighted judgment on the central processing unit and the graphics processing unit according to the result of the first temperature difference judgment and the result of the first temperature standard deviation judgment when the result of the first temperature standard deviation judgment indicates passing; a determination unit configured to determine that the cold plate meets the first heat dissipation performance condition when the result of the first weighted judgment indicates passing.

[0007] According to still another embodiment of the present application, a computer-readable storage medium is further provided. A computer program is stored in the computer-readable storage medium, wherein the computer program is configured to execute the steps in any one of the above method embodiments when running.

[0008] According to still another embodiment of the present application, an electronic device is further provided, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0009] Through the embodiments provided by the present application, temperature data of the central processing unit and the graphics processing unit are acquired during operation, not only monitoring whether the temperature exceeds the standard, but also further performing temperature difference judgment, temperature standard deviation judgment, and weighted judgment in sequence, establishing a comprehensive and accurate multi-dimensional heat dissipation performance evaluation system, significantly improving the diagnostic accuracy of the heat dissipation performance of the cold plate, thereby achieving the technical effect of improving the accuracy of the heat dissipation performance evaluation of the cold plate and solving the technical problem of low accuracy of the heat dissipation performance evaluation of the cold plate in the related art. Description of the Drawings

[0010] Figure 1 is a hardware structure block diagram of a heat dissipation performance evaluation method for a cold plate according to an embodiment of the present application;

[0011] Figure 2 is a flowchart of a heat dissipation performance evaluation method for a cold plate according to an embodiment of the present application;

[0012] Figure 3It is a flowchart of the implementation of the heat dissipation performance diagnosis of the cold plate of a liquid-cooled server according to an embodiment of the present application;

[0013] Figure 4 It is a flowchart of the implementation of the heat dissipation performance diagnosis of the cold plate of a liquid-cooled server according to an embodiment of the present application;

[0014] Figure 5 It is a block diagram of the structure of a heat dissipation performance evaluation device for a cold plate according to an embodiment of the present application. Detailed implementation manners

[0015] In the following, embodiments of the present application will be described in detail with reference to the drawings and in conjunction with the embodiments.

[0016] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0017] The method embodiments provided in the embodiments of the present application can be executed on a computer terminal or a similar computing device. Taking running on a computer terminal as an example, Figure 1 It is a hardware structure block diagram of a computer terminal for a heat dissipation performance evaluation method of a cold plate according to an embodiment of the present application. As Figure 1 shown, the computer terminal may include one or more ( Figure 1 only one is shown in the figure) processors 102 (the processors 102 may include, but are not limited to, processing devices such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. Among them, the above-mentioned computer terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned computer terminal. For example, the computer terminal may further include more or fewer components than those shown in Figure 1 the figure, or have a different configuration from that shown in Figure 1 the figure.

[0018] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for determining the mapping relationship in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the computer programs stored in the memory 104, that is, implements the above-mentioned method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory 104 may further include a memory remotely disposed relative to the processor 102, and these remote memories can be connected to the computer terminal through a network. Examples of the above-mentioned network include but are not limited to the Internet, enterprise intranet, local area network, mobile communication network, and combinations thereof.

[0019] The transmission device 106 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (abbreviated as NIC), which can be connected to other network devices through a base station and thus communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0020] As an alternative solution, the method for evaluating the heat dissipation performance of the cold plate specifically includes the following steps:

[0021] S202, obtain a plurality of first temperature values of the central processing unit and a plurality of second temperature values of the graphics processing unit when the cold plate is running, where the cold plate is used to dissipate heat for the central processing unit and the graphics processing unit;

[0022] S204, when each temperature value in the plurality of first temperature values and the plurality of second temperature values is less than the first temperature threshold, perform a first temperature difference judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values;

[0023] S206, when the result of the first temperature difference judgment indicates passing, perform a first temperature standard deviation judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values;

[0024] S208, when the result of the first temperature standard deviation judgment indicates passing, perform a first weighted judgment on the central processing unit and the graphics processing unit according to the result of the first temperature difference judgment and the result of the first temperature standard deviation judgment;

[0025] S210. When the result of the first weighted judgment indicates passing, it is determined that the cold plate meets the first heat dissipation performance condition.

[0026] Optionally, in this embodiment, the multiple first temperature values and the multiple second temperature values may, but are not limited to, being multiple temperature readings collected for the central processing unit (hereinafter referred to as CPU) and the graphics processing unit (hereinafter referred to as GPU) when the cold plate is operating under the first load condition (low load condition), and may, but are not limited to, being data continuously collected multiple times at the first sampling frequency.

[0027] Optionally, in this embodiment, the first temperature threshold is the set upper safety temperature for the CPU and GPU. Exceeding this upper limit will be regarded as insufficient heat dissipation performance.

[0028] Optionally, in this embodiment, the first temperature difference judgment is used to evaluate whether the maximum difference in the surface temperatures of the CPU and GPU under low load conditions meets a preset limit value, for checking the uniformity of the temperature distribution.

[0029] Optionally, in this embodiment, the first temperature standard deviation judgment is used to calculate the standard deviation of the CPU and GPU temperature data, thereby evaluating the stability of the heat dissipation performance, that is, the degree of temperature fluctuation.

[0030] Optionally, in this embodiment, the first weighted judgment is used to combine the evaluation results of the temperature difference and the temperature standard deviation, and conduct a comprehensive analysis through predefined weights to determine whether the overall heat dissipation performance of the cold plate reaches the first heat dissipation performance standard.

[0031] Optionally, in this embodiment, analyze the multiple first temperature values (CPU temperature) and the multiple second temperature values (GPU temperature) obtained under the current load condition (such as low load condition), and confirm whether each reading is lower than the preset first temperature threshold. If all readings do not exceed the first temperature threshold, the cold plate has initially passed the temperature safety check and can proceed to the next judgment.

[0032] For example, the preset temperature threshold is 90°C. During the low load test, the temperature readings of the CPU and GPU have always remained below 85°C and have not exceeded the limit.

[0033] Further analyze the temperature data that has passed the temperature threshold check, calculate the maximum difference in the surface temperatures of the CPU and GPU, and determine whether this difference is less than the temperature difference limit defined in the configuration file. For example, calculate the temperature differences at each measurement point of the CPU and the temperature differences at each measurement point of the GPU, and evaluate the heat dissipation uniformity of the cold plate by comparing the differences with the limit.

[0034] For example, if the temperature difference limits for the CPU and GPU are set to 5°C and 7°C respectively, under low load conditions, the maximum temperature differences between the CPU and GPU are 3°C and 6°C respectively, which do not exceed the limits, and the temperature difference judgment passes.

[0035] Calculate the standard deviation of the CPU and GPU temperature data to check the stability of the heat dissipation performance, that is, the degree of fluctuation of the temperature readings. If the standard deviation of the temperature data is less than the standard deviation limit defined in the configuration file, the cold plate also meets the requirements in terms of temperature stability.

[0036] For example, the standard deviation limit for the CPU temperature data is set to 2°C, while that for the GPU is 3°C. Under low load conditions, the standard deviations of the CPU and GPU temperatures are 1.7°C and 2.5°C respectively, both within the limit range, and the standard deviation judgment also passes.

[0037] Weight and fuse the results of the temperature difference judgment and the temperature standard deviation judgment to form a comprehensive evaluation index. For example, if the weights of the temperature difference and the standard deviation are 0.6 and 0.4 respectively, then the comprehensive heat dissipation performance score of the cold plate will be calculated based on the evaluation results of these two indicators. If the comprehensive score meets the preset multi-index fusion limit, the cold plate will be considered to meet the heat dissipation performance standard under low load conditions.

[0038] For example, assume that the temperature difference judgment score (based on the comparison result with the limit) is 80, the temperature standard deviation judgment score (based on the comparison result with the limit) is 90, and the weights are 0.6 and 0.4 respectively. Then the comprehensive score is calculated as: 80 * 0.6 + 90 * 0.4 = 84. If the multi-index fusion limit is set to 85, since the comprehensive score of 84 is slightly lower than the limit, it is necessary to re-evaluate or adjust the cold plate design.

[0039] Finally, if the cold plate performs well in each step of the temperature threshold, temperature difference, temperature standard deviation, and the first weighted judgment, meeting all the set standards, then it can be determined that the heat dissipation performance of the cold plate under the current load conditions meets the first heat dissipation performance standard.

[0040] Through the embodiments provided in this application, through multi-stage temperature data evaluation, it is ensured that the cold plate can not only operate under the safe temperature threshold, but also maintain good temperature distribution uniformity and stability, comprehensively considering multiple aspects of the heat dissipation performance. The setting of this process makes the evaluation of the cold plate more scientific and comprehensive, avoiding the deviation that may be brought by single-index evaluation, ensuring that when the server is running at low load, the cold plate can effectively manage heat, prevent overheating, maintain the normal working state of the CPU and GPU, and thus ensure the stability of the server and extend the hardware life.

[0041] As an alternative solution, based on multiple first temperature values and multiple second temperature values, a first temperature difference judgment is performed on the central processing unit and the graphics processing unit, including:

[0042] From the multiple first temperature values, determine the first temperature values of each central processing unit in the central processing unit, and from the multiple second temperature values, determine the second temperature values of each graphics processing unit in the graphics processing unit. Among them, the first temperature values of each central processing unit are used to determine the average temperature value of each central processing unit, and the second temperature values of each graphics processing unit are used to determine the average temperature value of each graphics processing unit;

[0043] According to the first average temperature value with the highest average temperature value and the second average temperature value with the lowest average temperature value among the average temperature values of each central processing unit, determine the first temperature difference parameter corresponding to the central processing unit;

[0044] According to the third average temperature value with the highest average temperature value and the fourth average temperature value with the lowest average temperature value among the average temperature values of each graphics processing unit, determine the second temperature difference parameter corresponding to the graphics processing unit;

[0045] In the case where the first temperature difference parameter is less than the temperature difference threshold and the second temperature difference parameter is less than the temperature difference threshold, determine that the result indication of the first temperature difference judgment passes.

