Liquid cooling system and method for computer chip

Abstract

The invention relates to the technical field of chip heat dissipation, in particular to a computer chip liquid cooling heat dissipation system and method. The heat dissipation control module is used for outputting a control instruction according to the temperature difference value, constructing a chip temperature prediction model, obtaining a chip temperature prediction value, adjusting a judgment result of a heat dissipation mode according to the chip temperature prediction value, and correcting a safe working temperature value according to the use duration of the chip; and the execution feedback module is used for controlling the cooling liquid adjusting device according to the control quantity, outputting alarm information according to the cooling liquid flow, constructing a preset flow dynamic adjusting model and obtaining preset flow automatic adjusting information according to the real-time environment temperature data and the real-time computer load data. According to the invention, the potential overheating risk can be found in time, and the stability and reliability of the system are improved.

Description

Technical Field

[0001] The present invention relates to the field of chip heat dissipation technology, and in particular to a liquid cooling and heat dissipation system and method for a computer chip. Background Art

[0002] Computer chips generate a lot of heat during operation. If the heat cannot be dissipated in time, the chip temperature will rise, which will affect its performance and life. Although the liquid cooling system can effectively reduce the chip temperature, it will be difficult to detect temperature abnormalities in time and take corresponding measures if the chip temperature is not predicted and monitored in real time. Chip temperature prediction can help operation and maintenance personnel understand the temperature change trend of the chip in advance, so as to take corresponding heat dissipation measures to avoid performance degradation or damage caused by chip overheating. Through temperature prediction, the operation strategy of the cooling system can also be optimized to improve heat dissipation efficiency and reduce energy consumption.

[0003] Chinese Patent Publication No.: CN104216492A discloses a computer chip liquid cooling immersion structure heat dissipation system, including a coolant pump, a coolant tank, a liquid cooling radiator, and a liquid pipe. The coolant pump, the coolant tank and the liquid cooling radiator are connected in sequence through the liquid pipe to form a loop; insulating coolant is used in the loop, and the cooling object is placed in the coolant tank; it can be seen that the computer chip liquid cooling immersion structure heat dissipation system has the following problems: lack of abnormal operating status alarm information, lack of chip temperature prediction process. Summary of the invention

[0004] To this end, the present invention provides a computer chip liquid cooling and heat dissipation system and method to overcome the problems in the prior art of lacking abnormal operating status alarm information and lacking a chip temperature prediction process.

[0005] To achieve the above objectives, on the one hand, the present invention provides a computer chip liquid cooling and heat dissipation system, comprising:

[0006] A data acquisition module is used to collect chip temperature data, ambient temperature data and computer load data;

[0007] The heat dissipation control module is used to judge the heat dissipation mode according to the temperature difference value, and output the control instruction according to the judgment result, and is also used to build a chip temperature prediction model to obtain the chip temperature prediction value, and is also used to judge whether the chip is within the safe operating temperature range according to the chip temperature prediction value, and adjust the judgment result of the heat dissipation mode according to the judgment result, and is also used to correct the safe operating temperature value according to the chip usage time;

[0008] The execution feedback module is used to transmit the control instructions output by the heat dissipation control module to the coolant regulating device, control the coolant regulating device according to the control amount, judge the operating status of the coolant regulating device according to the coolant flow rate, and output alarm information according to the judgment result. It is also used to build a preset flow dynamic adjustment model to obtain the preset flow automatic adjustment information according to the real-time ambient temperature data and the real-time computer load data.

[0009] Furthermore, the heat dissipation control module includes:

[0010] A heat dissipation mode matching unit, used to compare the temperature difference value T with a preset minimum temperature difference value T0min and a preset maximum temperature difference value T0max, judge the heat dissipation mode according to the comparison result, and output a control instruction according to the judgment result;

[0011] The chip temperature prediction unit is used to construct a chip temperature prediction model, input computer load data into the chip temperature prediction model, obtain a chip temperature prediction value, compare the chip temperature prediction value A with the safe operating temperature value A0, judge whether the chip is within the safe operating temperature range according to the comparison result, and adjust the judgment result of the heat dissipation mode according to the judgment result, and compare the chip usage time D with the preset chip usage time D0, judge the practicality of the chip according to the comparison result, and correct the safe operating temperature value A0 according to the judgment result.

[0012] Further, the heat dissipation mode matching unit obtains the temperature difference T between the chip temperature and the ambient temperature, sets T=T1-T2, where T1 is the chip temperature and T2 is the ambient temperature, compares the temperature difference T with the preset minimum temperature difference T0min and the preset maximum temperature difference T0max, judges the heat dissipation mode according to the comparison result, and outputs the control instruction according to the judgment result, wherein:

[0013] When T<T0min, the heat dissipation mode matching unit determines that the heat dissipation mode is a low power consumption heat dissipation mode and outputs a low power consumption control instruction;

[0014] When T0min≤T≤T0max, the heat dissipation mode matching unit determines that the heat dissipation mode is a medium power dissipation mode and outputs a medium power dissipation control instruction;

[0015] When T>T0max, the heat dissipation mode matching unit determines that the heat dissipation mode is a high power consumption heat dissipation mode, and outputs a high power consumption control instruction.

