Multi-core server heat dissipation control method, system and device and storage medium

Abstract

The invention provides a multi-core server heat dissipation control method, system and device and a storage medium. The method comprises the steps that firstly, flow information of cooling liquid at different environment temperatures is obtained; therefore, the turbulent flow fluctuation characteristics of the cooling liquid at different environment temperatures are extracted, and the cold energy loss of the cooling liquid at different environment temperatures is determined. Further determining the difference entropy of the liquid flow rate of the cooling liquid, and then determining the cold load capacity of the target multi-core server in the heat dissipation process; determining a heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information, and then determining the heat transfer strength of each central processing unit in the heat dissipation process; and further determining the cooling cost of each central processing unit in the heat dissipation process, and carrying out dynamic cooling regulation and control on the cooling liquid in the target multi-core server based on all the cooling costs. By adopting the scheme of the invention, the influence of a thermal asymmetry phenomenon on the heat dissipation effect of the liquid cooling multi-core server can be reduced, so that the efficiency of the cooling system is improved.

Description

Technical Field

[0001] The present application relates to the field of heat dissipation control technology, and more specifically, to a heat dissipation control method, system, device and storage medium for a multi-core server. Background Art

[0002] Heat dissipation control is crucial in the computer field. Its main purpose is to effectively manage and discharge heat during high-load operation to ensure the performance and life of the equipment. It includes air cooling, liquid cooling, and heat pipes. Air cooling uses a combination of heat sinks and fans to remove heat through air convection. It has a simple structure but limited effect in high-performance equipment. The liquid cooling system removes heat through the circulation of coolant and has good heat dissipation efficiency. It is suitable for high-end equipment. Heat pipe technology uses liquid evaporation and condensation phase change to conduct heat. With the continuous improvement of computing power, heat dissipation technology is moving towards more efficient, low-noise and energy-saving directions, such as hybrid systems of liquid cooling and air cooling, advanced artificial intelligence algorithm control, and cold plate technology for extreme environments. These technologies work together to provide a comprehensive heat dissipation solution for modern computing devices.

[0003] The heat dissipation control of liquid-cooled multi-core servers absorbs and conducts heat through circulating coolant, and is mainly implemented by a combination of cooling blocks, pumps, radiators and fans. The coolant absorbs heat from the heat source through the cooling block, is pushed by the pump to flow through the radiator for heat dissipation, and then circulates back to the cooling block for continued cooling. However, due to the different power consumption of different central processing units inside the liquid-cooled multi-core server, especially in high-performance computing tasks, excessive heat will be concentrated in the area where some central processing units are located in the liquid-cooled multi-core server, resulting in thermal asymmetry. Among them, the thermal asymmetry phenomenon refers to the uneven heat dissipation caused by the different heat loads of each central processing unit inside the liquid-cooled multi-core server, resulting in local hotspots in the area where the central processing unit is located, causing heat accumulation, further aggravating the temperature rise in this area. Therefore, how to reduce the impact of thermal asymmetry on the heat dissipation effect of liquid-cooled multi-core servers and thereby improve the efficiency of the cooling system has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a multi-core server heat dissipation control method, system, device and storage medium, which can reduce the impact of thermal asymmetry on the heat dissipation effect of a liquid-cooled multi-core server, thereby improving the efficiency of the cooling system.

[0005] In a first aspect, the present application provides a multi-core server heat dissipation control method, comprising the following steps: Collect the flow information of the coolant of the cooling block in the target multi-core server under different ambient temperatures; Extracting the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all the flow information, and determining the cooling energy loss of the coolant at different ambient temperatures according to all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process; Determine the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block, and determine the cooling load of the target multi-core server during the heat dissipation process according to the differential entropy and all cooling energy losses; Acquire heat exchange information between the cooling block and each central processing unit in the target multi-core server, and then determine the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information, and determine the heat transfer intensity of each central processing unit in the heat dissipation process through all the heat exchange coefficients and the temperature gradient of each central processing unit; The cooling cost of each central processing unit in the heat dissipation process is determined according to all heat transfer intensities and the cooling load, and the cooling liquid in the target multi-core server is dynamically cooled and regulated based on all cooling costs.

[0006] In some embodiments, extracting the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all flow information specifically includes: Determine an inlet dynamic flow rate interval and an outlet dynamic flow rate interval of the coolant in the target multi-core server according to all flow information; Performing differential processing on each flow information to obtain a flow velocity fluctuation group of the coolant at different ambient temperatures; The turbulent flow fluctuation characteristics of the coolant at different ambient temperatures are determined according to the inlet dynamic flow velocity interval, the outlet dynamic flow velocity interval and all flow velocity fluctuation groups.

[0007] In some embodiments, determining the cooling energy loss of the coolant at different ambient temperatures according to all the disturbance fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process specifically includes: Perform linear fitting on all the spoiler fluctuation characteristics to obtain spoiler fluctuation fitting curve; Determining the condensation heat exchange degree of the coolant at different ambient temperatures according to the turbulence fluctuation fitting curve; The cooling energy loss of the coolant at different ambient temperatures is determined through all condensation heat exchange degrees and the temperature change trends of each central processing unit in the target multi-core server during the heat dissipation process.

[0008] In some embodiments, determining the cooling load of the target multi-core server during the heat dissipation process according to the differential entropy and all cooling energy losses specifically includes: Determine the global cooling loss of the target multi-core server during the cooling process according to the preset sliding window and all cooling energy losses; The cooling load of the target multi-core server during the heat dissipation process is determined according to the difference entropy and the global cooling loss.

[0009] In some embodiments, determining the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information specifically includes: Determine the heat diffusion characteristic sequence and thermal boundary offset between the cooling block and each central processing unit according to all the heat exchange information; The heat transfer coefficient between the cooling block and each central processing unit is determined by all heat diffusion feature sequences and all thermal boundary offsets.

[0010] In some embodiments, determining the heat transfer intensity of each CPU during the heat dissipation process through all heat transfer coefficients and the temperature gradient of each CPU specifically includes: Determine the thermal attenuation of the target multi-core server during the heat dissipation process based on all heat transfer coefficients; The heat transfer intensity of each CPU during the heat dissipation process is determined by the heat attenuation degree and the temperature gradient of each CPU.

[0011] In some embodiments, determining the cooling cost of each CPU during the heat dissipation process according to all heat transfer intensities and the cooling load specifically includes: Convert all heat transfer intensities into a heat transfer intensity sequence; Determining the heat dissipation cost of each central processing unit according to the heat transfer intensity sequence and the cooling load; Obtain the cooling energy loss of the coolant at different ambient temperatures; The cooling cost of each CPU during the heat dissipation process is determined through all heat dissipation costs and the cooling energy loss of the coolant at different ambient temperatures.