[0046] Optionally, in this embodiment, the first temperature difference parameter and the second temperature difference parameter respectively represent the difference ratio between the highest and lowest average temperatures in the CPU and GPU under the first load condition, so as to evaluate the uniformity of the cold plate heat dissipation performance. The temperature difference threshold is used to determine whether the temperature difference parameter is within an acceptable range, so as to determine whether the cold plate heat dissipation performance meets the standard.

[0047] Optionally, in this embodiment, extract the temperature data belonging to each CPU from the multiple first temperature values, and calculate the average temperature value of each CPU; similarly, extract the temperature data belonging to each GPU from the multiple second temperature values, and calculate the average temperature value of each GPU.

[0048] For example, under low load conditions, the system monitors the continuous temperature data of 4 CPUs and 2 GPUs, which are collected once every 1 minute for a total of 60 times. For each CPU, we obtain 60 temperature readings and calculate the average temperature value. For example, the average temperature value of the first CPU is 85 °C, the second is 84 °C, and so on. For the GPU, similarly, the average temperature value of the first GPU is 80 °C, and the second is 82 °C.

[0049] Find the highest and lowest values among the average temperature values of the CPU and GPU, and record them as the first average temperature value and the second average temperature value for the CPU; and the third average temperature value and the fourth average temperature value for the GPU. Then, calculate the first temperature difference parameter and the second temperature difference parameter, that is, the ratio of the difference between the highest average temperature value and the lowest average temperature value.

[0050] For example, assume that the average temperature value of the first CPU is the highest value of 85°C (the first average temperature value), and the average temperature value of the second CPU is the lowest value of 84°C (the second average temperature value). Calculate the first temperature difference parameter: (85 - 84) / 84.5 is approximately 1.18%. Similarly, assume that the average temperature value of the first GPU is 82°C (the third average temperature value), and the average temperature value of the second GPU is 80°C (the fourth average temperature value). Calculate the second temperature difference parameter: (82 - 80) / 81 is approximately 2.47%.

[0051] Compare the calculated first temperature difference parameter and the second temperature difference parameter with the temperature difference threshold preset in the configuration file. If both of these parameters are less than the threshold, it indicates that under the current load conditions, the cold plate heat dissipation performance is uniform and there is no serious local overheating phenomenon.

[0052] Through the embodiments provided in this application, by comparing the first temperature difference parameter, the second temperature difference parameter with the temperature difference threshold, it is possible to accurately identify deficiencies in design or manufacturing, and ensure that the cold plate heat dissipation performance of the liquid-cooled server meets the expected consistency and stability standards before actual deployment.

[0053] As an alternative solution, according to the first average temperature value with the highest average temperature value and the second average temperature value with the lowest average temperature value among the average temperature values of each central processing unit, determine the first temperature difference parameter corresponding to the central processing unit, including:

[0054] Accumulate the average temperature values of each central processing unit to obtain the first accumulated temperature value;

[0055] Use the first difference between the first average temperature value and the second average temperature value, and divide it by the first accumulated temperature value to obtain the first temperature difference parameter;

[0056] According to the third average temperature value with the highest average temperature value and the fourth average temperature value with the lowest average temperature value among the average temperature values of each graphics processing unit, determine the second temperature difference parameter corresponding to the graphics processing unit, including:

[0057] Accumulate the average temperature values of each graphics processing unit to obtain the second accumulated temperature value;

[0058] The second temperature difference parameter is obtained by dividing the second difference between the third average temperature value and the fourth average temperature value by the second cumulative temperature value.

[0059] Optionally, in this embodiment, the first cumulative temperature value is the sum of the average temperature values of all CPUs under low load conditions, which is used as the base for calculating the first temperature difference parameter. The second cumulative temperature value is the sum of the average temperature values of all GPUs under the same low load conditions, which is used as the base for calculating the second temperature difference parameter.

[0060] Optionally, in this embodiment, the first difference and the second difference respectively represent the difference between the highest average temperature value and the lowest average temperature value in the CPU (the difference between the first average temperature value and the second average temperature value), and the difference between the highest average temperature value and the lowest average temperature value in the GPU (the difference between the third average temperature value and the fourth average temperature value).

[0061] Optionally, in this embodiment, the first temperature difference parameter and the second temperature difference parameter are used to quantitatively evaluate the consistency of the surface temperature distribution of the CPU and the GPU by calculating the ratios of the first difference and the second difference to their corresponding cumulative temperature values.

[0062] Optionally, in this embodiment, the average temperature values of each CPU are added together to obtain the first cumulative temperature value; similarly, the average temperature values of each GPU are added together to obtain the second cumulative temperature value.

[0063] For example, assume that under low load conditions, the average temperature values of 4 CPUs are 85°C, 84°C, 86°C, and 83°C respectively. Then the first cumulative temperature value = 85 + 84 + 86 + 83 = 338°C. At the same time, the average temperature values of 2 GPUs are 82°C and 80°C respectively. Then the second cumulative temperature value = 82 + 80 = 162°C.

[0064] Calculate the first difference (i.e., the difference between the first average temperature value and the second average temperature value, which is the difference between the highest and lowest average temperature values here), and then divide it by the first cumulative temperature value to obtain the first temperature difference parameter. For the GPU, calculate the second difference (i.e., the difference between the third average temperature value and the fourth average temperature value), and then divide it by the second cumulative temperature value to obtain the second temperature difference parameter.

[0065] For example, as obtained from the above steps, the highest average temperature value of the CPU is 86 °C (the first average temperature value), the lowest average temperature value is 83 °C (the second average temperature value), and the first difference = 86 - 83 = 3 °C. Calculate the first temperature difference parameter = (the first difference / the first cumulative temperature value) × 100% = (3 / 338) × 100% ≈ 0.89%. For the GPU, the highest average temperature value is 82 °C (the third average temperature value), the lowest average temperature value is 80 °C (the fourth average temperature value), and the second difference = 82 - 80 = 2 °C. Calculate the second temperature difference parameter = (the second difference / the second cumulative temperature value) × 100% = (2 / 162) × 100% ≈ 1.23%.

[0066] Through the embodiments provided by the present application, the calculation process of the temperature difference parameter is further refined. By accumulating the average temperature values, a reference benchmark is formed. Then, by calculating the ratio of the difference to the accumulated value, a quantitative evaluation of the surface temperature distribution uniformity of the CPU and GPU is realized. The calculation method of this parameter not only considers the influence of the maximum temperature difference but also evaluates it under the sum of all average temperature values, which can more accurately reflect the consistency level of the cold plate heat dissipation performance and avoid the one-sidedness that may be brought about by simply relying on the maximum temperature difference judgment.

[0067] As an alternative solution, based on multiple first temperature values and multiple second temperature values, perform a first temperature standard deviation judgment on the central processing unit and the graphics processing unit, including:

[0068] When the average temperature values of each central processing unit in the central processing unit are obtained, perform a standard deviation operation on the average temperature values of each central processing unit to obtain a first temperature standard deviation parameter;

[0069] When the average temperature values of each graphics processing unit in the graphics processing unit are obtained, perform a standard deviation operation on the average temperature values of each graphics processing unit to obtain a second temperature standard deviation parameter;

[0070] When the first temperature standard deviation parameter is less than the temperature standard deviation threshold and the second temperature standard deviation parameter is less than the temperature standard deviation threshold, determine that the result indication of the first temperature standard deviation judgment passes.

[0071] Optionally, in this embodiment, the first temperature standard deviation parameter is the standard deviation of the average temperature values of all CPUs under the current load condition, which is used to evaluate the volatility of the CPU surface temperature and the consistency of the heat dissipation system. The second temperature standard deviation parameter is the standard deviation of the average temperature values of all GPUs under the same current load condition, which is used to evaluate the stability of the GPU surface temperature and the uniformity of the heat dissipation system. The temperature standard deviation threshold is a maximum allowable standard deviation value defined in the configuration file, which is used to determine whether the temperature fluctuations of the CPU and GPU are within an acceptable range.

[0072] Optionally, in this embodiment, after obtaining the average temperature values of the CPU and GPU, the standard deviation of these values is calculated respectively to obtain the first temperature standard deviation parameter and the second temperature standard deviation parameter.

[0073] For example, assume that under low load conditions, the average temperature values of 4 CPUs are 85°C, 84°C, 86°C, and 83°C respectively. The standard deviation of these values is an index to measure their volatility. Suppose the calculated first temperature standard deviation parameter is 1.5°C. For the GPU, assume that the average temperature values of 2 GPUs are 82°C and 80°C respectively, and the calculated second temperature standard deviation parameter is 1.4°C.

[0074] Compare the calculated first and second temperature standard deviation parameters with the temperature standard deviation threshold in the configuration file respectively. If both parameters are less than or equal to the threshold, it indicates that the temperature fluctuations of the CPU and GPU are within the normal range, and the cold plate heat dissipation performance is stable and uniform.

[0075] For example, assume that the temperature standard deviation threshold set in the configuration file is 2°C. Since the first temperature standard deviation parameter 1.5°C is less than the threshold, and the second temperature standard deviation parameter 1.4°C is also less than the threshold, the result of the temperature standard deviation judgment indicates passing.

[0076] Through the embodiment provided by this application, by calculating the first and second temperature standard deviation parameters, the fluctuations of the CPU and GPU surface temperatures can be quantitatively evaluated, ensuring that the cold plate can provide a stable heat dissipation effect and avoiding potential problems caused by excessive temperature fluctuations, such as local overheating or unstable heat dissipation efficiency.