[0016] Furthermore, when predicting the chip temperature, the chip temperature prediction unit constructs a chip temperature prediction model, divides the preset chip temperature database into 80% of the chip temperature training set, 10% of the chip temperature verification set and 10% of the chip temperature test set, inputs the chip temperature training set into the recurrent neural network model for training to obtain the trained recurrent neural network model, inputs the chip temperature verification set into the trained recurrent neural network model to iteratively optimize the parameters of the trained recurrent neural network model, inputs the chip temperature test set into the iteratively optimized recurrent neural network model, tests the iteratively optimized convolutional network neural model to obtain the prediction accuracy R, compares the prediction accuracy R with the preset prediction accuracy R0, judges the training compliance of the recurrent neural network model according to the comparison result, and outputs the recurrent neural network model according to the judgment result, wherein:

[0017] When R≥R0, it is determined that the recurrent neural network model training has reached the standard, and the recurrent neural network model is output as a chip temperature prediction model;

[0018] When R<R0, it is determined that the recurrent neural network model training does not meet the standard, the recurrent neural network model parameters are iteratively optimized, and the training is repeated until the recurrent neural network model training meets the standard;

[0019] The computer load data is input into the chip temperature prediction model to obtain the chip temperature prediction value.

[0020] Further, the chip temperature prediction unit compares the chip temperature prediction value A with the safe operating temperature value A0, judges whether the chip is within the safe operating temperature range according to the comparison result, and adjusts the judgment result of the heat dissipation mode according to the judgment result, wherein:

[0021] When A≥A0, the chip temperature prediction unit determines that the chip is not within the safe operating temperature range, and adjusts the judgment result of the heat dissipation mode, wherein:

[0022] If the cooling mode is a low power cooling mode, the adjusted cooling mode is a medium power cooling mode;

[0023] If the cooling mode is a medium power cooling mode, the adjusted cooling mode is a high power cooling mode;

[0024] If the heat dissipation mode is a high power consumption heat dissipation mode, the adjusted heat dissipation mode is a high power consumption heat dissipation mode;

[0025] When A<A0, the chip temperature prediction unit determines that the chip is within a safe operating temperature range and does not adjust the determination result of the heat dissipation mode.

[0026] Furthermore, the chip temperature prediction unit compares the chip usage time D with the preset chip usage time D0, judges the practicality of the chip according to the comparison result, and corrects the safe operating temperature value A0 according to the judgment result, wherein:

[0027] When D≥D0, the chip temperature prediction unit determines that the practicality of the chip is weak, sets a correction coefficient a, and corrects the safe operating temperature value A0 according to the correction coefficient. The corrected safe operating temperature value is Aa, and Aa=A0×a is set;

[0028] When D<D0, the chip temperature prediction unit determines that the chip has high practicality and corrects the safe operating temperature value A0.

[0029] Furthermore, the execution feedback module obtains the control instruction output by the heat dissipation control module, transmits the control instruction to the coolant regulating device, analyzes the coolant flow g, and sets Among them, A is the cooling pump speed, a is the efficiency factor, S is the cross-sectional area of ​​the liquid cooling system, b is the correction coefficient, and i is the flow rate of the coolant. is the coolant heat exchange efficiency coefficient, c is the specific heat capacity of the coolant, F is the flow resistance of the liquid cooling and heat dissipation system, d is the influence coefficient of the coolant temperature on the flow rate, the execution feedback module calculates the coolant flow difference e(t) according to the coolant flow g and the real-time coolant flow g1, and sets e(t)=|g-g1|, calculates the control quantity u(t) according to the coolant flow difference e(t) and PID parameters, and sets u(t)=Kp×e(t)+Ki×∫e(t)dt+Kd×[de(t) / dt], the PID parameters include the proportional coefficient Kp, the integral coefficient Ki and the differential coefficient Kd, and the execution feedback module controls the coolant regulating device according to the control quantity u(t).

[0030] Further, the execution feedback module obtains the coolant flow rate, compares the coolant flow rate G with the preset minimum coolant flow rate G0min and the preset maximum coolant flow rate G0max, judges the operating state of the coolant regulating device according to the comparison result, and outputs an alarm message according to the judgment result, wherein:

[0031] When G<G0min, the execution feedback module determines that the operation state of the coolant regulating device is abnormal, and outputs device blockage alarm information;

[0032] When G0min≤G≤G0max, the execution feedback module determines that the operating state of the coolant regulating device is normal and does not output an alarm message;

[0033] When G>G0max, the execution feedback module determines that the operation state of the coolant regulating device is abnormal, and outputs device leakage alarm information.

[0034] Furthermore, when the execution feedback module dynamically adjusts the preset minimum coolant flow G0min and the preset maximum coolant flow G0max, a preset flow dynamic adjustment model is constructed, the preset flow database is divided into 70% flow training set, 15% flow verification set and 15% flow test set, the flow training set is input into the random forest model for training to obtain the trained random forest model, the flow verification set is input into the trained random forest model to iteratively optimize the parameters of the trained random forest model, and the flow test set is input into the iteratively optimized random forest model to test the iteratively optimized recurrent network neural model to obtain the dynamic adjustment accuracy U, the dynamic adjustment accuracy U is compared with the preset dynamic adjustment accuracy U0, the training compliance of the random forest model is judged according to the comparison result, and the random forest model is output according to the judgment result, wherein:

[0035] When U≥U0, the random forest model training is determined to have met the standard, and the random forest model is output as the preset traffic automatic adjustment model;

[0036] When U<U0, it is determined that the random forest model training does not meet the standard, the random forest model parameters are iteratively optimized, and the training is repeated until the random forest model training meets the standard;

[0037] The real-time ambient temperature data and the real-time computer load data are input into the preset flow automatic adjustment model to obtain the preset flow automatic adjustment information.

[0038] In another aspect, the present invention provides a liquid cooling method for a computer chip, comprising:

[0039] Step S1, collecting chip temperature data, ambient temperature data and computer load data;

[0040] Step S2, judging the heat dissipation mode according to the temperature difference value, and outputting a control instruction according to the judgment result;

[0041] Step S3, constructing a chip temperature prediction model to obtain a chip temperature prediction value;

[0042] Step S4, judging whether the chip is within a safe operating temperature range according to the chip temperature prediction value, and adjusting the judgment result of the heat dissipation mode according to the judgment result;

[0043] Step S5, correcting the safe operating temperature value according to the chip usage time;

[0044] Step S6, transmitting the control instruction output by the heat dissipation control module to the coolant regulating device, and controlling the coolant regulating device according to the control amount;

[0045] Step S7, judging the operating state of the coolant regulating device according to the coolant flow rate, and outputting alarm information according to the judgment result;

[0046] Step S8, constructing a preset flow dynamic adjustment model, and obtaining preset flow automatic adjustment information according to real-time ambient temperature data and real-time computer load data.