[0012] In a second aspect, the present application provides a multi-core server heat dissipation control system, comprising: A collection module, used to collect flow information of coolant of a cooling block in a target multi-core server at different ambient temperatures; A processing module, used to extract the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all the flow information, and determine the cooling energy loss of the coolant at different ambient temperatures according to all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process; The processing module is further used to determine the difference entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block, and determine the cooling load of the target multi-core server during the heat dissipation process according to the difference entropy and all cooling energy losses; The processing module is further used to obtain heat exchange information between the cooling block and each central processing unit in the target multi-core server, and then determine the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information, and determine the heat transfer intensity of each central processing unit in the heat dissipation process through all the heat exchange coefficients and the temperature gradient of each central processing unit; The execution module is used to determine the cooling cost of each central processing unit during the heat dissipation process according to all heat transfer intensities and the cooling load, and dynamically cool the coolant in the target multi-core server based on all cooling costs.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the multi-core server heat dissipation control method described above.

[0014] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the multi-core server heat dissipation control method described above is implemented.

[0015] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects: In the multi-core server heat dissipation control method, system, device and storage medium provided by the present application, flow information of the coolant of the cooling block in the target multi-core server at different ambient temperatures is first collected; the turbulence fluctuation characteristics of the coolant at different ambient temperatures are extracted from all the flow information, and the cold energy loss of the coolant at different ambient temperatures is determined according to all the turbulence fluctuation characteristics and the temperature change trend of each central processing unit in the target multi-core server during the heat dissipation process; the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block is determined, and the cold load of the target multi-core server during the heat dissipation process is determined according to the differential entropy and all the cold energy losses; the heat exchange information between the cooling block and each central processing unit in the target multi-core server is obtained, and then the heat exchange coefficient between the cooling block and each central processing unit is determined according to all the heat exchange information, and the heat transfer intensity of each central processing unit during the heat dissipation process is determined by all the heat exchange coefficients and the temperature gradient of each central processing unit; the cooling cost of each central processing unit during the heat dissipation process is determined according to all the heat transfer intensities and the cold load, and the coolant in the target multi-core server is dynamically cooled and regulated based on all the cooling costs.

[0016] It can be seen that in the present application, the cooling cost of each central processing unit in the heat dissipation process can be determined according to all the heat transfer intensities and the cooling load; wherein, firstly, the turbulence fluctuation characteristics of the coolant at different ambient temperatures are extracted from all the flow information, and the turbulence fluctuation characteristics can help to adjust the flow path and flow distribution of the coolant in the multi-core server in real time to ensure that the high heat load area is sufficiently cooled while avoiding excessive cooling of the low heat load area, and then all the turbulence fluctuation characteristics are fitted, and the fitting results are quantified with the temperature change trend of each central processing unit in the multi-core server during the heat dissipation process, and the cooling energy loss of the coolant at different ambient temperatures is obtained, wherein The cooling energy loss indicates that the energy consumed by the coolant in the cooling system when passing through the central processing unit fails to effectively reduce the energy lost by the heat load of the central processing unit. By reducing the cooling energy loss, the overall performance of the cooling system can be improved to ensure that the coolant can effectively take away the heat generated by the processor. Secondly, the difference entropy of the liquid flow rate between the coolant inlet and the coolant outlet is used to extract from all the cooling energy losses the total heat accumulated due to thermal asymmetry when the liquid cooling system cools each central processing unit in a multi-core server. The cooling load helps to evaluate and optimize the ability of the cooling system to cope with thermal asymmetry. Through effective cooling control, the formation of local hot spots can be reduced to avoid excessive temperatures caused by heat accumulation. Then, the heat transfer coefficient between the cooling block and each central processing unit is determined, wherein the heat transfer coefficient reflects the heat dissipation efficiency of the cooling block at the central processing unit. Then, all the heat transfer coefficients and the temperature gradients of each central processing unit are quantified to obtain the heat transfer intensity that measures the heat transfer of the central processing unit during the heat dissipation process. Again, the heat transfer intensity of each central processing unit during the heat dissipation process is extracted from all the heat transfer intensities. The cooling cost in the process, wherein the cooling cost refers to the energy and resources required to achieve effective heat dissipation of a multi-core server at different ambient temperatures, including the energy consumption of the coolant, the flow regulation overhead, and the additional cost required to overcome the thermal asymmetry of the system during the heat dissipation process. The cooling cost can be used to identify the thermal asymmetry areas that require additional attention, thereby reducing local overheating through technical means, and then determining the cooling cost of each central processing unit in the heat dissipation process through all the heat dissipation costs and the cold energy loss of the coolant at different ambient temperatures. The cooling cost refers to the cost required to achieve the best cooling effect for each central processing unit in a multi-core server at different ambient temperatures. The cooling cost can be used to evaluate the cooling needs of each central processing unit, dynamically adjust the flow and distribution of the coolant, thereby effectively reducing the impact of thermal asymmetry on the overall heat dissipation effect. Finally, based on all the cooling costs, the coolant in the target multi-core server is dynamically cooled and regulated; in summary, the solution of the present application can reduce the impact of thermal asymmetry on the heat dissipation effect of a liquid-cooled multi-core server, thereby improving the efficiency of the cooling system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is an exemplary flow chart of a multi-core server heat dissipation control method according to some embodiments of the present application; Figure 2 is a schematic diagram of a process for determining cooling energy loss in some embodiments of the present application; Figure 3 is a schematic diagram of a process for determining cooling cost in some embodiments of the present application; Figure 4 is a structural block diagram of a multi-core server heat dissipation control system in some embodiments of the present application; Figure 5 It is a structural schematic diagram of a computer device for implementing a multi-core server heat dissipation control method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] The core of this application is to first collect the flow information of the coolant of the cooling block in the target multi-core server at different ambient temperatures; extract the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all the flow information, and determine the cold energy loss of the coolant at different ambient temperatures according to all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process; determine the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block, and determine the cold load of the target multi-core server during the heat dissipation process according to the differential entropy and all the cold energy losses; obtain the heat exchange information between the cooling block and each central processor in the target multi-core server, and then determine the heat transfer coefficient between the cooling block and each central processor according to all the heat exchange information, and determine the heat transfer intensity of each central processor during the heat dissipation process through all the heat transfer coefficients and the temperature gradient of each central processor; determine the cooling cost of each central processor during the heat dissipation process according to all the heat transfer intensities and the cold load, and dynamically cool the coolant in the target multi-core server based on all the cooling costs. The scheme of this application can reduce the influence of thermal asymmetry on the heat dissipation effect of the liquid-cooled multi-core server, thereby improving the efficiency of the cooling system.