[0077] As an optional solution, according to the results of the temperature difference judgment and the temperature standard deviation judgment, a first weighted judgment is performed on the central processing unit and the graphics processing unit, including:

[0078] According to the result of the first temperature difference judgment, obtain the first temperature difference parameter of the central processing unit and the second temperature difference parameter of the graphics processing unit, and according to the result of the first temperature standard deviation judgment, obtain the first temperature standard deviation parameter of the central processing unit and the second temperature standard deviation parameter of the graphics processing unit;

[0079] Perform a first weighted fusion operation on the first temperature difference parameter and the first temperature standard deviation parameter to obtain the first comprehensive parameter of the central processing unit;

[0080] Perform a second weighted fusion operation on the second temperature difference parameter and the second temperature standard deviation parameter to obtain the second comprehensive parameter of the graphics processing unit;

[0081] In the case where the first comprehensive parameter is less than the comprehensive parameter threshold and the second comprehensive parameter is less than the comprehensive parameter threshold, determine that the result of the first weighted judgment indicates passing.

[0082] Optionally, in this embodiment, the first comprehensive parameter and the second comprehensive parameter are the results obtained by weighted fusion of the first and second temperature difference parameters and the first and second temperature standard deviation parameters, and are used to comprehensively evaluate the heat dissipation performance of the CPU and GPU under low load conditions.

[0083] Optionally, in this embodiment, the comprehensive parameter threshold is a numerical standard for judging whether the first and second comprehensive parameters indicate that the cold plate heat dissipation performance is qualified.

[0084] Optionally, in this embodiment, first, according to the result of the temperature difference judgment, obtain the first temperature difference parameter of each CPU and the second temperature difference parameter of each GPU. Then, according to the result of the temperature standard deviation judgment, obtain the first temperature standard deviation parameter of each CPU and the second temperature standard deviation parameter of each GPU.

[0085] For example, assume that after the temperature difference judgment, the first temperature difference parameter of the CPU is obtained as 1.0% and the second temperature difference parameter of the GPU is obtained as 1.2%; after the temperature standard deviation judgment, the first temperature standard deviation parameter of the CPU is obtained as 1.5 °C and the second temperature standard deviation parameter of the GPU is obtained as 1.4 °C.

[0086] Perform weighted fusion on the first temperature difference parameter and the first temperature standard deviation parameter to form the first comprehensive parameter. Perform weighted fusion on the second temperature difference parameter and the second temperature standard deviation parameter to form the second comprehensive parameter. The weighted fusion can assign different weights according to the importance of different parameters.

[0087] For example, when calculating the first comprehensive parameter, the weight of the first temperature difference parameter is 0.7, and the weight of the first temperature standard deviation parameter is 0.3; when calculating the second comprehensive parameter, the weight of the second temperature difference parameter is also 0.7, and the weight of the second temperature standard deviation parameter is 0.3. Therefore, the first comprehensive parameter = 1.0% × 0.7 + 1.5°C × 0.3; the second comprehensive parameter = 1.2% × 0.7 + 1.4°C × 0.3.

[0088] Compare the first and second comprehensive parameters with the comprehensive parameter thresholds set in the configuration file respectively. If both comprehensive parameters are less than the thresholds, then the result of the first weighted judgment indicates passing, indicating that the comprehensive evaluation of the heat dissipation performance of the CPU and GPU under low load conditions meets the standards.

[0089] For example, assume that the comprehensive parameter thresholds set in the configuration file are 2.0% and 2.5°C. According to the calculation in step 2, the first comprehensive parameter is approximately 1.35%, and the second comprehensive parameter is approximately 1.18% + 0.42°C. Considering the nature of the parameters after weighted fusion, if it is regarded as a comprehensive percentage form, the values can be approximately calculated or directly compared. In this case, both the first and second comprehensive parameters are less than the corresponding thresholds, so it is determined that the result of the first weighted judgment indicates passing.

[0090] Through the embodiments provided in this application, by fusing the temperature difference parameter and the temperature standard deviation parameter, a comprehensive and quantitative heat dissipation performance evaluation index is provided. The temperature difference parameter reflects the consistency of temperatures at different positions, while the temperature standard deviation parameter measures the stability of temperature data. By weighted fusion of these two parameters, the comprehensive level of the cold plate heat dissipation performance can be evaluated more accurately, ensuring that the heat dissipation of the CPU and GPU is both uniform and stable under low load conditions, meeting the expected heat dissipation performance requirements.

[0091] As an alternative solution, perform a first weighted fusion operation on the first temperature difference parameter and the first temperature standard deviation parameter to obtain the first comprehensive parameter of the central processing unit, including:

[0092] Obtain the first weighted parameter corresponding to the central processing unit and the second weighted parameter corresponding to the graphics processing unit, and obtain the weighted parameter sum value of the first weighted parameter and the second weighted parameter;

[0093] Obtain the first cumulative parameter obtained by multiplying the first weighted parameter by the first temperature difference parameter, and obtain the second cumulative parameter obtained by multiplying the second weighted parameter by the first temperature standard deviation parameter;

[0094] Determine the result of dividing the sum of the first cumulative parameter and the second cumulative parameter by the weighted parameter sum value as the first comprehensive parameter;

[0095] Perform a second weighted fusion operation on the second temperature difference parameter and the second temperature standard deviation parameter to obtain a second comprehensive parameter of the graphics processing unit, including:

[0096] Obtain a third cumulative parameter obtained by multiplying the first weighted parameter by the second temperature difference parameter, and obtain a fourth cumulative parameter obtained by multiplying the second weighted parameter by the second temperature standard deviation parameter;

[0097] Determine the sum of the third cumulative parameter and the fourth cumulative parameter, and divide the result by the sum of the weighted parameters as the second comprehensive parameter.

[0098] Optionally, in this embodiment, the product of the first weighted parameter and the second weighted parameter is a weight value preset when calculating the first comprehensive parameter and the second comprehensive parameter, which is used to highlight the importance of one parameter relative to another parameter. The sum of the weighted parameters is the sum of the first weighted parameter and the second weighted parameter, which is used to standardize the calculation result in the weighted fusion process.

[0099] Optionally, in this embodiment, preset a first weighted parameter (for the first temperature difference parameter) and a second weighted parameter (for the first temperature standard deviation parameter) for the CPU. For example, the first weighted parameter is set to 0.7 and the second weighted parameter is set to 0.3 to ensure that the importance of the temperature difference is higher than that of the temperature standard deviation.

[0100] Obtain the sum of the weighted parameters. Add the first weighted parameter and the second weighted parameter to obtain the sum of the weighted parameters. In this example, the sum of the weighted parameters = 0.7 + 0.3 = 1.

[0101] Obtain a first cumulative parameter obtained by multiplying the first weighted parameter by the first temperature difference parameter, and a second cumulative parameter obtained by multiplying the second weighted parameter by the first temperature standard deviation parameter. For example, if the first temperature difference parameter is 1.0%, the first cumulative parameter = 0.7 × 1.0% = 0.7%; if the first temperature standard deviation parameter is 1.5 °C, the second cumulative parameter = 0.3 × 1.5 °C = 0.45 °C.

[0102] Divide the sum of the first cumulative parameter and the second cumulative parameter by the sum of the weighted parameters to obtain the first comprehensive parameter. In this example, the first comprehensive parameter = (0.7% + 0.45 °C) / 1. Since the parameter units are different, it is usually necessary to convert them to the same unit (such as converting them all to percentage form) for calculation, or take other appropriate standardization measures.

[0103] Similarly, set a first weighting parameter (for the second temperature difference parameter) and a second weighting parameter (for the second temperature standard deviation parameter) for the GPU. For example, the first weighting parameter is set to 0.7 and the second weighting parameter is set to 0.3. The sum of the weighting parameters is also 1 (0.7 + 0.3 = 1), which is consistent with the calculation of the CPU. Obtain the third cumulative parameter by multiplying the first weighting parameter by the second temperature difference parameter, and the fourth cumulative parameter by multiplying the second weighting parameter by the second temperature standard deviation parameter. For example, if the second temperature difference parameter is 1.2%, the third cumulative parameter = 0.7 × 1.2% = 0.84%; if the second temperature standard deviation parameter is 1.4 °C, the fourth cumulative parameter = 0.3 × 1.4 °C = 0.42 °C. Divide the sum of the third cumulative parameter and the fourth cumulative parameter by the sum of the weighting parameters to obtain the second comprehensive parameter. Similar to the calculation of the first comprehensive parameter, it is necessary to ensure that the parameter units are consistent or take appropriate standardization measures.

[0104] Through the embodiments provided in the present application, by calculating the first and second comprehensive parameters and comparing them with the comprehensive parameter threshold, it can be determined whether the heat dissipation performance of the cold plate under low load conditions meets high standards, ensuring that the server can maintain a good thermal management level in actual applications. Especially in environments with strict requirements for heat dissipation such as high-performance computing and data centers, the introduction of this evaluation mechanism greatly improves the effectiveness and accuracy of the diagnostic system.

[0105] As an alternative solution, before making the first temperature difference judgment on the central processing unit and the graphics processing unit according to multiple first temperature values and multiple second temperature values, the method further includes:

[0106] Obtain the target material type of the cold plate;

[0107] Determine the temperature difference threshold corresponding to the target material type, the temperature standard deviation threshold corresponding to the target material type, and the comprehensive parameter threshold corresponding to the target material type from the configuration file, where the temperature difference threshold is used for the first temperature difference judgment, the temperature standard deviation threshold is used for the first temperature standard deviation judgment, and the comprehensive parameter threshold is used for the first weighting judgment. Multiple sets of corresponding relationships between the material type and the threshold are stored in the configuration file.

[0108] Optionally, in this embodiment, the target material type refers to the material type to which the cold plate of the liquid-cooled server being tested belongs, and different material types correspond to different heat dissipation performance requirements and standards.