[0047] Compared with the prior art, the beneficial effect of the present invention is that the system is applied to the chip heat dissipation process of mobile communication base station equipment, and by obtaining the predicted value of the chip temperature, it is helpful to timely discover potential overheating risks and take corresponding preventive measures to improve the stability and reliability of the system. The system ensures the timeliness and accuracy of data through the data acquisition module, realizes automatic acquisition, reduces manual intervention, and improves work efficiency. The system flexibly adapts to different working environments and load conditions through the heat dissipation control module, thereby effectively controlling the chip temperature, and obtaining the predicted value of the chip temperature in advance, which helps to timely discover potential overheating risks and take corresponding preventive measures to avoid damage to the chip due to overheating, thereby extending the service life of the equipment, ensuring that the chip can maintain the best working state at different use stages, and helping to maintain the best performance state of the system. The system ensures that the heat dissipation effect is optimal by accurately controlling the coolant regulating device through the execution feedback module. When an abnormal operating state is found, an alarm message can be immediately output to remind the operator to take measures in time to avoid overheating or damage to the equipment, and ensure that the heat dissipation system can maintain the best performance under different working conditions. Through precise control and real-time status monitoring, the stability and reliability of the heat dissipation system are improved.

[0048] In particular, the data acquisition module monitors the temperature of the chip to ensure the timeliness and accuracy of the data, helps to detect temperature anomalies in a timely manner, prevents hardware damage due to overheating, realizes automated collection, reduces manual intervention, improves work efficiency, avoids data errors caused by human operational errors, and improves system stability.

[0049] In particular, the heat dissipation control module intelligently determines the heat dissipation mode according to the temperature difference value and outputs corresponding control instructions. It can flexibly adapt to different working environments and load conditions, so as to more effectively control the chip temperature. By constructing a chip temperature prediction model, the heat dissipation control module can obtain the predicted value of the chip temperature in advance, which is helpful to timely discover potential overheating risks and take corresponding preventive measures to avoid chip damage due to overheating, thereby extending the service life of the equipment. By real-time judgment of whether the chip is within the safe operating temperature range and adjusting the judgment result of the heat dissipation mode according to the judgment result, it is ensured that the chip operates at the optimal operating temperature, thereby improving the stability and reliability of the system. The safe operating temperature value is corrected according to the chip usage time. As the chip usage time increases, its performance will gradually decline, resulting in changes in the safe operating temperature range. By adaptively adjusting the safe operating temperature value, the module can ensure that the chip can maintain the best working state at different usage stages, which helps to maintain the optimal performance state of the system.

[0050] In particular, the execution feedback module can quickly and accurately transmit the control instructions output by the heat dissipation control module to the coolant regulating device. According to different control quantities, it can realize precise control of the coolant regulating device to ensure the optimal heat dissipation effect. By real-time monitoring of the coolant flow rate and judging the operating status of the coolant regulating device according to the coolant flow rate, once an abnormal operating status is found, an alarm message can be output immediately to remind the operator to take timely measures to avoid overheating or damage to the equipment. By constructing a preset flow dynamic adjustment model, the preset flow rate of the coolant is automatically adjusted to ensure that the heat dissipation system can maintain optimal performance under different working conditions. Through precise control and real-time status monitoring, the execution feedback module can improve the stability and reliability of the heat dissipation system, help ensure that the computer can maintain a stable working state under long-term and high-load operation, and help reduce equipment operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a schematic diagram of the structure of the liquid cooling and heat dissipation system for computer chips in this embodiment;

[0052] Figure 2 This is a schematic diagram of the structure of the heat dissipation control module of this embodiment;

[0053] Figure 3 Schematic diagram of the process of the liquid cooling and heat dissipation method for computer chips in this embodiment. DETAILED DESCRIPTION

[0054] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0056] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0057] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0058] See also Figure 1 As shown, it is a schematic diagram of the structure of the liquid cooling and heat dissipation system of a computer chip in this embodiment, including:

[0059] A data acquisition module is used to collect chip temperature data, ambient temperature data and computer load data;

[0060] The heat dissipation control module is used to judge the heat dissipation mode according to the temperature difference value, and output the control instruction according to the judgment result, and is also used to build a chip temperature prediction model to obtain the chip temperature prediction value, and is also used to judge whether the chip is within the safe operating temperature range according to the chip temperature prediction value, and adjust the judgment result of the heat dissipation mode according to the judgment result, and is also used to correct the safe operating temperature value according to the chip usage time;

[0061] The execution feedback module is used to transmit the control instructions output by the heat dissipation control module to the coolant regulating device, control the coolant regulating device according to the control amount, judge the operating status of the coolant regulating device according to the coolant flow rate, and output alarm information according to the judgment result. It is also used to build a preset flow dynamic adjustment model to obtain the preset flow automatic adjustment information according to the real-time ambient temperature data and the real-time computer load data.