[0019] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 , which is an exemplary flow chart of a multi-core server heat dissipation control method according to some embodiments of the present application. The multi-core server heat dissipation control method 100 mainly includes the following steps: In step 101, flow information of a cooling liquid of a cooling block in a target multi-core server at different ambient temperatures is collected.

[0020] In specific implementation, the flow information of the coolant of the cooling block within a specified time period under different ambient temperatures is obtained from the database of the cooling system in the target multi-core server, where the specified time period is the most recent hour. Other specified time periods may be used in other embodiments without limitation.

[0021] It should be noted that the cooling block described in the present application is a core component of the liquid cooling system in the target multi-core server, which is in direct contact with each central processing unit in the target multi-core server. The cooling block is usually made of a high thermal conductivity material (such as copper, aluminum, etc.). The cooling block is composed of a plurality of independent tortuous channels and each tortuous channel is in direct contact with the central processing unit, wherein each tortuous channel is connected to the main channel, and the coolant flow control valve arranged on the tortuous channel controls the coolant in the main channel to flow to the channel, absorbs heat from the central processing unit, and then takes the heat away.

[0022] In addition, it should be noted that the flow information described in the present application is a set consisting of an inlet velocity sequence and an outlet velocity sequence of the coolant at a specified ambient temperature, wherein the inlet velocity sequence represents an ordered set of inlet velocities collected at an interval of 10 seconds at the coolant inlet, and the outlet velocity sequence represents an ordered set of outlet velocities collected at an interval of 10 seconds at the coolant outlet. In addition, one ambient temperature corresponds to one flow information, and the total number of inlet velocities of the inlet velocity sequence in each flow information is the same, and the total number of outlet velocities of the outlet velocity sequence in each flow information is also the same.

[0023] In step 102, the turbulence fluctuation characteristics of the coolant at different ambient temperatures are extracted from all the flow information, and the cooling energy loss of the coolant at different ambient temperatures is determined based on all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process.

[0024] In some embodiments, extracting the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all flow information can be achieved by using the following steps: Determine an inlet dynamic flow rate interval and an outlet dynamic flow rate interval of the coolant in the target multi-core server according to all flow information; Performing differential processing on each flow information to obtain a flow velocity fluctuation group of the coolant at different ambient temperatures; The turbulent flow fluctuation characteristics of the coolant at different ambient temperatures are determined according to the inlet dynamic flow velocity interval, the outlet dynamic flow velocity interval and all flow velocity fluctuation groups.

[0025] In specific implementation, the inlet dynamic flow rate interval and outlet dynamic flow rate interval of the coolant in the target multi-core server are determined according to all flow information, which can be implemented in the following manner, namely: first, the DBSCAN clustering algorithm is used to cluster all the inlet speeds in all the flow information, so as to obtain multiple clusters, and the obtained clusters are all used as inlet speed clusters, and the maximum inlet speed and the minimum inlet speed are selected from each inlet speed cluster, and the mean of all the selected maximum inlet speeds is calculated, and the obtained mean is used as the maximum inlet speed, and then the mean of all the selected minimum inlet speeds is calculated, and the obtained mean is used as the minimum inlet speed, so that the range consisting of the minimum inlet speed and the maximum inlet speed is used as the cooling fluid in the target multi-core server. The inlet dynamic flow velocity interval of the liquid is then clustered by the DBSCAN clustering algorithm to cluster all the outlet velocities in all the flow information, so as to obtain multiple clusters, and the obtained clusters are all used as outlet velocity clusters, and the maximum outlet velocity and the minimum outlet velocity are selected from each inlet velocity cluster, and the mean of all the selected maximum outlet velocities is calculated, and the obtained mean is used as the maximum outlet velocity, and then the mean of all the selected minimum outlet velocities is calculated, and the obtained mean is used as the minimum outlet velocity, so that the range composed of the minimum outlet velocity and the maximum outlet velocity is used as the outlet dynamic flow velocity interval of the coolant in the target multi-core server. In other embodiments, other methods can also be used to achieve this, which is not limited here.

[0026] In specific implementation, each flow information is differentially processed to obtain the flow velocity fluctuation group of the coolant at different ambient temperatures, which can be achieved in the following manner, namely: select an ambient temperature as the selected ambient temperature, perform first-order difference on the inlet velocity sequence and the outlet velocity sequence in the flow information corresponding to the selected ambient temperature, and use the sequence obtained after the first-order difference of the inlet velocity sequence as the inlet velocity difference sequence, and use the sequence obtained after the first-order difference of the outlet velocity sequence as the outlet velocity difference sequence, then multiply the kurtosis of the inlet velocity sequence and the maximum value of the inlet velocity difference sequence, and use the multiplied value as the inlet fluctuation characteristic value, and multiply the kurtosis of the outlet velocity difference sequence and the maximum value of the outlet velocity difference sequence, and use the multiplied value as the outlet fluctuation characteristic value, and use the binary group consisting of the inlet fluctuation characteristic value and the outlet fluctuation characteristic value as the flow velocity fluctuation group of the coolant at the selected ambient temperature, and continue to determine the flow velocity fluctuation groups at the remaining ambient temperatures. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0027] It should be noted that each flow rate fluctuation group in the present application corresponds to an ambient temperature.

[0028] In specific implementation, the turbulence fluctuation characteristics of the coolant at different ambient temperatures are determined according to the inlet dynamic flow velocity interval, the outlet dynamic flow velocity interval and all flow velocity fluctuation groups, which can be achieved in the following manner, namely: selecting a flow velocity fluctuation group as the selected flow velocity fluctuation group, multiplying the inlet fluctuation characteristic value in the selected flow velocity fluctuation group by the difference between the maximum value and the minimum value in the inlet dynamic flow velocity interval, and using the multiplied value as the inlet turbulence characteristic, then multiplying the outlet fluctuation characteristic value in the selected flow velocity fluctuation group by the difference between the maximum value and the minimum value in the outlet dynamic flow velocity interval, and using the multiplied value as the outlet turbulence characteristic, thereby dividing the inlet turbulence characteristic by the outlet turbulence characteristic, and using the divided value as the turbulence fluctuation characteristic of the coolant at the ambient temperature corresponding to the selected flow velocity fluctuation group, and continuing to determine the turbulence fluctuation characteristics of the coolant at the ambient temperatures corresponding to the remaining flow velocity fluctuation groups. Other methods can also be used in other embodiments, which will not be repeated here.

[0029] It should be noted that the turbulence fluctuation characteristics described in the present application represent the behavioral characteristics of flow disturbances caused by factors such as flow instability, changes in flow channel shape, and differences in the thermal load of the central processing unit during the flow of the coolant. The turbulence fluctuation characteristics can help adjust the flow path and flow distribution of the coolant in a multi-core server in real time to ensure that high heat load areas are adequately cooled while avoiding overcooling of low heat load areas.

