Heat dissipation control system, method and device, medium and program product

Abstract

The invention discloses a heat dissipation control system, method and device, a medium and a program product in the technical field of computers. According to the invention, the heat dissipation control module in the master control node is in network connection with the slave management controller in any cabinet node; therefore, the heat dissipation control module obtains equipment operation information of at least one piece of computing equipment in the corresponding cabinet node through the slave management controller in any cabinet node, operation risk prediction data is obtained according to the equipment operation information, and a fan regulation and control strategy corresponding to the operation risk prediction data is configured to the slave management controller; and the slave management controller performs rotating speed control on the fans around the at least one computing device according to the fan regulation and control strategy. Therefore, temperature risk prediction is carried out in time according to the actual operation condition of the computing equipment, and accurate heat dissipation control is carried out on the fans in the area where the corresponding computing equipment is located, so that the master control node can predict the temperature and heat dissipation condition of each cabinet node, and heat dissipation early warning and management and control of each cabinet are realized.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a heat dissipation control system, method, device, medium and program product. Background Art

[0002] Currently, multiple BOX (cabinet) nodes can be connected to the same host. Each cabinet has its own management controller, such as BMC; the management controller in each cabinet manages the cooling fans of the corresponding cabinet. The host cannot predict the cooling conditions of each cabinet and cannot jointly control the cooling of each cabinet.

[0003] Therefore, how to provide early warning and control of the heat dissipation of each cabinet is a problem that technical personnel in this field need to solve. Summary of the invention

[0004] In view of this, the purpose of this application is to provide a heat dissipation control system, method, device, medium and program product to provide early warning and control of the heat dissipation of each cabinet. The specific scheme is as follows:

[0005] In a first aspect, the present application provides a heat dissipation control system, comprising: a master control node and at least one cabinet node; the master control node comprises a heat dissipation control module; the at least one cabinet node comprises a slave management controller, at least one computing device and a plurality of fans;

[0006] Wherein, the heat dissipation control module is connected to the slave management controller network in any cabinet node;

[0007] The heat dissipation control module is used to: obtain device operation information of at least one computing device in the corresponding cabinet node through a slave management controller in any cabinet node, obtain operation risk prediction data according to the device operation information, and configure the fan control strategy corresponding to the operation risk prediction data to the slave management controller;

[0008] The slave management controller is used to control the rotation speed of the fans around the at least one computing device according to the fan regulation strategy.

[0009] Optionally, the heat dissipation control module is used to: obtain the core temperature, memory temperature and power consumption data of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node to obtain the device operation information; normalize the core temperature, the memory temperature and the power consumption data; calculate and obtain comprehensive indicator data based on the normalized core temperature, the normalized memory temperature and the normalized power consumption data; and obtain the operation risk prediction data based on the comprehensive indicator data.

[0010] Optionally, the heat dissipation control module is used for: if the rising change rate of the comprehensive index data and the previous comprehensive index data is greater than a preset index threshold, the rising change rate of the normalized core temperature and the previous core temperature is greater than the preset core temperature threshold, the rising change rate of the normalized memory temperature and the previous memory temperature is greater than the preset memory temperature threshold, or the rising change rate of the normalized power consumption data and the previous power consumption data is greater than the preset power consumption data threshold, then it is determined that the operation risk prediction data is high-risk data, and the fan control strategy corresponding to the operation risk prediction data is determined to be a rapid cooling strategy.

[0011] Optionally, the slave management controller is used to: determine identification information of fans around the at least one computing device, and control the fans with the identification information to run at full speed according to the rapid cooling strategy.

[0012] Optionally, the heat dissipation control module is used to: determine a corresponding task volume control strategy based on the operation risk prediction data, and send the task volume control strategy to the operating system of the master control node, so that the operating system controls the business volume transmission speed reaching the at least one computing device according to the task volume control strategy.

[0013] Optionally, the master control node includes a master management controller and a plurality of fans;

[0014] Accordingly, the heat dissipation control module is used to: obtain the device operation information in the main control node through the main management controller, and update the fan control strategy in the main management controller according to the device operation information;

[0015] The main management controller is used to control the rotation speed of the fan in the main control node according to the updated fan control strategy.

[0016] Optionally, the master control node includes a processor, a master management controller and a target port; the master management controller is connected to a slave management controller in any cabinet node;

[0017] The at least one cabinet node comprises a switching module; the switching module is connected to at least one computing device in the corresponding cabinet node;

[0018] The processor is connected to a switch module in any cabinet node through the target port.