[0062] Specifically, the system is applied to the chip heat dissipation process of mobile communication base station equipment. By obtaining the predicted value of the chip temperature, it is helpful to timely discover potential overheating risks and take corresponding preventive measures to improve the stability and reliability of the system. The system ensures the timeliness and accuracy of the data through the data acquisition module, realizes automatic acquisition, reduces manual intervention, and improves work efficiency. The system flexibly adapts to different working environments and load conditions through the heat dissipation control module, thereby effectively controlling the chip temperature and obtaining the predicted value of the chip temperature in advance, which helps to timely discover potential overheating risks and take corresponding preventive measures to avoid damage to the chip due to overheating, thereby extending the service life of the equipment, ensuring that the chip can maintain the best working state at different stages of use, and helping to maintain the best performance state of the system. The system ensures that the heat dissipation effect is optimal by precisely controlling the coolant regulating device through the execution feedback module. If an abnormal operating state is found, an alarm message can be output immediately to remind the operator to take timely measures to avoid overheating or damage to the equipment, ensuring that the heat dissipation system can maintain the best performance under different working conditions, and improving the stability and reliability of the heat dissipation system through precise control and real-time status monitoring.

[0063] Specifically, when the data acquisition module collects chip temperature data, ambient temperature data and computer load data, the chip temperature data refers to the temperature of the core chip in the computer, such as the CPU and GPU, the ambient temperature data refers to the temperature of the environment in which the computer is located, and the computer load data refers to indicators of the workload and pressure faced by computer systems and components during operation, including CPU usage, memory occupancy, disk I / O activity and network traffic. This embodiment does not limit the method for obtaining chip temperature data, ambient temperature data and computer load data. Relevant technical personnel in this field can freely set it according to actual needs, and only need to meet the requirements for obtaining chip temperature data, ambient temperature data and computer load data. For example, the temperature sensor can be installed at the chip position to obtain chip temperature data, the temperature sensor can be installed outside the computer chassis to obtain ambient temperature data, and the computer load data can be obtained through the hardware monitoring module.

[0064] Specifically, the data acquisition module monitors the temperature of the chip to ensure the timeliness and accuracy of the data, helps to detect temperature anomalies in a timely manner, prevents hardware damage due to overheating, realizes automated collection, reduces manual intervention, improves work efficiency, avoids data errors caused by human operational errors, and improves system stability.

[0065] See also Figure 2 As shown, it is a schematic diagram of the structure of the heat dissipation control module of this embodiment, including:

[0066] A heat dissipation mode matching unit, used to compare the temperature difference value T with a preset minimum temperature difference value T0min and a preset maximum temperature difference value T0max, judge the heat dissipation mode according to the comparison result, and output a control instruction according to the judgment result;

[0067] The chip temperature prediction unit is used to construct a chip temperature prediction model, input computer load data into the chip temperature prediction model, obtain a chip temperature prediction value, compare the chip temperature prediction value A with the safe operating temperature value A0, judge whether the chip is within the safe operating temperature range according to the comparison result, and adjust the judgment result of the heat dissipation mode according to the judgment result, and compare the chip usage time D with the preset chip usage time D0, judge the practicality of the chip according to the comparison result, and correct the safe operating temperature value A0 according to the judgment result.

[0068] Specifically, the heat dissipation mode matching unit obtains the temperature difference T between the chip temperature and the ambient temperature, sets T=T1-T2, where T1 is the chip temperature and T2 is the ambient temperature, compares the temperature difference T with the preset minimum temperature difference T0min and the preset maximum temperature difference T0max, judges the heat dissipation mode according to the comparison result, and outputs the control instruction according to the judgment result, wherein:

[0069] When T<T0min, the heat dissipation mode matching unit determines that the heat dissipation mode is a low power consumption heat dissipation mode and outputs a low power consumption control instruction;

[0070] When T0min≤T≤T0max, the heat dissipation mode matching unit determines that the heat dissipation mode is a medium power dissipation mode and outputs a medium power dissipation control instruction;

[0071] When T>T0max, the heat dissipation mode matching unit determines that the heat dissipation mode is a high power consumption heat dissipation mode, and outputs a high power consumption control instruction.

[0072] Specifically, the preset minimum temperature difference refers to the minimum temperature difference value set to ensure that the coolant can effectively take away the heat generated by the chip and prevent the system from overheating and performance degradation, such as setting T0min=10°C. The preset maximum temperature difference refers to the maximum temperature difference value set to ensure that the coolant can effectively take away the heat generated by the chip and prevent the system from overheating and performance degradation, such as setting T0max=20°C. The low-power cooling mode refers to a mode suitable for scenes with low loads and small temperature changes. It reduces energy consumption and noise by reducing the power consumption of the cooling system. The medium-power cooling mode refers to a mode suitable for scenes with medium loads and moderate temperature changes. It balances the cooling effect and energy consumption. The high-power cooling mode refers to a mode suitable for scenes with high loads and large temperature changes. Scene, providing powerful heat dissipation capability to ensure stable operation of the system. This embodiment does not limit the output method of the control instruction. Relevant technical personnel in this field can freely set it according to actual needs, and only need to meet the output requirements of the control instruction. For example, the control instruction can be set to output through network communication, such as TCP / IP, UDP and MQTT. This embodiment does not limit the specific methods of low-power control instructions, medium-power control instructions and high-power control instructions. Relevant technical personnel in this field can freely set it according to actual needs, and only need to meet the heat dissipation requirements of computer chips. For example, the low-power control instruction can be set to reduce the flow rate of the coolant, the medium-power control instruction can be set to control the flow rate of the coolant, and the high-power control instruction can be set to increase the flow rate of the coolant and the flow rate of the coolant.