[0030] In some embodiments, reference Figure 2 As shown, this figure is a schematic diagram of the process of determining the cooling energy loss in some embodiments of the present application. In this embodiment, the cooling energy loss of the coolant at different ambient temperatures can be determined according to all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process. The following steps can be used to achieve: First, in step 1021, linear fitting is performed on all spoiler fluctuation characteristics to obtain spoiler fluctuation fitting curves; Secondly, in step 1022, the condensation heat exchange degree of the coolant at different ambient temperatures is determined according to the turbulence fluctuation fitting curve; Then, in step 1023, the cooling energy loss of the coolant at different ambient temperatures is determined through all condensation heat exchange degrees and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process.

[0031] In the specific implementation, linear fitting is performed on all the spoiler fluctuation characteristics to obtain the spoiler fluctuation fitting curve, which can be achieved in the following manner, namely: first, all the spoiler fluctuation characteristics are sorted in ascending order, and the sequence obtained after sorting is used as the spoiler fluctuation feature sequence; then, the spoiler fluctuation feature sequence is fit using the least squares support vector machine algorithm in the prior art, and the fitted curve is used as the spoiler fluctuation fitting curve. Other fitting algorithms can also be used in other embodiments, which are not limited here.

[0032] It should be noted that in the present application, each value of the spoiler fluctuation fitting curve is used as a spoiler fluctuation fitting value, and each spoiler fluctuation feature corresponds to a spoiler fluctuation fitting value, wherein each spoiler fluctuation feature corresponds to an ambient temperature, that is, each spoiler fluctuation fitting value corresponds to an ambient temperature.

[0033] In specific implementation, the condensation heat exchange degree of the coolant at different ambient temperatures can be determined according to the turbulence fluctuation fitting curve in the following manner, namely: taking the derivative of the position corresponding to each turbulence fluctuation fitting value on the turbulence fluctuation fitting curve, and then multiplying all the obtained derivatives by the corresponding turbulence fluctuation fitting value, and taking all the multiplied values ​​as the condensation heat exchange degree, wherein each condensation heat exchange degree corresponds to an ambient temperature, thereby obtaining the condensation heat exchange degree of the coolant at different ambient temperatures. In other embodiments, other methods can also be used for implementation, which are not limited here.

[0034] It should be noted that the condensation heat exchange degree described in this application represents a parameter value that measures the ability of the cooling system of the target multi-core server to remove heat during the condensation process. The greater the condensation heat exchange degree, the greater the ability of the cooling system of the target multi-core server to remove heat during the condensation process, and vice versa. In addition, each condensation heat exchange degree corresponds to an ambient temperature.

[0035] In addition, it should be noted that the temperature change trend of each central processing unit in the heat dissipation process in the present application represents a trend curve of the temperature change of the central processing unit over time under the influence of the cooling system during the heat dissipation process. During the heat dissipation process, the temperature change trend can reflect the temperature rise, fall and stable state change process of the central processing unit from startup to stable operation to load change or cooling system adjustment. The temperature data can be obtained through a temperature sensor near the central processing unit, where the temperature value is collected once every one second, and then a temperature change trend curve is generated using a data analysis tool (such as Excel, MATLAB, Python, etc.).

[0036] In specific implementation, the following method is used to determine the cold energy loss of the coolant at different ambient temperatures through all condensation heat exchange degrees and the temperature change trends of each central processor in the target multi-core server during the heat dissipation process, namely: first, select a temperature change trend as the selected temperature change trend, select multiple extreme values ​​from the selected temperature change trend, average all extreme values, and use the obtained average as the temperature transfer variable, select the maximum value and the minimum value from the selected temperature change trend, divide the difference between the maximum value and the minimum value by the temperature transfer variable, and use the obtained value as the temperature residual value of the selected temperature change trend, and continue to determine the temperature residual value of the remaining temperature change trend; secondly, sum all the temperature residual values, and use the summed value as the temperature loss amount; then, select an ambient temperature as the selected ambient temperature, multiply the condensation heat exchange degree corresponding to the selected ambient temperature by the temperature loss amount, and use the multiplied value as the cold energy loss of the coolant at the selected ambient temperature, and continue to determine the cold energy loss of the coolant at the remaining ambient temperature; other methods can also be used in other embodiments, which will not be repeated here.

[0037] It should be noted that the cooling energy loss described in the present application refers to the energy consumed by the coolant in the cooling system when passing through the central processing unit, which fails to effectively reduce the energy lost by the heat load of the central processing unit. By reducing the cooling energy loss, the overall performance of the cooling system can be improved to ensure that the coolant can effectively take away the heat generated by the processor.

[0038] In step 103, the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block is determined, and the cooling load of the target multi-core server during the heat dissipation process is determined based on the differential entropy and all cooling energy losses.

[0039] In some embodiments, determining the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block may be achieved by the following steps: Acquire an inlet velocity sequence and an outlet velocity sequence in the flow information of the coolant of the cooling block at different ambient temperatures, and then obtain a plurality of inlet velocity sequences and a plurality of outlet velocity sequences; Selecting an import speed sequence as a selected import speed sequence; Determine the flow velocity gradient of the selected inlet velocity sequence at the corresponding ambient temperature according to the selected inlet velocity sequence and all outlet velocity sequences; Continue to determine the velocity gradients of the remaining inlet velocity sequences at the corresponding ambient temperature; The differential entropy of the liquid flow rate between the cooling liquid inlet and the cooling liquid outlet of the cooling block is determined based on all flow rate gradients.

[0040] In specific implementation, the velocity gradient at the corresponding ambient temperature of the selected inlet velocity sequence can be determined according to the selected inlet velocity sequence and all the outlet velocity sequences. The following method can be used for implementation, namely: first, the ambient temperature of the selected inlet velocity sequence is used as the dependent variable, and then all the outlet velocity sequences are used as the independent variables, and a regression model (such as polynomial regression, ridge regression, etc., which are not limited here) is used for training to obtain a regression model of the training process, and then the mean of multiple regression coefficients of the regression model is calculated, and then the reciprocal of the exponential function with the natural logarithm as the base is taken for the obtained mean, and the value obtained by taking the reciprocal of the exponential function with the natural logarithm as the base is taken as the temperature correction coefficient. number; secondly, select the outlet velocity sequence with the largest variance as the selected outlet velocity sequence, and subtract each outlet velocity in the selected outlet velocity sequence from the corresponding inlet velocity in the selected inlet velocity sequence. For example: outlet velocity sequence = [3, 4], inlet velocity sequence = [4, 6], then subtract 3 from 4, subtract 4 from 6, and then divide all the subtracted values ​​by the length of the flow channel in the cooling block, further calculate the average of all the divided values, and then multiply the average by the temperature correction coefficient, and use the multiplied value as the flow velocity gradient at the corresponding ambient temperature of the selected inlet velocity sequence. In other embodiments, other methods can also be used for implementation, which is not limited here.