[0109] Optionally, in this embodiment, the configuration file stores a file of the corresponding relationships between multiple cold plate material types and each threshold, including the temperature difference threshold, the temperature standard deviation threshold, and the comprehensive parameter threshold, which is used to guide the parameter acquisition and comparison in the heat dissipation performance judgment process.

[0110] Optionally, in this embodiment, the temperature difference threshold, the temperature standard deviation threshold, and the comprehensive parameter threshold are the heat dissipation performance standards defined in the configuration file for a specific material type, and are respectively used for the first temperature difference judgment, the first temperature standard deviation judgment, and the first weighted judgment to ensure that the heat dissipation performance of the cold plate of the liquid-cooled server meets the design requirements.

[0111] Optionally, in this embodiment, before starting the heat dissipation performance judgment, it is first necessary to determine the material type of the cold plate to be tested, and this step is crucial for subsequently calling the correct threshold from the configuration file.

[0112] For example, assume that the currently tested cold plate is of model V0020FZ, and its target material type is V0020FZ.

[0113] Once the target material type is determined, then retrieve the temperature difference threshold, the temperature standard deviation threshold, and the comprehensive parameter threshold related to this material type from the configuration file to provide a benchmark for the upcoming heat dissipation performance judgment.

[0114] After obtaining the thresholds corresponding to the target material type, the first heat dissipation performance judgment of the cold plate can be performed based on the collected first temperature data and second temperature data, and the heat dissipation performance of the cold plate can be evaluated by comparing with the thresholds.

[0115] Through the embodiment provided by the present application, the importance of obtaining the target material type of the cold plate and its corresponding threshold before performing the heat dissipation performance judgment is shown. This process ensures the pertinence and accuracy of the heat dissipation performance test, and avoids the deviation of the test results that may be caused by using general thresholds. Through the configuration file, a series of heat dissipation performance standards matching the target material type, including the temperature difference threshold, the temperature standard deviation threshold, and the comprehensive parameter threshold, can be quickly located and called before the test starts. This not only simplifies the test process, reduces human errors, but also can adapt to the specific heat dissipation requirements of cold plates of different material types, improving the efficiency and accuracy of the test.

[0116] As an alternative solution, the method further includes:

[0117] In the case where there are temperature values greater than or equal to the preset threshold among multiple first temperature values and multiple second temperature values, or in the case where the result of the first temperature difference judgment indicates failure, or in the case where the result of the first temperature standard deviation judgment indicates failure, or in the case where the result of the first weighted judgment indicates failure, it is determined that the result of the first heat dissipation performance judgment indicates that the cold plate does not meet the first heat dissipation performance standard.

[0118] Optionally, in this embodiment, among the multiple first temperature values (CPU average temperature) and multiple second temperature values (GPU average temperature) collected, it is checked whether there is any temperature value exceeding the preset threshold. If such a situation exists, it is preliminarily determined that there may be a serious problem with the heat dissipation performance of the cold plate and the heat dissipation requirement is not basically met, and the test result will indicate that the first heat dissipation performance standard is not passed.

[0119] After the calculation of the temperature difference parameter is completed, if the first temperature difference parameter exceeds the temperature difference threshold defined in the configuration file for the target material type, the result of the first temperature difference judgment also indicates non-pass, indicating that the temperature distribution on the surface of the cold plate is uneven and the heat dissipation performance is poor.

[0120] After calculating the temperature standard deviation parameter, if the first temperature standard deviation parameter or the second temperature standard deviation parameter exceeds the corresponding temperature standard deviation threshold, this indicates that the temperature of the CPU or GPU fluctuates greatly and the stability of the heat dissipation system is insufficient, and the result of the temperature standard deviation judgment also indicates non-pass.

[0121] When the first comprehensive parameter or the second comprehensive parameter is compared with the comprehensive parameter threshold in the configuration file, if any comprehensive parameter exceeds the threshold, the result of the first weighted judgment indicates non-pass, meaning that after comprehensively considering the temperature difference and the temperature standard deviation, the overall heat dissipation performance of the cold plate fails to reach the qualified level.

[0122] Through the embodiments provided in this application, through a series of conditional judgments, it is ensured that the cold plate can be accurately evaluated and effectively identified in the face of various heat dissipation challenges such as high temperature, uneven temperature distribution, large temperature fluctuations, and poor comprehensive performance. This design not only strengthens the reliability of the test results, avoids potential heat dissipation performance problems being ignored, but also provides a multi-level judgment mechanism, making the test process more rigorous and comprehensive.

[0123] As an alternative solution, obtaining multiple first temperature values of the central processing unit and multiple second temperature values of the graphics processing unit when the cold plate is running includes:

[0124] Obtaining multiple first temperature values of the central processing unit and multiple second temperature values of the graphics processing unit when the cold plate is running under the first load condition according to the first sampling frequency;

[0125] After determining that the cold plate meets the first heat dissipation performance condition, the method further includes:

[0126] At the second sampling frequency, when the cold plate is operating under the second load condition, obtain multiple third temperature values of the central processing unit and multiple fourth temperature values of the graphics processing unit, where the second sampling frequency is higher than the first sampling frequency, and the multiple third temperature values and the multiple fourth temperature values are used to determine whether the cold plate meets the second heat dissipation performance condition.

[0127] Optionally, in this embodiment, the first sampling frequency and the second sampling frequency are used to indicate the speed of data acquisition, expressed as the number of readings per unit time. In this embodiment, the first sampling frequency is used for temperature data acquisition under low load conditions, and the second sampling frequency is used under high load conditions, and its frequency is higher than the first sampling frequency to obtain denser temperature data points and more carefully monitor the temperature changes under heavy load.

[0128] Optionally, in this embodiment, the multiple first temperature values and the multiple second temperature values are multiple temperature readings collected for the CPU and GPU under low load conditions, usually data continuously collected multiple times at the first sampling frequency. The multiple third temperature values and the multiple fourth temperature values are temperature readings collected for the CPU and GPU under high load conditions, but collected at a higher second sampling frequency to ensure that temperature changes can be monitored more timely and comprehensively under performance limits.

[0129] Optionally, in this embodiment, when the server is operating under low load, a lower first sampling frequency (for example, once every 1 minute) is used to read the temperature data of the CPU and GPU. This can provide a stable temperature baseline to judge the heat dissipation performance of the cold plate under light load.

[0130] For example, when the server is powered on and running lightweight tasks, the temperature of the CPU and GPU is read through the IPMI interface once every 1 minute for a total of 1 hour, thus obtaining 60 groups of multiple first and second temperature values, including each reading of the CPU temperature and each reading of the GPU temperature.

[0131] Once the preliminary test of the cold plate under low load conditions passes, the system will enter the high load test phase. In this phase, a higher second sampling frequency (for example, once every 30 seconds) is used to collect the temperature data of the CPU and GPU in order to more accurately monitor and capture the temperature changes under high power consumption.

[0132] For example, the server enters the high load mode and starts performing high-performance computing tasks. At the same time, the system reads the temperature data of the CPU and GPU at a frequency of once every 30 seconds for 10 hours, thus obtaining 1200 groups of multiple third and fourth temperature values, including each CPU and GPU temperature reading during the high load test.

[0133] Through the embodiments provided in this application, the heat dissipation performance of the cold plate is evaluated by setting different sampling frequencies under low-load and high-load conditions. This method can more comprehensively reflect the heat dissipation efficiency of the cold plate. Under low-load conditions, since the temperature change is relatively gentle, a relatively low first sampling frequency is used to collect temperature data, which can not only meet the needs of performance evaluation but also reduce the amount of data processing and improve the test efficiency. Under high-load conditions, since the temperature change may be more drastic and rapid, a higher second sampling frequency is used to more accurately capture the instantaneous temperature change, so as to more accurately evaluate the heat dissipation capacity and stability of the cold plate in high-power consumption scenarios.

[0134] As an alternative solution, after obtaining multiple third temperature values of the central processing unit and multiple fourth temperature values of the graphics processing unit when the cold plate is operating under the second load condition according to the second sampling frequency, the method further includes:

[0135] In the case where each temperature value among the multiple third temperature values and the multiple fourth temperature values is less than the second temperature threshold, a second temperature difference judgment is performed on the central processing unit and the graphics processing unit according to the multiple third temperature values and the multiple fourth temperature values;

[0136] In the case where the result of the second temperature difference judgment indicates passing, a second temperature standard deviation judgment is performed on the central processing unit and the graphics processing unit according to the multiple third temperature values and the multiple fourth temperature values;

[0137] In the case where the result of the second temperature standard deviation judgment indicates passing, a second weighted judgment is performed on the central processing unit and the graphics processing unit according to the result of the second temperature difference judgment and the result of the second temperature standard deviation judgment;

[0138] In the case where the result of the second weighted judgment indicates passing, it is determined that the cold plate meets the second heat dissipation performance condition. Optionally, in this embodiment, the multiple third temperature values and the multiple fourth temperature values may but are not limited to multiple temperature readings collected for the central processing unit and the graphics processing unit when the cold plate is operating under the second load condition (high-load condition), and may but are not limited to data continuously collected multiple times at the second sampling frequency.

[0139] Optionally, in this embodiment, the second temperature threshold is the upper limit of the safe temperature set for the CPU and GPU. Exceeding this upper limit will be regarded as insufficient heat dissipation performance. The second temperature threshold may but is not limited to be higher than the first temperature threshold.

[0140] Optionally, in this embodiment, the second temperature difference judgment is used to evaluate whether the maximum difference in the surface temperatures of the CPU and GPU under high-load conditions meets a preset limit, and is used to check the uniformity of the temperature distribution.

[0141] Optionally, in this embodiment, the second temperature standard deviation judgment is used to calculate the standard deviation of the CPU and GPU temperature data, so as to evaluate the stability of the heat dissipation performance, that is, the degree of temperature fluctuation.