[0073] Specifically, when predicting the chip temperature, the chip temperature prediction unit constructs a chip temperature prediction model, divides the preset chip temperature database into 80% of the chip temperature training set, 10% of the chip temperature verification set and 10% of the chip temperature test set, inputs the chip temperature training set into the recurrent neural network model for training to obtain the trained recurrent neural network model, inputs the chip temperature verification set into the trained recurrent neural network model to iteratively optimize the parameters of the trained recurrent neural network model, inputs the chip temperature test set into the iteratively optimized recurrent neural network model, tests the iteratively optimized convolutional network neural model to obtain the prediction accuracy R, compares the prediction accuracy R with the preset prediction accuracy R0, judges the training compliance of the recurrent neural network model according to the comparison result, and outputs the recurrent neural network model according to the judgment result, wherein:

[0074] When R≥R0, it is determined that the recurrent neural network model training has reached the standard, and the recurrent neural network model is output as a chip temperature prediction model;

[0075] When R<R0, it is determined that the recurrent neural network model training does not meet the standard, the recurrent neural network model parameters are iteratively optimized, and the training is repeated until the recurrent neural network model training meets the standard;

[0076] The computer load data is input into the chip temperature prediction model to obtain the chip temperature prediction value.

[0077] Specifically, the chip temperature prediction model refers to a mathematical model that provides a basis for thermal design, heat dissipation optimization and reliability evaluation of the chip by analyzing and predicting the temperature changes of the chip during operation. The preset chip temperature database refers to a historical data repository for storing, managing and analyzing computer load data and corresponding chip temperature changes. The chip temperature training set refers to a data set used to train the recurrent neural network model. The chip temperature verification set refers to a data set used to optimize and evaluate the performance of the recurrent neural network model. The chip temperature test set refers to a data set used to test the chip temperature prediction model. The prediction accuracy R refers to the proportion of the number of samples Y correctly output by the recurrent neural network model to the total number of samples Y0 in the chip temperature test set, that is, R=Y / Y0. The preset prediction accuracy R0 refers to a preset value for judging the training compliance of the recurrent neural network model. This embodiment does not limit the value of the preset prediction accuracy R0. Relevant technicians in this field can freely set it according to actual needs, as long as the need to judge the training compliance of the recurrent neural network model is met. For example, the preset prediction accuracy R0 can be set to 0.98. The chip temperature prediction value refers to the predicted value of the temperature change of the chip over time under specific working conditions.

[0078] Specifically, the chip temperature prediction unit compares the chip temperature prediction value A with the safe operating temperature value A0, judges whether the chip is within the safe operating temperature range according to the comparison result, and adjusts the judgment result of the heat dissipation mode according to the judgment result, wherein:

[0079] When A≥A0, the chip temperature prediction unit determines that the chip is not within the safe operating temperature range, and adjusts the judgment result of the heat dissipation mode, wherein:

[0080] If the cooling mode is a low power cooling mode, the adjusted cooling mode is a medium power cooling mode;

[0081] If the cooling mode is a medium power cooling mode, the adjusted cooling mode is a high power cooling mode;

[0082] If the heat dissipation mode is a high power consumption heat dissipation mode, the adjusted heat dissipation mode is a high power consumption heat dissipation mode;

[0083] When A<A0, the chip temperature prediction unit determines that the chip is within a safe operating temperature range and does not adjust the determination result of the heat dissipation mode.

[0084] Specifically, the safe operating temperature value refers to a temperature value at which the chip will not be damaged when working for a long time, such as setting A0=30°C.

[0085] Specifically, the chip temperature prediction unit compares the chip usage time D with the preset chip usage time D0, judges the practicality of the chip according to the comparison result, and corrects the safe operating temperature value A0 according to the judgment result, wherein:

[0086] When D≥D0, the chip temperature prediction unit determines that the practicality of the chip is weak, sets a correction coefficient a, and corrects the safe operating temperature value A0 according to the correction coefficient. The corrected safe operating temperature value is Aa, and Aa=A0×a is set;

[0087] When D<D0, the chip temperature prediction unit determines that the chip has high practicality and corrects the safe operating temperature value A0.

[0088] Specifically, the chip usage time refers to the length of time from when the chip starts to be used to the current moment, and the preset chip usage time refers to a preset value for measuring the practicality of the chip, such as setting D0=3 years.

[0089] Specifically, the heat dissipation control module intelligently determines the heat dissipation mode according to the temperature difference value and outputs corresponding control instructions. It can flexibly adapt to different working environments and load conditions, so as to more effectively control the chip temperature. By constructing a chip temperature prediction model, the heat dissipation control module can obtain the predicted value of the chip temperature in advance, which helps to timely discover potential overheating risks and take corresponding preventive measures to avoid chip damage due to overheating, thereby extending the service life of the equipment. By judging in real time whether the chip is within the safe operating temperature range and adjusting the judgment result of the heat dissipation mode according to the judgment result, it is ensured that the chip operates at the optimal operating temperature and improves the stability and reliability of the system. The safe operating temperature value is corrected according to the chip usage time. As the chip usage time increases, its performance will gradually decline, resulting in changes in the safe operating temperature range. By adaptively adjusting the safe operating temperature value, the module can ensure that the chip can maintain the best working state at different stages of use, which helps to maintain the optimal performance state of the system.

[0090] Specifically, the execution feedback module obtains the control instruction output by the heat dissipation control module, transmits the control instruction to the coolant regulating device, analyzes the coolant flow g, and sets Among them, A is the cooling pump speed, a is the efficiency factor, S is the cross-sectional area of ​​the liquid cooling system, b is the correction coefficient, and i is the flow rate of the coolant. is the coolant heat exchange efficiency coefficient, c is the specific heat capacity of the coolant, F is the flow resistance of the liquid cooling and heat dissipation system, d is the influence coefficient of the coolant temperature on the flow rate, the execution feedback module calculates the coolant flow difference e(t) according to the coolant flow g and the real-time coolant flow g1, and sets e(t)=|g-g1|, calculates the control quantity u(t) according to the coolant flow difference e(t) and PID parameters, and sets u(t)=Kp×e(t)+Ki×∫e(t)dt+Kd×[de(t) / dt], the PID parameters include the proportional coefficient Kp, the integral coefficient Ki and the differential coefficient Kd, and the execution feedback module controls the coolant regulating device according to the control quantity u(t).