[0041] It should be noted that the flow velocity gradient described in the present application measures the degree of change in the coolant flow velocity in space. The flow velocity gradient can show the difference in flow velocity of the coolant at different positions and help identify whether the flow is uniform.

[0042] In specific implementation, determining the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet of the cooling block based on all flow velocity gradients can be achieved in the following manner, namely: using the information entropy formula to calculate all flow velocity gradients, and using the calculated value as the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet of the cooling block. In other embodiments, other methods can also be used to achieve this, which will not be repeated here.

[0043] It should be noted that the differential entropy described in the present application represents the distribution uncertainty of the flow velocity gradient of the coolant between the inlet and outlet of the cooling block. Through the differential entropy, the uneven cooling effect caused by thermal asymmetry can be identified, thereby helping to optimize the coolant flow of the cooling system to improve the cooling efficiency.

[0044] In some embodiments, determining the cooling load of the target multi-core server during the heat dissipation process according to the differential entropy and all cooling energy losses can be achieved by using the following steps: Determine the global cooling loss of the target multi-core server during the cooling process according to the preset sliding window and all cooling energy losses; The cooling load of the target multi-core server during the heat dissipation process is determined according to the difference entropy and the global cooling loss.

[0045] It should be noted that the sliding window described in the present application represents a sequence for processing all cooling energy losses, and can be set using a fixed-length sequence, wherein the fixed length can be set using a value obtained by taking the square root of the total number of all cooling energy losses and rounding it down.

[0046] In specific implementation, determining the global cooling loss amount of the target multi-core server during the heat dissipation process according to the preset sliding window and all cooling energy losses can be implemented in the following manner, namely: sorting all cooling energy losses in ascending order, and taking the sequence obtained after sorting as the cooling energy loss sequence, then aligning the sliding window with the first cooling energy loss in the cooling energy loss sequence, selecting the maximum cooling energy loss and the minimum cooling energy loss from all cooling energy losses covered by the sliding window, and taking the sum of the maximum cooling energy loss and the minimum cooling energy loss as the local cooling loss amount, moving the sliding window, aligning the sliding window with the second cooling energy loss in the cooling energy loss sequence, selecting the maximum cooling energy loss and the minimum cooling energy loss from all cooling energy losses covered by the sliding window, and taking the sum of the maximum cooling energy loss and the minimum cooling energy loss as the local cooling loss amount, repeating the above steps, moving the sliding window in sequence until the sliding window is aligned with the last cooling energy loss in the cooling energy loss sequence, then summing all the local cooling loss amounts, and taking the summed value as the global cooling loss amount of the target multi-core server during the heat dissipation process, and other methods can also be used in other embodiments, which are not limited here.

[0047] It should be noted that the global cooling loss amount described in the present application represents the total cooling energy loss of the entire cooling system in the target multi-core server due to the uneven distribution of heat loads of each central processing unit.

[0048] It should also be noted that the cooling load described in the present application represents the total additional heat accumulated by the liquid cooling system due to thermal asymmetry when cooling each central processing unit in a multi-core server. The cooling load helps to evaluate and optimize the ability of the cooling system to cope with thermal asymmetry. Through effective cooling control, the formation of local hot spots can be reduced to avoid excessive temperatures due to heat accumulation. As a preferred embodiment, the cooling load of the target multi-core server during the heat dissipation process is determined based on the differential entropy and the global cooling loss. The following method is used, namely: the product of the global cooling loss and the flow rate differential entropy is used as the cooling load of the target multi-core server during the heat dissipation process. Other methods can also be used in other embodiments, which will not be repeated here.

[0049] In step 104, the heat exchange information between the cooling block and each central processing unit in the target multi-core server is obtained, and then the heat exchange coefficient between the cooling block and each central processing unit is determined based on all the heat exchange information, and the heat transfer intensity of each central processing unit during the heat dissipation process is determined through all the heat exchange coefficients and the temperature gradient of each central processing unit.

[0050] In specific implementation, the temperature data between the cooling block and each central processing unit in the target multi-core server is collected by a temperature sensor, and then the heat flux density data between the cooling block and each central processing unit is collected by a heat flux meter, and then the set of temperature data and heat flux density data is used as the heat exchange information between the cooling block and each central processing unit, wherein one central processing unit corresponds to one piece of heat exchange information.

[0051] It should be noted that the temperature data described in this application represents an ordered sequence of temperature values ​​between the cooling block and the central processing unit collected at intervals of 10 seconds in the last hour, and the heat flux density data represents an ordered sequence of heat flux density values ​​between the cooling block and the central processing unit collected at intervals of 10 seconds in the last hour.

[0052] In some embodiments, determining the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information can be achieved by using the following steps: Determine the heat diffusion characteristic sequence and thermal boundary offset between the cooling block and each central processing unit according to all the heat exchange information; The heat transfer coefficient between the cooling block and each central processing unit is determined by all heat diffusion feature sequences and all thermal boundary offsets.

[0053] In specific implementation, determining the heat diffusion feature sequence and heat boundary offset between the cooling block and each central processing unit according to all the heat exchange information can be implemented in the following manner, namely: selecting a piece of heat exchange information as the selected heat exchange information, performing spatial interpolation (for example: Kriging interpolation, spline interpolation, etc., not limited here) on the temperature data and heat flux density data in the selected heat exchange information, thereby generating a continuous temperature field and heat flux density field, further generating images of the temperature field and the heat flux density field, and then obtaining a temperature field distribution map and a heat flux density field distribution map, using the Canny edge detection algorithm in the image processing technology to detect the boundaries of the temperature field distribution map and the heat flux density field distribution map, and extracting the boundaries of the temperature field distribution map by the Canny edge detection algorithm as the temperature diffusion boundary, and at the same time, extracting the boundaries of the heat flux density field distribution map by the Canny edge detection algorithm as the heat flux diffusion boundary, further, sorting all the values ​​on the temperature diffusion boundary in ascending order, and the sequence obtained after the ascending order is sorted. As a temperature diffusion boundary sequence, at the same time, all the numerical values ​​on the heat flux diffusion boundary are sorted in ascending order, and the sequence obtained after ascending sorting is used as the heat flux diffusion boundary sequence, then, each numerical value in the temperature diffusion boundary sequence and the numerical value at the corresponding position in the heat flux diffusion boundary sequence are added, and then all the added values ​​are sorted in ascending order, and the sequence obtained after sorting is used as the heat diffusion characteristic sequence between the central processing unit and the cooling block corresponding to the selected heat exchange information, wherein each value in the heat diffusion sequence is used as a heat flux diffusion characteristic value, and finally, each numerical value in the temperature diffusion boundary sequence is subtracted from the numerical value at the corresponding position in the heat flux diffusion boundary sequence, and then the average of all the subtracted values ​​is calculated, and the obtained value is used as the thermal boundary offset between the central processing unit and the cooling block corresponding to the selected heat exchange information, and the thermal diffusion characteristic sequence and thermal boundary offset between the central processing unit and the cooling block corresponding to the remaining heat exchange information are continued to be determined. In other embodiments, other methods can also be used to achieve this, which is not limited here.