[0142] Optionally, in this embodiment, the second weighted judgment is used to combine the evaluation results of the temperature difference and the temperature standard deviation, and perform a comprehensive analysis through predefined weights to determine whether the overall heat dissipation performance of the cold plate meets the second heat dissipation performance standard.

[0143] Optionally, in this embodiment, analyze multiple third temperature values (CPU temperature) and multiple fourth temperature values (GPU temperature) obtained under the current load condition (high load condition), and confirm whether each reading is lower than the preset second temperature threshold. If all readings do not exceed the second temperature threshold, the cold plate initially passes the temperature safety check and can proceed to the next judgment.

[0144] Further analyze the temperature data that has passed the temperature threshold check, calculate the maximum difference between the CPU and GPU surface temperatures, and determine whether this difference is less than the temperature difference limit defined in the configuration file. For example, calculate the temperature differences at each measurement point of the CPU and the temperature differences at each measurement point of the GPU, and evaluate the heat dissipation uniformity of the cold plate by comparing the differences with the limit.

[0145] Calculate the standard deviation of the CPU and GPU temperature data to check the stability of the heat dissipation performance, that is, the degree of fluctuation of the temperature readings. If the standard deviation of the temperature data is less than the standard deviation limit defined in the configuration file, the cold plate also meets the requirements in terms of temperature stability.

[0146] Fuse the results of the temperature difference judgment and the temperature standard deviation judgment by weighting to form a comprehensive evaluation index. For example, if the weights of the temperature difference and the standard deviation are 0.6 and 0.4 respectively, then the comprehensive heat dissipation performance score of the cold plate will be calculated based on the evaluation results of these two indicators. If the comprehensive score meets the preset multi-index fusion limit, the cold plate will be considered to meet the heat dissipation performance standard under low load conditions.

[0147] Finally, if the cold plate performs well in each step of the temperature threshold, temperature difference, temperature standard deviation, and the second weighted judgment, and meets all the set standards, then it can be determined that the heat dissipation performance of the cold plate meets the first heat dissipation performance standard under the current load condition.

[0148] Through the embodiments provided in this application, by collecting the temperature data of the central processing unit (CPU) and the graphics processing unit (GPU) under two working conditions of low load and high load respectively, a comprehensive evaluation of the heat dissipation performance of the cold plate is achieved. First, through the temperature data under the first load condition (such as a lower load), it can be preliminarily judged whether the basic heat dissipation performance of the cold plate meets the standard. If the cold plate has good heat dissipation performance under this light load condition, then further tests are carried out under a higher load condition (i.e., the second load condition) to examine its heat dissipation capacity and stability under heavy load. Through the third temperature data and the fourth temperature data obtained under the high load condition, a more rigorous second heat dissipation performance judgment is carried out to check the heat dissipation efficiency of the cold plate under the extreme working state.

[0149] Generally speaking, the embodiments provided in this application achieve a dynamic evaluation of the heat dissipation performance of the cold plate under different load conditions, ensuring that the cold plate can provide the optimal heat dissipation effect in various working environments, thereby achieving the technical effect of improving the accuracy of the heat dissipation performance evaluation of the cold plate.

[0150] As an alternative solution, the method further includes:

[0151] In the case where the cold plate does not meet the first heat dissipation performance condition or does not meet the second heat dissipation performance condition, it is determined that the cold plate does not meet the expected heat dissipation performance condition;

[0152] Perform a correction operation on the cold plate;

[0153] In the case where the cold plate meets the first heat dissipation performance condition and meets the second heat dissipation performance condition, it is determined that the cold plate meets the expected heat dissipation performance condition.

[0154] Optionally, in this embodiment, the correction operation is used to indicate the problem location, cause analysis, and design or manufacturing correction process when the cold plate fails to meet the expected heat dissipation performance condition, aiming to improve the heat dissipation performance of the cold plate and ensure that it meets the design requirements.

[0155] Optionally, in this embodiment, after the heat dissipation performance judgment under the low load and high load conditions is completed, if the result under any condition indicates that the cold plate does not meet the corresponding heat dissipation performance standard, it can be determined that the cold plate as a whole does not meet the expected heat dissipation performance condition.

[0156] For example, assume that in the low load test, the cold plate fails to pass the temperature difference judgment (i.e., does not meet the first heat dissipation performance condition), even if all indicators are qualified in the high load test; or in the high load test, the comprehensive parameters of the cold plate exceed the threshold (i.e., do not meet the second heat dissipation performance condition), even if it performs well in the low load test; the cold plate will generally be determined not to meet the expected heat dissipation performance condition, and a single problem under either the low load or high load condition will trigger the determination of overall performance failure.

[0157] Once it is determined that the cold plate does not meet the expected heat dissipation performance conditions, the next step is to conduct an in-depth review of the cold plate, analyze the specific reasons for failing the test, such as design defects, material problems, improper assembly, etc., and then make modifications or adjustments according to the review results to improve its heat dissipation performance.

[0158] In response to the review results, it may be necessary to optimize the cold plate structure, adjust the distribution of the heat-conducting medium, or improve the assembly process to ensure that the heat dissipation performance of the cold plate meets the standards under low-load conditions. It may be necessary to adjust the parameter settings of the liquid cooling system, optimize the selection of cold plate materials, or re-design the heat dissipation path to ensure that the heat dissipation performance of the cold plate can meet the expectations when the server is under high computing loads.

[0159] Through the embodiments provided in this application, the closed-loop mechanism of the liquid-cooled server cold plate heat dissipation performance diagnosis method, that is, during the diagnosis process, if the cold plate fails to meet the heat dissipation performance standards under low-load or high-load conditions, it will not only be determined that it does not meet the expected heat dissipation performance conditions, but also further trigger the review and modification process of the cold plate. This mechanism ensures that producers can discover and solve problems in a timely manner, avoiding non-compliant cold plates from entering the market, thereby guaranteeing the stable operation of the server and the safe use of users.

[0160] As an alternative solution, the above-mentioned cold plate heat dissipation performance evaluation method is applied to the liquid-cooled server cold plate heat dissipation performance diagnosis scenario. Under high-load and low-load conditions in the linux system, the sensor information of the CPU and GPU is automatically read through IPMI respectively, and the sensor network and algorithm are used to automatically match parameters to monitor the server heat dissipation status in real time, establishing a multi-dimensional quantitative evaluation system, rather than only monitoring a single temperature threshold, to determine whether the design and assembly of the server CPU liquid-cooled plate and GPU liquid-cooled plate meet the heat dissipation performance requirements.

[0161] Optionally, a flowchart of the implementation of the liquid-cooled server cold plate heat dissipation performance diagnosis is as Figure 3 shown, including:

[0162] S301, during the normal operation test after power-on, (through the IPMI command) read the CPU temperature and GPU temperature sensor (sensor) information, calculate each temperature, and determine whether it meets the heat dissipation performance requirements;

[0163] S302, after power-on, conduct a continuous stress test on the machine, (through the IPMI command) read the CPU temperature and GPU temperature sensor (sensor) information, calculate each temperature, and determine whether it meets the heat dissipation performance requirements.

[0164] Specifically, a flowchart of the implementation of the liquid-cooled server cold plate heat dissipation performance diagnosis is as Figure 4As shown in the figure, it includes:

[0165] S401, start the test;

[0166] S402, collect CPU and GPU temperature information under no-pressure state;

[0167] S403, determine whether the data collected in S402 exceeds the temperature threshold;

[0168] S404, when the data collected in S402 does not exceed the temperature threshold, calculate whether the temperature difference between each CPU and GPU meets the heat dissipation performance requirements;

[0169] S405, when the temperature difference between each CPU and GPU meets the heat dissipation performance requirements in S404, collect CPU and GPU temperature information under the pressure state;

[0170] S406, determine whether the data collected in S405 exceeds the temperature threshold;

[0171] S407, when the data collected in S406 does not exceed the temperature threshold, calculate whether the temperature difference between each CPU and GPU meets the heat dissipation performance requirements;

[0172] S408, when the temperature difference between each CPU and GPU meets the heat dissipation performance requirements in S407, determine that the test is passed.

[0173] It should be noted that in other cases during the above judgment process, it is determined that the test fails.

[0174] For further illustration, according to the cold plate material code and performance requirements, the temperature difference limits ( , ), standard deviation limits ( , ), and multi-index fusion limits ( , ) under low-load and high-load conditions are defined in the configuration file.

[0175] Perform low-load cold plate heat dissipation performance diagnosis to evaluate the cold plate heat dissipation performance when the server is running at low load. During the normal operation state after startup, read the temperature of each CPU and GPU through the ipmi command every 1 minute and output it to the log file, and collect for 1 hour, that is, 60 times. If the real-time temperature value exceeds the threshold during the data collection process then the test fails.

[0176] If the temperature collected in real time does not exceed the threshold , further perform temperature difference data analysis. If a large local temperature difference is measured at the surface of the heating element or different positions of the radiator (for example, the temperature at a certain point is significantly higher than other points), it indicates that the uneven heat distribution may cause local overheating or low heat dissipation efficiency. Therefore, the design evaluates whether the heat dissipation performance requirements are met by measuring the temperature differences at each component measurement point.