[0091] Specifically, the coolant regulating device refers to the precise control of the cooling effect by adjusting the circulation flow, temperature and pressure parameters of the coolant. The cooling pump speed refers to the number of revolutions per minute or the arc per second of the cooling pump, which is used to measure the working speed of the cooling pump. The efficiency factor refers to the efficiency of the cooling pump in converting mechanical energy into liquid kinetic energy. The cross-sectional area of ​​the liquid cooling and heat dissipation system refers to the cross-sectional area of ​​the coolant flowing in the pipeline. The correction coefficient refers to the coefficient used to adjust or correct the influence of factors such as pipeline shape, material, and length on the coolant flow rate. The flow rate of the coolant refers to the distance that the coolant passes through a certain cross-section of the pipeline per unit time. The coolant heat exchange efficiency coefficient refers to the efficiency of the coolant in absorbing heat during the heat exchange process. The specific heat capacity of the coolant refers to the amount of heat absorbed by a unit mass of coolant when the temperature rises by 1 degree Celsius. The flow resistance of the liquid cooling and heat dissipation system refers to the resistance encountered by the coolant when flowing in the cooling and heat dissipation system. The influence coefficient of the coolant temperature on the flow rate refers to the coefficient of the influence of the coolant temperature change on the flow rate.

[0092] Specifically, the execution feedback module obtains the coolant flow rate, compares the coolant flow rate G with the preset minimum coolant flow rate G0min and the preset maximum coolant flow rate G0max, judges the operating state of the coolant regulating device according to the comparison result, and outputs the alarm information according to the judgment result, wherein:

[0093] When G<G0min, the execution feedback module determines that the operation state of the coolant regulating device is abnormal, and outputs device blockage alarm information;

[0094] When G0min≤G≤G0max, the execution feedback module determines that the operating state of the coolant regulating device is normal and does not output an alarm message;

[0095] When G>G0max, the execution feedback module determines that the operation state of the coolant regulating device is abnormal, and outputs device leakage alarm information.

[0096] Specifically, the present embodiment does not limit the method for obtaining the coolant flow rate, and relevant technical personnel in the field can freely set it according to actual needs, and only need to meet the needs for obtaining the coolant flow rate. For example, a flow sensor can be installed at the coolant regulating device to obtain the coolant flow rate. The preset minimum coolant flow rate G0min refers to the minimum value of the coolant flow rate set in the liquid cooling and heat dissipation system of the computer chip to ensure that the chip temperature does not exceed the safety threshold, such as setting G0min=5L / min. The preset maximum coolant flow rate G0max refers to the maximum value of the coolant flow rate set in the liquid cooling and heat dissipation system of the computer chip to ensure that the chip temperature does not exceed the safety threshold, such as setting G0max=10L / min. The present embodiment does not limit the output method of the alarm information, and relevant technical personnel in the field can freely set it according to actual needs, and only need to meet the output requirements of the alarm information. For example, an LED indicator light can be installed on the coolant regulating device. When the operating state of the coolant regulating device is abnormal, the LED light will light up and change color to send out an alarm message.

[0097] Specifically, when the execution feedback module dynamically adjusts the preset minimum coolant flow G0min and the preset maximum coolant flow G0max, a preset flow dynamic adjustment model is constructed, and the preset flow database is divided into 70% of the flow training set, 15% of the flow verification set and 15% of the flow test set. The flow training set is input into the random forest model for training to obtain the trained random forest model, and the flow verification set is input into the trained random forest model to iteratively optimize the parameters of the trained random forest model, and the flow test set is input into the iteratively optimized random forest model, and the iteratively optimized recurrent network neural model is tested to obtain the dynamic adjustment accuracy U, and the dynamic adjustment accuracy U is compared with the preset dynamic adjustment accuracy U0, and the training compliance of the random forest model is judged according to the comparison result, and the random forest model is output according to the judgment result, wherein:

[0098] When U≥U0, the random forest model training is determined to have met the standard, and the random forest model is output as the preset traffic automatic adjustment model;

[0099] When U<U0, it is determined that the random forest model training does not meet the standard, the random forest model parameters are iteratively optimized, and the training is repeated until the random forest model training meets the standard;

[0100] The real-time ambient temperature data and the real-time computer load data are input into the preset flow automatic adjustment model to obtain the preset flow automatic adjustment information.

[0101] Specifically, the preset flow database refers to a historical database used to store, manage and analyze load data, ambient temperature data and corresponding coolant flow data, the flow training set refers to a data set used to train the random forest model, the flow verification set refers to a data set used to optimize and evaluate the performance of the random forest model, the flow test set refers to a data set used to test the preset flow automatic adjustment model, the random forest model refers to a neural network model used to process sequence data, the dynamic adjustment accuracy U refers to the proportion of the number of samples X correctly output by the random forest model to the total number of samples X0 in the flow test set, that is, U=X / X0, the preset dynamic adjustment accuracy U0 refers to a preset value for judging the training compliance of the random forest model, this embodiment does not limit the value of the preset dynamic adjustment accuracy U0, relevant technical personnel in this field can freely set it according to actual needs, as long as the need to judge the training compliance of the random forest model is met, such as the preset dynamic adjustment accuracy U0 can be set to 0.96, the preset flow automatic adjustment information refers to dynamically adjusting the preset minimum coolant flow G0min and the preset maximum coolant flow G0max according to the real-time ambient temperature data and the real-time computer load data.