[0054] It should be noted that the thermal diffusion feature sequence described in the present application is an ordered sequence composed of multiple heat flux diffusion feature values, wherein the heat flux diffusion feature value reflects the intensity of heat propagation during the thermal diffusion process. The larger the heat flux diffusion feature value, the stronger the intensity of heat propagation during the thermal diffusion process, and vice versa. In addition, it should be noted that in the present application, a thermal diffusion feature sequence and a thermal boundary offset correspond to the same central processor, that is, a thermal diffusion feature sequence corresponds to a thermal boundary offset.

[0055] In addition, it should be noted that the thermal boundary offset described in this application represents the balance of the difference between the heat distribution boundary and the temperature distribution boundary. The thermal boundary offset can reflect the uniformity of heat dissipation of the central processor in a multi-core server.

[0056] In specific implementation, the heat transfer coefficient between the cooling block and each central processing unit is determined by all heat diffusion characteristic sequences and all thermal boundary offsets, which can be implemented in the following manner, namely: a heat diffusion characteristic sequence is selected as a selected heat diffusion characteristic sequence, all heat flux diffusion characteristic values ​​in the selected heat diffusion characteristic sequence are summed, and the summed value is divided by the mean value of the selected heat diffusion characteristic sequence, and the value obtained by the division is further multiplied by the thermal boundary offset corresponding to the selected heat diffusion characteristic sequence, so as to take the square root of the multiplied value, and the value obtained after the square root is taken as the heat transfer coefficient between the cooling block and the central processing unit corresponding to the selected heat diffusion characteristic sequence, and the heat transfer coefficient between the cooling block and the central processing units corresponding to the remaining heat diffusion characteristic sequences is further determined. In other embodiments, other methods can also be used for implementation, which will not be repeated here.

[0057] It should be noted that the heat transfer coefficient described in the present application reflects the heat dissipation efficiency of the cooling block at the central processing unit. The larger the heat transfer coefficient, the higher the heat dissipation efficiency of the cooling block at the central processing unit, and vice versa. By adjusting the heat transfer coefficient, the thermal asymmetry caused by the difference in heat dissipation efficiency in different areas of the liquid cooling system can be reduced, ensuring that each part can maintain a relatively uniform temperature distribution, thereby improving the overall efficiency of the cooling system, wherein each heat transfer coefficient corresponds to a central processing unit.

[0058] In some embodiments, determining the heat transfer intensity of each CPU during the heat dissipation process through all heat transfer coefficients and the temperature gradient of each CPU can be achieved by the following steps: Determine the thermal attenuation of the target multi-core server during the heat dissipation process based on all heat transfer coefficients; The heat transfer intensity of each CPU during the heat dissipation process is determined by the heat attenuation degree and the temperature gradient of each CPU.

[0059] In specific implementation, the thermal attenuation degree of the target multi-core server during the heat dissipation process can be determined based on all the heat transfer coefficients in the following manner, namely: multiplying each heat transfer coefficient by the temperature value between the central processing unit and the cooling block corresponding to each heat transfer coefficient, then summing all the multiplied values, further dividing the summed value by the total number of heat transfer coefficients, and using the divided value as the thermal attenuation degree of the target multi-core server during the heat dissipation process. Other methods may also be used in other embodiments, which are not limited here.

[0060] It should be noted that the thermal attenuation described in this application represents the decay rate at which heat gradually decays from the central processing unit in a multi-core server to a stable state during the heat dissipation process. A good thermal attenuation helps to keep the device within a stable operating temperature range, thereby improving the stability and performance of the system and reducing the risk of overheating due to uneven temperature.

[0061] It should also be noted that in the present application, the temperature gradient of the central processing unit can obtain temperature data through a temperature sensor near the central processing unit, and then use the finite difference method to calculate the temperature gradient of each temperature value, and then select the largest temperature gradient from all temperature gradients, and use the largest temperature gradient as the temperature gradient of the central processing unit.

[0062] In addition, it should be noted that the heat transfer intensity described in the present application represents an indicator for measuring the heat transfer intensity of the central processing unit during the heat dissipation process. The heat transfer intensity can identify the area where the central processing unit is located where the heat transfer efficiency is low and the temperature is uneven under the thermal asymmetry phenomenon. As a preferred embodiment, the heat transfer intensity of each central processing unit during the heat dissipation process is determined by the thermal attenuation degree and the temperature gradient of each central processing unit. The following method can be used to achieve this, namely: dividing the thermal attenuation degree by the temperature gradient of each central processing unit, and using the value obtained by the division as the heat transfer intensity of the corresponding central processing unit during the heat dissipation process. In other embodiments, other methods can also be used to achieve this, which will not be repeated here.

[0063] In step 105, the cooling cost of each central processing unit in the heat dissipation process is determined according to all the heat transfer intensities and the cooling load, and the cooling liquid in the target multi-core server is dynamically cooled and regulated based on all the cooling costs.

[0064] In some embodiments, reference Figure 3 As shown, this figure is a schematic diagram of the process of determining the cooling cost in some embodiments of the present application. In this embodiment, the cooling cost of each central processor in the heat dissipation process is determined according to all heat transfer intensities and the cooling load, which can be achieved by the following steps: First, in step 1051, all heat transfer intensities are converted into a heat transfer intensity sequence; Next, in step 1052, the heat dissipation cost of each CPU is determined according to the heat transfer intensity sequence and the cooling load; Then, in step 1053, the cooling energy loss of the coolant at different ambient temperatures is obtained; Finally, in step 1054, the cooling cost of each CPU during the heat dissipation process is determined through all heat dissipation costs and the cooling energy loss of the coolant at different ambient temperatures.

[0065] In specific implementation, converting all heat transfer intensities into a heat transfer intensity sequence can be achieved in the following manner, namely: sorting all heat transfer intensities in ascending order, and using the sorted sequence as the heat transfer intensity sequence. Other methods can also be used in other embodiments, which will not be described in detail here.