[0177] During the above acquisition process, perform a temperature difference calculation on the acquired data every 5 minutes. Read the temperatures of each CPU from the log file and calculate the average value of the acquired temperatures for each CPU 、 、 …, calculate the sum of the average CPU temperatures , sort the average values to find the maximum value and the minimum value , calculate the difference between the maximum value and the minimum value , use the difference and the sum value to divide and calculate the percentage , obtain the material code of the current CPU cold plate from the BOM of the whole machine server production order, and match the corresponding , should be less than , otherwise the heat dissipation performance requirements are not met and the test fails. The specific calculation formula is as follows:

[0178]

[0179]

[0180]

[0181] Similarly, read the temperatures of each GPU from the log file and calculate the average value of the acquired temperatures for each GPU 、 、 …, calculate the sum of the average GPU temperatures , sort the average values to find the maximum value and the minimum value , calculate the difference between the maximum value and the minimum value , use the difference and the sum value to divide and calculate the percentage , obtain the material code of the current GPU cold plate from the BOM of the whole machine server production order, and match the corresponding , should be less than , otherwise the heat dissipation performance requirements are not met and the test fails. The specific calculation formula is as follows:

[0182]

[0183]

[0184]

[0185] If the temperature difference between each CPU and each GPU does not exceed the limit value as described above, further standard deviation data analysis is performed. The temperature standard deviation (σ) is a statistical indicator that measures the degree of dispersion of the temperature distribution and can reflect the uniformity and thermal management efficiency of the heat dissipation system. If there are design defects in the heat dissipation system (such as uneven heat conduction and poor air flow distribution), the temperature in some areas will be significantly higher than that in other areas, resulting in an increase in the standard deviation of the temperature distribution. The larger the standard deviation, the more dispersed the data distribution, and the smaller the standard deviation, the more concentrated the data. Therefore, the design evaluates whether the liquid cooling plate meets the heat dissipation performance requirements through the standard deviation.

[0186] During the above collection process, the standard deviation of each CPU temperature is calculated every 5 minutes for the collected data. The temperatures of each CPU are read from the log file, and the average value and variance of all the collected CPU temperatures are calculated. Average value And variance . The material code of the current CPU cold plate is obtained from the BOM, and the corresponding value is read from the configuration file according to the current material code. , Should be less than , otherwise the heat dissipation performance requirement test fails. The specific calculation formula is as follows:

[0187]

[0188] Similarly, to evaluate the standard deviation of the GPU temperature, the standard deviation of each GPU temperature is calculated every 5 minutes for the collected data. The temperatures of each GPU are read from the log file, and the average value and variance of all the collected GPU temperatures are calculated. Average value And variance , the material code of the current GPU cold plate is obtained from the BOM, and the corresponding value is read from the configuration file according to the current material code. , Should be less than , otherwise the heat dissipation performance requirement test fails. The specific calculation formula is as follows:

[0189]

[0190] Perform linear weighted fusion on the temperature difference and standard deviation calculated every 5 minutes as described above, and match the corresponding value read from the configuration file according to the current material code. , use comprehensive indicators to evaluate whether the heat dissipation performance requirements are met. The specific calculation formula is as follows:

[0191]

[0192] Perform high-load cold plate heat dissipation performance diagnosis to evaluate the cold plate heat dissipation performance when the server is running at high load and high heat generation. Conduct stress tests on the CPU and GPU. During the stress test, the overall load of the server will increase, and the temperatures of each CPU and GPU will also increase accordingly. Read the temperatures of each CPU and GPU once every 30s through the ipmi command and output them to a file. The stress test and data collection will last for 10 hours, that is, 1200 times. If the temperature value exceeds the threshold in the collected data, the test fails. If it does not exceed the threshold, proceed to the next heat dissipation performance analysis.

[0193] If the real-time collected temperatures of each component do not exceed the threshold, further perform temperature difference data analysis to evaluate whether the temperature differences at the measurement points of each component under high load meet the heat dissipation performance requirements.

[0194] During the above collection process, calculate the temperature difference of the collected data once every 5 minutes. Read the temperatures of each CPU from the log file and calculate the average value of the collected temperatures of each CPU , , …, calculate the sum of the average values of the CPU temperatures , sort the average values to find the maximum value and the minimum value , calculate the difference between the maximum value and the minimum value , use the difference divide by the sum value to calculate the percentage , obtain the material code of the current CPU cold plate from the BOM of the overall server production order, and match the corresponding , should be less than , otherwise the heat dissipation performance requirement test fails. The specific calculation formula is as follows:

[0195]

[0196]

[0197]

[0198] Similarly, read the temperatures of each GPU from the log file and calculate the average value of the collected temperatures of each GPU , , …, calculate the sum of the average values of each GPU temperature , sort each average value to find the maximum value and the minimum value , calculate the difference between the maximum value and the minimum value , use the difference to divide by the sum value to calculate the percentage , obtain the material code of the current GPU cold plate from the production order BOM of the whole machine server, and match the corresponding , should be less than , otherwise it does not meet the heat dissipation performance requirements and the test fails. The specific calculation formula is as follows:

[0199]

[0200]

[0201]

[0202] If the temperature differences between the above-mentioned CPUs and GPUs do not exceed the limit value, further standard deviation data analysis is carried out. The larger the standard deviation, the more dispersed the data distribution is, and the smaller the standard deviation, the more concentrated the data is. The liquid cooling plate is evaluated whether it meets the heat dissipation performance requirements through the standard deviation under high load.

[0203] During the above collection process, calculate the standard deviation of each CPU temperature every 5 minutes for the collected data, read the temperature of each CPU from the log file, and calculate the average value of all the collected CPU temperatures and the variance , obtain the material code of the current CPU cold plate from the BOM, and read the corresponding , should be less than , otherwise it does not meet the heat dissipation performance requirements and the test fails. The specific calculation formula is as follows:

[0204]

[0205] Similarly, evaluate the standard deviation of the GPU temperature. Calculate the standard deviation of each GPU temperature every 5 minutes for the collected data, read the temperature of each GPU from the log file, and calculate the average value of all the collected GPU temperatures and the variance , obtain the material code of the current GPU cold plate from the BOM, and read the corresponding , should be less than , otherwise it fails to meet the heat dissipation performance requirement test. The specific calculation formula is as follows:

[0206]

[0207] Perform linear weighted fusion on the temperature difference and standard deviation calculated every 5 minutes above, and use the comprehensive index to evaluate whether the heat dissipation performance requirements are met. The specific calculation formula is as follows:

[0208]

[0209] It should be noted that in the linux system of this embodiment, in combination with the sensor data (such as CPU / GPU temperature) in the high / low load scenarios of the entire server, the temperature difference Δ algorithm, standard deviation σ algorithm, and linear fusion value algorithm are used to analyze and evaluate the heat dissipation performance, and a multi-dimensional quantitative evaluation system is established, rather than only monitoring a single temperature threshold to determine whether the CPU liquid cooling plate and GPU liquid cooling plate of the server meet the heat dissipation performance requirements.

[0210] It can be understood that this embodiment designs a cold plate liquid cooling AI server heat dissipation performance diagnosis method and system, which is used in the server production test and inspection process, effectively intercepts the defects of the liquid cooling plate body and assembly defects, ensures the product quality, improves the test efficiency, test coverage rate and test automation level.

[0211] According to the cold plate material code and performance requirements, the temperature difference limits ( , ), standard deviation limits ( , ), and multi-index fusion limits ( , ) in the low-load and high-load conditions are defined in the configuration file. An example table of the relationship between the material code and limit parameters of a configuration file is shown in Table 1:

[0212] Table 1 Example table of the relationship between the material code and limit parameters of the configuration file

[0213]

[0214] Through the embodiments provided in this application, through multi-parameter monitoring and data analysis, real-time monitoring, fault diagnosis, and performance optimization of the cold plate liquid cooling system of the entire server are achieved, significantly improving the stability and reliability of the liquid-cooled server, which is of great significance for heat dissipation management in a high-performance computing environment. Conduct cold plate temperature uniformity testing and analysis to evaluate the temperature distribution uniformity on the surface of the cold plate to ensure that the cold plate can effectively cool the entire object to be cooled. This private energy can comprehensively and automatically and efficiently diagnose whether the server cold plate meets the heat dissipation performance requirements under the linux system, establish a multi-dimensional quantitative evaluation system, rather than only monitoring a single temperature threshold. Through this diagnosis, the testing efficiency and testing automation level in the mass production of servers are improved, the quality of the shipped server products is guaranteed, and the reliable operation of the server in high-performance computing tasks on the client side is ensured.

[0215] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases, the former is a better implementation method. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc), including several instructions for causing a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of various embodiments of this application.

[0216] In this embodiment, a device for evaluating the heat dissipation performance of a cold plate is also provided. This device is used to implement the above embodiments and preferred implementation manners, and those that have been described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0217] Figure 5 It is a structural block diagram of a device for evaluating the heat dissipation performance of a cold plate according to an embodiment of this application. As Figure 5 shown, the device includes:

[0218] An acquisition unit 502, configured to acquire a plurality of first temperature values of a central processing unit and a plurality of second temperature values of a graphics processing unit when the cold plate is running, where the cold plate is used to dissipate heat for the central processing unit and the graphics processing unit;

[0219] The first determination unit 504 is configured to perform a first temperature difference determination on the central processing unit and the graphics processing unit according to a plurality of first temperature values and a plurality of second temperature values when each of the temperature values among the plurality of first temperature values and the plurality of second temperature values is less than the first temperature threshold;

[0220] The second determination unit 506 is configured to perform a first temperature standard deviation determination on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values when the result indicated by the first temperature difference determination passes;

[0221] The third determination unit 508 is configured to perform a first weighted determination on the central processing unit and the graphics processing unit according to the result of the first temperature difference determination and the result of the first temperature standard deviation determination when the result indicated by the first temperature standard deviation determination passes;

[0222] The determination unit 510 is configured to determine that the cold plate meets the first heat dissipation performance condition when the result indicated by the first weighted determination passes.