[0102] Specifically, the execution feedback module can quickly and accurately transmit the control instructions output by the heat dissipation control module to the coolant regulating device. According to the different control quantities, the coolant regulating device can be precisely controlled to ensure the optimal heat dissipation effect. The coolant flow rate is monitored in real time, and the operating status of the coolant regulating device is judged according to the coolant flow rate. Once an abnormal operating status is found, an alarm message can be output immediately to remind the operator to take timely measures to avoid overheating or damage to the equipment. The preset flow rate of the coolant is automatically adjusted by constructing a preset flow rate dynamic adjustment model to ensure that the heat dissipation system can maintain optimal performance under different working conditions. Through precise control and real-time status monitoring, the execution feedback module can improve the stability and reliability of the heat dissipation system, help ensure that the computer can maintain a stable working state under long-term and high-load operation, and help reduce equipment operating costs.

[0103] See also Figure 3 As shown, it is a schematic diagram of the process of the liquid cooling and heat dissipation method of a computer chip in this embodiment, including:

[0104] Step S1, collecting chip temperature data, ambient temperature data and computer load data;

[0105] Step S2, judging the heat dissipation mode according to the temperature difference value, and outputting a control instruction according to the judgment result;

[0106] Step S3, constructing a chip temperature prediction model to obtain a chip temperature prediction value;

[0107] Step S4, judging whether the chip is within a safe operating temperature range according to the chip temperature prediction value, and adjusting the judgment result of the heat dissipation mode according to the judgment result;

[0108] Step S5, correcting the safe operating temperature value according to the chip usage time;

[0109] Step S6, transmitting the control instruction output by the heat dissipation control module to the coolant regulating device, and controlling the coolant regulating device according to the control amount;

[0110] Step S7, judging the operating state of the coolant regulating device according to the coolant flow rate, and outputting an alarm message according to the judgment result;

[0111] Step S8, constructing a preset flow dynamic adjustment model, and obtaining preset flow automatic adjustment information according to real-time ambient temperature data and real-time computer load data.

[0112] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A liquid cooling and heat dissipation system for a computer chip, characterized in that: include: A data acquisition module is used to collect chip temperature data, ambient temperature data and computer load data; The heat dissipation control module is used to judge the heat dissipation mode according to the temperature difference value, and output the control instruction according to the judgment result, and is also used to build a chip temperature prediction model to obtain the chip temperature prediction value, and is also used to judge whether the chip is within the safe operating temperature range according to the chip temperature prediction value, and adjust the judgment result of the heat dissipation mode according to the judgment result, and is also used to correct the safe operating temperature value according to the chip usage time; The execution feedback module is used to transmit the control instructions output by the heat dissipation control module to the coolant regulating device, control the coolant regulating device according to the control amount, judge the operating status of the coolant regulating device according to the coolant flow rate, and output alarm information according to the judgment result. It is also used to build a preset flow dynamic adjustment model to obtain the preset flow automatic adjustment information according to the real-time ambient temperature data and the real-time computer load data.

2. The computer chip liquid cooling and heat dissipation system according to claim 1, characterized in that: The heat dissipation control module comprises: A heat dissipation mode matching unit, used to compare the temperature difference value T with a preset minimum temperature difference value T0min and a preset maximum temperature difference value T0max, judge the heat dissipation mode according to the comparison result, and output a control instruction according to the judgment result; The chip temperature prediction unit is used to construct a chip temperature prediction model, input computer load data into the chip temperature prediction model, obtain a chip temperature prediction value, compare the chip temperature prediction value A with the safe operating temperature value A0, judge whether the chip is within the safe operating temperature range according to the comparison result, and adjust the judgment result of the heat dissipation mode according to the judgment result, and compare the chip usage time D with the preset chip usage time D0, judge the practicality of the chip according to the comparison result, and correct the safe operating temperature value A0 according to the judgment result.

3. The computer chip liquid cooling and heat dissipation system according to claim 2, characterized in that: The heat dissipation mode matching unit obtains the temperature difference T between the chip temperature and the ambient temperature, sets T=T1-T2, where T1 is the chip temperature and T2 is the ambient temperature, compares the temperature difference T with the preset minimum temperature difference T0min and the preset maximum temperature difference T0max, judges the heat dissipation mode according to the comparison result, and outputs a control instruction according to the judgment result, wherein: When T<T0min, the heat dissipation mode matching unit determines that the heat dissipation mode is a low power consumption heat dissipation mode and outputs a low power consumption control instruction; When T0min≤T≤T0max, the heat dissipation mode matching unit determines that the heat dissipation mode is a medium power dissipation mode and outputs a medium power dissipation control instruction; When T>T0max, the heat dissipation mode matching unit determines that the heat dissipation mode is a high power consumption heat dissipation mode, and outputs a high power consumption control instruction.

4. The computer chip liquid cooling and heat dissipation system according to claim 2, characterized in that: When predicting the chip temperature, the chip temperature prediction unit constructs a chip temperature prediction model, divides the preset chip temperature database into 80% of the chip temperature training set, 10% of the chip temperature verification set and 10% of the chip temperature test set, inputs the chip temperature training set into the recurrent neural network model for training to obtain the trained recurrent neural network model, inputs the chip temperature verification set into the trained recurrent neural network model to iteratively optimize the parameters of the trained recurrent neural network model, inputs the chip temperature test set into the iteratively optimized recurrent neural network model, tests the iteratively optimized convolutional network neural model to obtain the prediction accuracy R, compares the prediction accuracy R with the preset prediction accuracy R0, judges the training compliance of the recurrent neural network model according to the comparison result, and outputs the recurrent neural network model according to the judgment result, wherein: When R≥R0, it is determined that the recurrent neural network model training has reached the standard, and the recurrent neural network model is output as a chip temperature prediction model; When R<R0, it is determined that the recurrent neural network model training does not meet the standard, the recurrent neural network model parameters are iteratively optimized, and the training is repeated until the recurrent neural network model training meets the standard; The computer load data is input into the chip temperature prediction model to obtain the chip temperature prediction value.