[0066] In specific implementation, the heat dissipation cost of each central processing unit is determined according to the heat transfer intensity sequence and the cooling load. The following method is used, namely: first, the heat transfer intensity sequence is curve-fitted using Gaussian kernel regression, and the fitted curve is used as the heat transfer intensity curve. At the same time, each value on the heat transfer intensity curve is used as the heat transfer intensity fitting value, wherein each heat transfer intensity corresponds to a heat transfer intensity fitting value on the heat transfer intensity curve. In addition, one heat transfer intensity corresponds to one central processing unit, namely, each central processing unit corresponds to one heat transfer intensity fitting value, and an ambient temperature is selected as the selected ambient temperature. Then, the central processing unit with the largest difference between the mean value of the temperature data and the selected ambient temperature is selected from all the central processing units, and the central processing unit is used as the target central processing unit. The heat transfer intensity fitting value corresponding to the target central processing unit is divided by the cooling load, and the value obtained by the division is used as the heat dissipation cost of the target central processing unit under the selected ambient temperature. The central processing units under the remaining ambient temperatures are continuously processed to obtain the heat dissipation cost of each central processing unit. Other methods may also be used in other embodiments, which will not be described in detail here.

[0067] It should be noted that the heat dissipation cost mentioned in this application refers to the energy and resources required to achieve effective heat dissipation in a multi-core server under different ambient temperatures, including the energy consumption of the coolant, the flow regulation overhead, and the additional cost required to overcome the thermal asymmetry of the system during the heat dissipation process. The cooling cost can be used to identify thermally asymmetric areas that require additional attention, thereby reducing local overheating through technical means.

[0068] In addition, it should be noted that the cooling cost described in the present application refers to the cost required to achieve the best cooling effect for each central processing unit in a multi-core server at different ambient temperatures. The cooling cost can be used to evaluate the cooling requirements of each central processing unit, and the flow and distribution of the coolant can be dynamically adjusted, thereby effectively reducing the impact of thermal asymmetry on the overall heat dissipation effect. As a preferred embodiment, the cooling cost of each central processing unit in the heat dissipation process is determined by all heat dissipation costs and the cold energy loss of the coolant at different ambient temperatures. It can be achieved in the following way, namely: first, the maximum ambient temperature and the minimum ambient temperature are selected from all ambient temperatures, the cold energy loss at the maximum ambient temperature is added to the cold energy loss at the minimum ambient temperature, and the added value is used as the ambient temperature cold energy loss, and then, the heat dissipation cost of each central processing unit is added to the ambient temperature cold energy loss, and the added value is used as the cooling cost of each central processing unit in the heat dissipation process, wherein each cooling cost corresponds to a central processing unit, and other methods can also be used in other embodiments, which are not limited here.

[0069] In some embodiments, dynamically adjusting and controlling the cooling of the coolant in the target multi-core server based on all cooling costs may be implemented by the following steps: Determine a cooling cost threshold based on all cooling costs; Select a cooling cost as the selected cooling cost; Determining a coolant dynamic adjustment coefficient according to a selected cooling cost and the cooling cost threshold; The coolant flow control valve of the tortuous channel on the central processing unit corresponding to the selected cooling cost is adjusted by the coolant dynamic adjustment coefficient; Continue to adjust the flow of the coolant flow control valve of the tortuous channel on the central processing unit corresponding to the remaining cooling cost.

[0070] In specific implementation, the cooling cost threshold value can be determined based on all cooling costs in the following manner, namely: calculating the average of all cooling costs, and using the average as the cooling cost threshold value; the dynamic adjustment coefficient of the coolant can be determined based on the selected cooling cost and the cooling cost threshold value in the following manner, namely: subtracting the selected cooling cost from the cooling cost threshold value, and using the value obtained by the difference as the dynamic adjustment coefficient of the coolant.

[0071] In specific implementation, the flow regulation of the coolant flow control valve of the tortuous channel on the central processing unit corresponding to the selected cooling cost through the coolant dynamic regulation coefficient can be achieved in the following manner, namely: if the coolant dynamic regulation coefficient is greater than 0, the coolant flow control valve of the tortuous channel on the central processing unit corresponding to the selected cooling cost is used to increase the coolant flow by the same multiple of the value of the coolant dynamic regulation coefficient; if the coolant dynamic regulation coefficient is less than 0, the coolant flow control valve of the tortuous channel on the central processing unit corresponding to the selected cooling cost is used to reduce the coolant flow by the same multiple of the value of the coolant dynamic regulation coefficient; if the coolant dynamic regulation coefficient is equal to 0, no operation is performed, for example: when the coolant dynamic regulation coefficient is equal to 0.4, the coolant flow of the coolant flow control valve of the tortuous channel on the central processing unit corresponding to the selected cooling cost is increased to 0.4 times the original.

[0072] In addition, in another aspect of the present application, in some embodiments, the present application provides a multi-core server heat dissipation control system, referring to Figure 4 , which is a schematic diagram of exemplary hardware and / or software of a multi-core server heat dissipation control system according to some embodiments of the present application, the multi-core server heat dissipation control system 400 includes: a collection module 401, a processing module 402 and an execution module 403, which are described as follows: The acquisition module 401 in this application is mainly used to collect the flow information of the coolant of the cooling block in the target multi-core server under different ambient temperatures; Processing module 402, in the present application, is used to extract the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all flow information, and determine the cooling energy loss of the coolant at different ambient temperatures according to all turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process; It should be noted that the processing module 402 in the present application is also used to determine the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block, and determine the cooling load of the target multi-core server during the heat dissipation process according to the differential entropy and all cooling energy losses; In addition, the processing module 402 in the present application is also used to obtain heat exchange information between the cooling block and each central processing unit in the target multi-core server, and then determine the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information, and determine the heat transfer intensity of each central processing unit in the heat dissipation process through all the heat exchange coefficients and the temperature gradient of each central processing unit; Execution module 403, in this application, execution module 403 is mainly used to determine the cooling cost of each central processing unit during the heat dissipation process according to all heat transfer intensities and the cooling load, and dynamically adjust the cooling of the coolant in the target multi-core server based on all cooling costs.

[0073] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned multi-core server heat dissipation control method.

[0074] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device using a multi-core server heat dissipation control method according to some embodiments of the present application. The multi-core server heat dissipation control method in the above embodiment can be Figure 5 The computer device 500 shown in the figure is implemented, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503 and at least one communication interface 504.

[0075] The processor 501 may be a general-purpose central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more processors for controlling the execution of the multi-core server heat dissipation control method in the present application.