[0223] As an optional solution, the first determination unit 504 includes:

[0224] The first determination module is configured to determine, from the plurality of first temperature values, the first temperature values of the respective central processing units in the central processing unit, and to determine, from the plurality of second temperature values, the second temperature values of the respective graphics processing units in the graphics processing unit, wherein the first temperature values of the respective central processing units are used to determine the average temperature values of the respective central processing units, and the second temperature values of the respective graphics processing units are used to determine the average temperature values of the respective graphics processing units;

[0225] The second determination module is configured to determine a first temperature difference parameter corresponding to the central processing unit according to the highest first average temperature value and the lowest second average temperature value among the average temperature values of the respective central processing units;

[0226] The third determination module is configured to determine a second temperature difference parameter corresponding to the graphics processing unit according to the highest third average temperature value and the lowest fourth average temperature value among the average temperature values of the respective graphics processing units;

[0227] The fourth determination module is configured to determine that the result indicated by the first temperature difference determination passes when the first temperature difference parameter is less than the temperature difference threshold and the second temperature difference parameter is less than the temperature difference threshold.

[0228] As an optional solution, the second determination module includes:

[0229] The first calculation sub-module is used to accumulate the average temperature values of each central processing unit to obtain a first accumulated temperature value;

[0230] The second calculation sub-module is used to divide the first difference between the first average temperature value and the second average temperature value by the first accumulated temperature value to obtain a first temperature difference parameter;

[0231] The third determination module includes:

[0232] The third calculation sub-module is used to accumulate the average temperature values of each graphics processing unit to obtain a second accumulated temperature value;

[0233] The fourth calculation sub-module is used to divide the second difference between the third average temperature value and the fourth average temperature value by the second accumulated temperature value to obtain a second temperature difference parameter.

[0234] As an optional solution, the second judgment unit 506 includes:

[0235] The first calculation module is used to perform a standard deviation operation on the average temperature values of each central processing unit in the central processing unit to obtain a first temperature standard deviation parameter when the average temperature values of each central processing unit in the central processing unit are obtained;

[0236] The second calculation module is used to perform a standard deviation operation on the average temperature values of each graphics processing unit in the graphics processing unit to obtain a second temperature standard deviation parameter when the average temperature values of each graphics processing unit in the graphics processing unit are obtained;

[0237] The fifth determination module is used to determine that the result indication of the first temperature standard deviation judgment passes when the first temperature standard deviation parameter is less than the temperature standard deviation threshold and the second temperature standard deviation parameter is less than the temperature standard deviation threshold.

[0238] As an optional solution, the third judgment unit 508 includes:

[0239] The first acquisition module is used to acquire the first temperature difference parameter of the central processing unit and the second temperature difference parameter of the graphics processing unit according to the result of the first temperature difference judgment, and acquire the first temperature standard deviation parameter of the central processing unit and the second temperature standard deviation parameter of the graphics processing unit according to the result of the first temperature standard deviation judgment;

[0240] The first weighting module is used to perform a first weighted fusion operation on the first temperature difference parameter and the first temperature standard deviation parameter to obtain a first comprehensive parameter of the central processing unit;

[0241] The second weighting module is used to perform a second weighted fusion operation on the second temperature difference parameter and the second temperature standard deviation parameter to obtain a second comprehensive parameter of the graphics processing unit;

[0242] A sixth determination module, configured to determine that the result indication of the first weighted determination passes when the first comprehensive parameter is less than the comprehensive parameter threshold and the second comprehensive parameter is less than the comprehensive parameter threshold.

[0243] As an optional solution, the first weighting module includes:

[0244] A first acquisition sub-module, configured to acquire a first weighting parameter corresponding to a central processing unit and a second weighting parameter corresponding to a graphics processing unit, and acquire the sum value of the weighting parameters of the first weighting parameter and the second weighting parameter;

[0245] A second acquisition sub-module, configured to acquire a first cumulative parameter obtained by multiplying the first weighting parameter by a first temperature difference parameter, and acquire a second cumulative parameter obtained by multiplying the second weighting parameter by a first temperature standard deviation parameter;

[0246] A fifth calculation sub-module, configured to determine the sum value of the first cumulative parameter and the second cumulative parameter divided by the sum value of the weighting parameters as the first comprehensive parameter;

[0247] The second weighting module includes:

[0248] A third acquisition sub-module, configured to acquire a third cumulative parameter obtained by multiplying the first weighting parameter by a second temperature difference parameter, and acquire a fourth cumulative parameter obtained by multiplying the second weighting parameter by a second temperature standard deviation parameter;

[0249] A sixth calculation sub-module, configured to determine the sum value of the third cumulative parameter and the fourth cumulative parameter divided by the sum value of the weighting parameters as the second comprehensive parameter.

[0250] As an optional solution, the device further includes:

[0251] A second acquisition module, configured to acquire the target material type of the cold plate before performing a first temperature difference determination on the central processing unit and the graphics processing unit according to a plurality of first temperature values and a plurality of second temperature values;

[0252] A seventh determination module, configured to determine, before performing a first temperature difference determination on the central processing unit and the graphics processing unit according to a plurality of first temperature values and a plurality of second temperature values, a temperature difference threshold corresponding to the target material type, a temperature standard deviation threshold corresponding to the target material type, and a comprehensive parameter threshold corresponding to the target material type from a configuration file, where the temperature difference threshold is used for the first temperature difference determination, the temperature standard deviation threshold is used for the first temperature standard deviation determination, and the comprehensive parameter threshold is used for the first weighted determination, and the configuration file stores multiple sets of corresponding relationships between material types and thresholds.

[0253] As an optional solution, the device further includes:

[0254] An eighth determination module, configured to determine that the cold plate does not meet the first heat dissipation performance condition when there is a temperature value greater than or equal to a preset threshold among multiple first temperature values and multiple second temperature values, or when the result of the first temperature difference determination indicates failure, or when the result of the first temperature standard deviation determination indicates failure, or when the result of the first weighted determination indicates failure.

[0255] As an optional solution, the acquisition unit includes:

[0256] A third acquisition module, configured to acquire multiple first temperature values of the central processing unit and multiple second temperature values of the graphics processing unit of the cold plate when the cold plate operates under a first load condition according to a first sampling frequency;

[0257] The device further includes:

[0258] A fourth acquisition module, configured to, after determining that the cold plate meets the first heat dissipation performance condition, acquire multiple third temperature values of the central processing unit and multiple fourth temperature values of the graphics processing unit of the cold plate when the cold plate operates under a second load condition according to a second sampling frequency, where the second sampling frequency is higher than the first sampling frequency, and the multiple third temperature values and the multiple fourth temperature values are used to determine whether the cold plate meets the second heat dissipation performance condition.

[0259] As an optional solution, the device further includes:

[0260] A first judgment module, configured to, after acquiring multiple third temperature values of the central processing unit and multiple fourth temperature values of the graphics processing unit of the cold plate when the cold plate operates under a second load condition according to the second sampling frequency, perform a second temperature difference judgment on the central processing unit and the graphics processing unit according to the multiple third temperature values and the multiple fourth temperature values when each temperature value among the multiple third temperature values and the multiple fourth temperature values is less than a second temperature threshold;

[0261] A second judgment module, configured to, after acquiring multiple third temperature values of the central processing unit and multiple fourth temperature values of the graphics processing unit of the cold plate when the cold plate operates under a second load condition according to the second sampling frequency, perform a second temperature standard deviation judgment on the central processing unit and the graphics processing unit according to the multiple third temperature values and the multiple fourth temperature values when the result of the second temperature difference judgment indicates passing;

[0262] A third judgment module, configured to, after obtaining multiple third temperature values of a central processing unit and multiple fourth temperature values of a graphics processing unit when a cold plate operates under a second load condition according to a second sampling frequency, and when the result indication of the second temperature standard deviation judgment passes, perform a second weighted judgment on the central processing unit and the graphics processing unit according to the result of the second temperature difference judgment and the result of the second temperature standard deviation judgment;

[0263] A ninth determination module, configured to determine that the cold plate meets the second heat dissipation performance condition when the result indication of the second weighted judgment passes.

[0264] As an optional solution, the device further includes:

[0265] A tenth determination module, configured to determine that the cold plate does not meet the expected heat dissipation performance condition when the cold plate does not meet the first heat dissipation performance condition or does not meet the second heat dissipation performance condition;

[0266] A correction module, configured to perform a correction operation on the cold plate;

[0267] An eleventh determination module, configured to determine that the cold plate meets the expected heat dissipation performance condition when the cold plate meets the first heat dissipation performance condition and meets the second heat dissipation performance condition.

[0268] Specific examples in this embodiment may refer to the examples described in the above embodiments and exemplary implementation manners, and will not be elaborated herein.

[0269] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation manner. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc), and includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of the various embodiments of the present application.

[0270] It should be noted that the above-mentioned various modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited thereto: the above-mentioned modules are all located in the same processor; or, the above-mentioned various modules are respectively located in different processors in any combination form.

[0271] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored, and the computer program is configured to execute the steps in any one of the above method embodiments when running.

[0272] In an exemplary embodiment, the above computer-readable storage medium may include, but is not limited to: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), external hard drives, magnetic disks, or optical discs that can store computer programs.

[0273] An embodiment of the present application also provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0274] In an exemplary embodiment, the above electronic device may further include a transmission device and input / output devices. Among them, the transmission device is connected to the above processor, and the input / output devices are connected to the above processor.

[0275] An embodiment of the present application also provides a computer program product, including a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium stores a computer program product. When the computer program is executed by a processor, the steps of the methods in various embodiments of the present application are implemented.

[0276] Specific examples in this embodiment may refer to the examples described in the above embodiments and exemplary embodiments, and will not be repeated here.

[0277] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. They can be implemented by program codes executable by the computing device. Thus, they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a different order than here, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module to be implemented. In this way, the present application is not limited to any specific combination of hardware and software.