5. The computer chip liquid cooling and heat dissipation system according to claim 4, characterized in that: The chip temperature prediction unit compares the chip temperature prediction value A with the safe operating temperature value A0, judges whether the chip is within the safe operating temperature range according to the comparison result, and adjusts the judgment result of the heat dissipation mode according to the judgment result, wherein: When A≥A0, the chip temperature prediction unit determines that the chip is not within the safe operating temperature range, and adjusts the judgment result of the heat dissipation mode, wherein: If the cooling mode is a low power cooling mode, the adjusted cooling mode is a medium power cooling mode; If the cooling mode is a medium power cooling mode, the adjusted cooling mode is a high power cooling mode; If the heat dissipation mode is a high power consumption heat dissipation mode, the adjusted heat dissipation mode is a high power consumption heat dissipation mode; When A<A0, the chip temperature prediction unit determines that the chip is within a safe operating temperature range and does not adjust the determination result of the heat dissipation mode.

6. The computer chip liquid cooling and heat dissipation system according to claim 5, characterized in that: The chip temperature prediction unit compares the chip usage time D with the preset chip usage time D0, judges the practicality of the chip according to the comparison result, and corrects the safe operating temperature value A0 according to the judgment result, wherein: When D≥D0, the chip temperature prediction unit determines that the practicality of the chip is weak, sets a correction coefficient a, and corrects the safe operating temperature value A0 according to the correction coefficient. The corrected safe operating temperature value is Aa, and Aa=A0×a is set; When D<D0, the chip temperature prediction unit determines that the chip has high practicality and corrects the safe operating temperature value A0.

7. The computer chip liquid cooling and heat dissipation system according to claim 1, characterized in that: The execution feedback module obtains the control instruction output by the heat dissipation control module, transmits the control instruction to the coolant regulating device, analyzes the coolant flow g, and sets g=(A×a×S b ×i×η o ) / (c×F d ), where A is the cooling pump speed, a is the efficiency factor, S is the cross-sectional area of ​​the liquid cooling system, b is the correction coefficient, i is the flow rate of the coolant, η o is the coolant heat exchange efficiency coefficient, c is the specific heat capacity of the coolant, F is the flow resistance of the liquid cooling and heat dissipation system, d is the influence coefficient of the coolant temperature on the flow rate, the execution feedback module calculates the coolant flow difference e(t) according to the coolant flow g and the real-time coolant flow g1, and sets e(t)=|g-g1|, calculates the control quantity u(t) according to the coolant flow difference e(t) and PID parameters, and sets u(t)=Kp×e(t)+Ki×∫e(t)dt+Kd×[de(t) / dt], the PID parameters include the proportional coefficient Kp, the integral coefficient Ki and the differential coefficient Kd, and the execution feedback module controls the coolant regulating device according to the control quantity (ut).

8. The computer chip liquid cooling and heat dissipation system according to claim 1, characterized in that: The execution feedback module obtains the coolant flow rate, compares the coolant flow rate G with the preset minimum coolant flow rate G0min and the preset maximum coolant flow rate G0max, judges the operating state of the coolant regulating device according to the comparison result, and outputs alarm information according to the judgment result, wherein: When G<G0min, the execution feedback module determines that the operation state of the coolant regulating device is abnormal, and outputs device blockage alarm information; When G0min≤G≤G0max, the execution feedback module determines that the operating state of the coolant regulating device is normal and does not output an alarm message; When G>G0max, the execution feedback module determines that the operation state of the coolant regulating device is abnormal, and outputs device leakage alarm information.

9. The computer chip liquid cooling and heat dissipation system according to claim 1, characterized in that: When the execution feedback module dynamically adjusts the preset minimum coolant flow G0min and the preset maximum coolant flow G0max, a preset flow dynamic adjustment model is constructed, the preset flow database is divided into 70% of the flow training set, 15% of the flow verification set and 15% of the flow test set, the flow training set is input into the random forest model for training to obtain the trained random forest model, the flow verification set is input into the trained random forest model to iteratively optimize the parameters of the trained random forest model, and the flow test set is input into the iteratively optimized random forest model to test the iteratively optimized recurrent network neural model to obtain the dynamic adjustment accuracy U, the dynamic adjustment accuracy U is compared with the preset dynamic adjustment accuracy U0, the training compliance of the random forest model is judged according to the comparison result, and the random forest model is output according to the judgment result, wherein: When U≥U0, the random forest model training is determined to have met the standard, and the random forest model is output as the preset traffic automatic adjustment model; When U<U0, it is determined that the random forest model training does not meet the standard, the random forest model parameters are iteratively optimized, and the training is repeated until the random forest model training meets the standard; The real-time ambient temperature data and the real-time computer load data are input into the preset flow automatic adjustment model to obtain the preset flow automatic adjustment information.

10. The computer chip liquid cooling and heat dissipation method according to claim 1, characterized in that: include: Step S1, collecting chip temperature data, ambient temperature data and computer load data; Step S2, judging the heat dissipation mode according to the temperature difference value, and outputting a control instruction according to the judgment result; Step S3, constructing a chip temperature prediction model to obtain a chip temperature prediction value; Step S4, judging whether the chip is within a safe operating temperature range according to the chip temperature prediction value, and adjusting the judgment result of the heat dissipation mode according to the judgment result; Step S5, correcting the safe operating temperature value according to the chip usage time; Step S6, transmitting the control instruction output by the heat dissipation control module to the coolant regulating device, and controlling the coolant regulating device according to the control amount; Step S7, judging the operating state of the coolant regulating device according to the coolant flow rate, and outputting an alarm message according to the judgment result; Step S8, constructing a preset flow dynamic adjustment model, and obtaining preset flow automatic adjustment information according to real-time ambient temperature data and real-time computer load data.

Images