[0076] The communication bus 502 may be used to transmit information between the above-mentioned components.

[0077] The memory 503 may be a read only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read only memory (EEPROM), a compact disc read only memory (CD ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 503 may exist independently and be connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0078] The memory 503 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 501. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The method described in the above method embodiment can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0079] The communication interface 504 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0080] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-core (single CPU) processor or a multi-core (multi CPU) processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0081] The above-mentioned computer device may be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0082] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the multi-core server heat dissipation control method described above is implemented.

[0083] In summary, in the multi-core server heat dissipation control method, system, device and storage medium disclosed in the embodiments of the present application, first, the flow information of the coolant of the cooling block in the target multi-core server at different ambient temperatures is collected; the turbulence fluctuation characteristics of the coolant at different ambient temperatures are extracted from all the flow information, and the cold energy loss of the coolant at different ambient temperatures is determined according to all the turbulence fluctuation characteristics and the temperature change trend of each central processing unit in the target multi-core server during the heat dissipation process; the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block is determined, and the cold load of the target multi-core server during the heat dissipation process is determined according to the differential entropy and all the cold energy losses; the heat exchange information between the cooling block and each central processing unit in the target multi-core server is obtained, and then the heat exchange coefficient between the cooling block and each central processing unit is determined according to all the heat exchange information, and the heat transfer intensity of each central processing unit during the heat dissipation process is determined by all the heat transfer coefficients and the temperature gradient of each central processing unit; the cooling cost of each central processing unit during the heat dissipation process is determined according to all the heat transfer intensities and the cold load, and the coolant in the target multi-core server is dynamically cooled and regulated based on all the cooling costs. The solution of the present application can reduce the impact of thermal asymmetry on the heat dissipation effect of a liquid-cooled multi-core server, thereby improving the efficiency of the cooling system.

[0084] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A multi-core server heat dissipation control method, characterized in that: The steps include: Collect the flow information of the coolant of the cooling block in the target multi-core server under different ambient temperatures; Extracting the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all the flow information, and determining the cooling energy loss of the coolant at different ambient temperatures according to all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process; Determine the differential entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block, and determine the cooling load of the target multi-core server during the heat dissipation process according to the differential entropy and all cooling energy losses; Acquire heat exchange information between the cooling block and each central processing unit in the target multi-core server, and then determine the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information, and determine the heat transfer intensity of each central processing unit in the heat dissipation process through all the heat exchange coefficients and the temperature gradient of each central processing unit; The cooling cost of each central processing unit in the heat dissipation process is determined according to all heat transfer intensities and the cooling load, and the cooling liquid in the target multi-core server is dynamically cooled and regulated based on all cooling costs.

2. The method according to claim 1, characterized in that Extracting the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all the flow information specifically includes: Determine an inlet dynamic flow rate interval and an outlet dynamic flow rate interval of the coolant in the target multi-core server according to all flow information; Performing differential processing on each flow information to obtain a flow velocity fluctuation group of the coolant at different ambient temperatures; The turbulent flow fluctuation characteristics of the coolant at different ambient temperatures are determined according to the inlet dynamic flow velocity interval, the outlet dynamic flow velocity interval and all flow velocity fluctuation groups.

3. The method according to claim 1, characterized in that According to all the disturbance fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process, the cooling energy loss of the coolant at different ambient temperatures is determined, specifically including: Perform linear fitting on all the spoiler fluctuation characteristics to obtain spoiler fluctuation fitting curve; Determining the condensation heat exchange degree of the coolant at different ambient temperatures according to the turbulence fluctuation fitting curve; The cooling energy loss of the coolant at different ambient temperatures is determined through all condensation heat exchange degrees and the temperature change trends of each central processing unit in the target multi-core server during the heat dissipation process.

4. The method according to claim 1, characterized in that Determining the cooling load of the target multi-core server during the heat dissipation process according to the differential entropy and all cooling energy losses specifically includes: Determine the global cooling loss of the target multi-core server during the cooling process according to the preset sliding window and all cooling energy losses; The cooling load of the target multi-core server during the heat dissipation process is determined according to the difference entropy and the global cooling loss.

5. The method according to claim 1, characterized in that Determining the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information specifically includes: Determine the heat diffusion characteristic sequence and thermal boundary offset between the cooling block and each central processing unit according to all the heat exchange information; The heat transfer coefficient between the cooling block and each central processing unit is determined by all heat diffusion feature sequences and all thermal boundary offsets.

6. The method according to claim 1, characterized in that The heat transfer intensity of each CPU during the heat dissipation process is determined by all heat transfer coefficients and the temperature gradient of each CPU, including: Determine the thermal attenuation of the target multi-core server during the heat dissipation process based on all heat transfer coefficients; The heat transfer intensity of each CPU during the heat dissipation process is determined by the heat attenuation degree and the temperature gradient of each CPU.

7. The method according to claim 1, characterized in that The cooling cost of each CPU in the heat dissipation process is determined according to all heat transfer intensities and the cooling load, and specifically includes: Convert all heat transfer intensities into a heat transfer intensity sequence; Determining the heat dissipation cost of each central processing unit according to the heat transfer intensity sequence and the cooling load; Obtain the cooling energy loss of the coolant at different ambient temperatures; The cooling cost of each CPU during the heat dissipation process is determined through all heat dissipation costs and the cooling energy loss of the coolant at different ambient temperatures.

8. A multi-core server heat dissipation control system, characterized in that: include: A collection module, used to collect flow information of coolant of a cooling block in a target multi-core server at different ambient temperatures; A processing module, used to extract the turbulence fluctuation characteristics of the coolant at different ambient temperatures from all the flow information, and determine the cooling energy loss of the coolant at different ambient temperatures according to all the turbulence fluctuation characteristics and the temperature change trend of each central processor in the target multi-core server during the heat dissipation process; The processing module is further used to determine the difference entropy of the liquid flow rate between the coolant inlet and the coolant outlet in the cooling block, and determine the cooling load of the target multi-core server during the heat dissipation process according to the difference entropy and all cooling energy losses; The processing module is further used to obtain heat exchange information between the cooling block and each central processing unit in the target multi-core server, and then determine the heat exchange coefficient between the cooling block and each central processing unit according to all the heat exchange information, and determine the heat transfer intensity of each central processing unit in the heat dissipation process through all the heat exchange coefficients and the temperature gradient of each central processing unit; The execution module is used to determine the cooling cost of each central processing unit during the heat dissipation process according to all heat transfer intensities and the cooling load, and dynamically cool the coolant in the target multi-core server based on all cooling costs.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the multi-core server heat dissipation control method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the multi-core server heat dissipation control method according to any one of claims 1 to 7 is implemented.

Images