[0278] The above are only the preferred embodiments of the present application and are not used to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for evaluating the heat dissipation performance of a cold plate, characterized in that, Including: Obtaining a plurality of first temperature values of a central processing unit and a plurality of second temperature values of a graphics processing unit when a cold plate is operating, wherein the cold plate is used to dissipate heat for the central processing unit and the graphics processing unit; When each temperature value among the plurality of first temperature values and the plurality of second temperature values is less than a first temperature threshold, making a first temperature difference judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values; When the result of the first temperature difference judgment indicates passing, making a first temperature standard deviation judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values; When the result of the first temperature standard deviation judgment indicates passing, making a first weighted judgment on the central processing unit and the graphics processing unit according to the result of the first temperature difference judgment and the result of the first temperature standard deviation judgment; When the result of the first weighted judgment indicates passing, determining that the cold plate meets the first heat dissipation performance condition.

2. The method according to claim 1, characterized in that, The making a first temperature difference judgment on the central processing unit and the graphics processing unit according to the plurality of first temperature values and the plurality of second temperature values includes: Determining, from the plurality of first temperature values, the first temperature values of each central processing unit in the central processing unit, and determining, from the plurality of second temperature values, the second temperature values of each graphics processing unit in the graphics processing unit, wherein the first temperature values of each central processing unit are used to determine the average temperature value of each central processing unit, and the second temperature values of each graphics processing unit are used to determine the average temperature value of each graphics processing unit; According to the first average temperature value with the highest average temperature value and the second average temperature value with the lowest average temperature value among the average temperature values of each central processing unit, determining a first temperature difference parameter corresponding to the central processing unit; According to the third average temperature value with the highest average temperature value and the fourth average temperature value with the lowest average temperature value among the average temperature values of each graphics processing unit, determining a second temperature difference parameter corresponding to the graphics processing unit; When the first temperature difference parameter is less than a temperature difference threshold and the second temperature difference parameter is less than the temperature difference threshold, determining that the result of the first temperature difference judgment indicates passing.

3. The method according to claim 2, wherein The determining a first temperature difference parameter corresponding to the central processing unit according to the first average temperature value with the highest average temperature value and the second average temperature value with the lowest average temperature value among the average temperature values of each central processing unit includes: Accumulating the average temperature values of each central processing unit to obtain a first accumulated temperature value; Dividing the first difference between the first average temperature value and the second average temperature value by the first accumulated temperature value to obtain the first temperature difference parameter; Determining the second temperature difference parameter corresponding to the graphics processor according to the third average temperature value with the highest average temperature value and the fourth average temperature value with the lowest average temperature value among the average temperature values of the respective graphics processors includes: Accumulating the average temperature values of the respective graphics processors to obtain a second accumulated temperature value; Dividing the second difference between the third average temperature value and the fourth average temperature value by the second accumulated temperature value to obtain the second temperature difference parameter.

4. The method according to claim 1, wherein Performing a first temperature standard deviation judgment on the central processing unit and the graphics processor according to the multiple first temperature values and the multiple second temperature values includes: When the average temperature values of the respective central processing units in the central processing unit are obtained, performing a standard deviation operation on the average temperature values of the respective central processing units to obtain a first temperature standard deviation parameter; When the average temperature values of the respective graphics processors in the graphics processor are obtained, performing a standard deviation operation on the average temperature values of the respective graphics processors to obtain a second temperature standard deviation parameter; When the first temperature standard deviation parameter is less than the temperature standard deviation threshold and the second temperature standard deviation parameter is less than the temperature standard deviation threshold, determining that the result indication of the first temperature standard deviation judgment passes.

5. The method according to claim 1, wherein Performing a first weighted judgment on the central processing unit and the graphics processor according to the result of the first temperature difference judgment and the result of the first temperature standard deviation judgment includes: According to the result of the first temperature difference judgment, obtaining the first temperature difference parameter of the central processing unit and the second temperature difference parameter of the graphics processor, and according to the result of the first temperature standard deviation judgment, obtaining the first temperature standard deviation parameter of the central processing unit and the second temperature standard deviation parameter of the graphics processor; Performing a first weighted fusion operation on the first temperature difference parameter and the first temperature standard deviation parameter to obtain a first comprehensive parameter of the central processing unit; Performing a second weighted fusion operation on the second temperature difference parameter and the second temperature standard deviation parameter to obtain a second comprehensive parameter of the graphics processor; When the first comprehensive parameter is less than the comprehensive parameter threshold and the second comprehensive parameter is less than the comprehensive parameter threshold, determining that the result indication of the first weighted judgment passes.

6. The method according to claim 5, wherein The performing a first weighted fusion operation on the first temperature difference parameter and the first temperature standard deviation parameter to obtain a first comprehensive parameter of the central processing unit includes: Obtaining a first weighted parameter corresponding to the central processing unit and a second weighted parameter corresponding to the graphics processor, and obtaining a weighted parameter sum value of the first weighted parameter and the second weighted parameter; Obtaining a first cumulative parameter obtained by multiplying the first weighted parameter by the first temperature difference parameter, and obtaining a second cumulative parameter obtained by multiplying the second weighted parameter by the first temperature standard deviation parameter; Determine the sum of the first cumulative parameter and the second cumulative parameter, and divide the result by the sum of the weighted parameters to obtain the first comprehensive parameter; The second weighted fusion operation on the second temperature difference parameter and the second temperature standard deviation parameter to obtain the second comprehensive parameter of the graphics processing unit includes: Obtain a third cumulative parameter obtained by multiplying the first weighted parameter by the second temperature difference parameter, and obtain a fourth cumulative parameter obtained by multiplying the second weighted parameter by the second temperature standard deviation parameter; Determine the sum of the third cumulative parameter and the fourth cumulative parameter, and divide the result by the sum of the weighted parameters to obtain the second comprehensive parameter.

7. The method according to claim 1, characterized in that, Before the first temperature difference judgment on the central processing unit and the graphics processing unit according to the multiple first temperature values and the multiple second temperature values, the method further includes: Obtain the target material type of the cold plate; Determine the temperature difference threshold corresponding to the target material type, the temperature standard deviation threshold corresponding to the target material type, and the comprehensive parameter threshold corresponding to the target material type from the configuration file, where the temperature difference threshold is used for the first temperature difference judgment, the temperature standard deviation threshold is used for the first temperature standard deviation judgment, and the comprehensive parameter threshold is used for the first weighted judgment. Multiple sets of corresponding relationships between material types and thresholds are stored in the configuration file.

8. The method according to claim 1, wherein The method further includes: When there is a temperature value greater than or equal to the first preset threshold among the multiple first temperature values and the multiple second temperature values, or when the result of the first temperature difference judgment indicates failure, or when the result of the first temperature standard deviation judgment indicates failure, or when the result of the first weighted judgment indicates failure, determine that the cold plate does not meet the first heat dissipation performance condition.

9. The method according to claim 1, wherein The step of obtaining multiple first temperature values of the central processing unit and multiple second temperature values of the graphics processing unit when the cold plate is running includes: Obtain the multiple first temperature values of the central processing unit and the multiple second temperature values of the graphics processing unit when the cold plate is running under the first load condition according to the first sampling frequency; After determining that the cold plate meets the first heat dissipation performance condition, the method further includes: Obtain multiple third temperature values of the central processing unit and multiple fourth temperature values of the graphics processing unit when the cold plate is running under the second load condition according to the second sampling frequency, where the second sampling frequency is higher than the first sampling frequency, and the multiple third temperature values and the multiple fourth temperature values are used to determine whether the cold plate meets the second heat dissipation performance condition.

10. The method according to claim 9, characterized in that, After obtaining the multiple third temperature values of the central processing unit and the multiple fourth temperature values of the graphics processing unit when the cold plate is running under the second load condition according to the second sampling frequency, the method further includes: When each of the multiple third temperature values and the multiple fourth temperature values is less than a second temperature threshold, a second temperature difference judgment is performed on the central processing unit and the graphics processing unit according to the multiple third temperature values and the multiple fourth temperature values; When the result of the second temperature difference judgment indicates passing, a second temperature standard deviation judgment is performed on the central processing unit and the graphics processing unit according to the multiple third temperature values and the multiple fourth temperature values; When the result of the second temperature standard deviation judgment indicates passing, a second weighted judgment is performed on the central processing unit and the graphics processing unit according to the result of the second temperature difference judgment and the result of the second temperature standard deviation judgment; When the result of the second weighted judgment indicates passing, it is determined that the cold plate meets the second heat dissipation performance condition.

11. The method according to claim 10, wherein, The method further includes: When the cold plate does not meet the first heat dissipation performance condition or does not meet the second heat dissipation performance condition, it is determined that the cold plate does not meet the expected heat dissipation performance condition; Perform a correction operation on the cold plate; When the cold plate meets the first heat dissipation performance condition and meets the second heat dissipation performance condition, it is determined that the cold plate meets the expected heat dissipation performance condition.

12. A heat dissipation performance evaluation device for a cold plate, characterized in that, Includes: An acquisition unit configured to acquire multiple first temperature values of a central processing unit and multiple second temperature values of a graphics processing unit when the cold plate is operating, where the cold plate is used to dissipate heat for the central processing unit and the graphics processing unit; A first judgment unit configured to perform a first temperature difference judgment on the central processing unit and the graphics processing unit according to the multiple first temperature values and the multiple second temperature values when each of the multiple first temperature values and the multiple second temperature values is less than a first temperature threshold; A second judgment unit configured to perform a first temperature standard deviation judgment on the central processing unit and the graphics processing unit according to the multiple first temperature values and the multiple second temperature values when the result of the first temperature difference judgment indicates passing; A third judgment unit configured to perform a first weighted judgment on the central processing unit and the graphics processing unit according to the result of the first temperature difference judgment and the result of the first temperature standard deviation judgment when the result of the first temperature standard deviation judgment indicates passing; A determination unit configured to determine that the cold plate meets the first heat dissipation performance condition when the result of the first weighted judgment indicates passing.

13. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, where the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 11.

14. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 11.

15. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 11.

Images