[0019] Optionally, the master management controller is used to: when confirming that the slave management controller in any cabinet node is in an activated state and the cabinet node is in a powered-on state, control the master control node to power on and activate the master management controller.

[0020] Optionally, the master management controller is used to: when confirming that the slave management controller in any cabinet node is in an inactivated state and the cabinet node is in a shutdown state, control the master control node to be shut down and control the master management controller to be in an inactivated state.

[0021] Optionally, the master management controller is used to confirm the state of the slave management controller in any cabinet node and the state of the corresponding cabinet node by reading a register in the slave management controller.

[0022] Optionally, the main management controller is used to: after the main control node is powered on, activate the heat dissipation control module through the operating system of the main control node.

[0023] Optionally, the main management controller is used to: after the main control node is powered on, load a configuration file stored in itself, and control the speed of the fan in the main control node according to the configuration file;

[0024] The slave management controller in any cabinet node is used to: after the corresponding cabinet node is powered on, load the configuration file stored in itself, and control the speed of the fan in the corresponding cabinet node according to the configuration file.

[0025] In a second aspect, the present application provides a heat dissipation control method, which is applied to a heat dissipation control module, wherein the heat dissipation control module is arranged in a master control node, and the master control node includes the heat dissipation control module; the heat dissipation control module is network-connected to a slave management controller in at least one cabinet node; the at least one cabinet node includes a slave management controller, at least one computing device, and a plurality of fans;

[0026] The method includes:

[0027] Obtaining, through a slave management controller in any cabinet node, device operation information of at least one computing device in the corresponding cabinet node;

[0028] Obtaining operation risk prediction data according to the equipment operation information;

[0029] The fan regulation policy corresponding to the operation risk prediction data is configured in the slave management controller, so that the slave management controller controls the rotation speed of the fan around the at least one computing device according to the fan regulation policy.

[0030] Optionally, the device operation information of at least one computing device in the corresponding cabinet node is obtained through the slave management controller in any cabinet node, including: obtaining the core temperature, memory temperature and power consumption data of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node to obtain the device operation information.

[0031] Optionally, obtaining operation risk prediction data according to the equipment operation information includes:

[0032] Normalizing the core temperature, the memory temperature, and the power consumption data;

[0033] The comprehensive index data is calculated based on the normalized core temperature, normalized memory temperature and normalized power consumption data;

[0034] The operational risk prediction data is obtained according to the comprehensive indicator data.

[0035] Optionally, obtaining the operation risk prediction data according to the comprehensive indicator data includes:

[0036] If the rising change rate of the comprehensive index data and the previous comprehensive index data is greater than the preset index threshold, it is determined that the operation risk prediction data belongs to high-risk data, and the fan control strategy corresponding to the operation risk prediction data is determined to be a rapid cooling strategy.

[0037] Optionally, if the rising rate of change between the normalized core temperature and the previous core temperature is greater than a preset core temperature threshold, the rising rate of change between the normalized memory temperature and the previous memory temperature is greater than a preset memory temperature threshold, or the rising rate of change between the normalized power consumption data and the previous power consumption data is greater than a preset power consumption data threshold, then it is determined that the operation risk prediction data is high-risk data, and the fan control strategy corresponding to the operation risk prediction data is determined to be a rapid cooling strategy.

[0038] Optionally, the management controller controls the speed of fans around the at least one computing device according to the fan regulation strategy, including: determining identification information of the fans around the at least one computing device from the management controller, and controlling the fans with the identification information to run at full speed according to the rapid cooling strategy.

[0039] Optionally, it also includes: a heat dissipation control module determines a corresponding task volume control strategy based on the operation risk prediction data, and sends the task volume control strategy to the operating system of the main control node, so that the operating system controls the business volume transmission speed reaching the at least one computing device according to the task volume control strategy.

[0040] Optionally, the main control node includes a main management controller and several fans; accordingly, the heat dissipation control module obtains the device operation information in the main control node through the main management controller, and updates the fan control strategy in the main management controller according to the device operation information, so that the main management controller controls the speed of the fans in the main control node according to the updated fan control strategy.

[0041] Optionally, the master control node includes a processor, a master management controller and a target port; the master management controller is connected to a slave management controller in any cabinet node; at least one cabinet node includes a switching module; the switching module is connected to at least one computing device in the corresponding cabinet node; the processor is connected to the switching module in any cabinet node through the target port.

[0042] In a third aspect, the present application provides an electronic device, including:

[0043] Memory for storing computer programs;

[0044] The processor is used to execute the computer program to implement the heat dissipation control method disclosed above.

[0045] In a fourth aspect, the present application provides a non-volatile storage medium for storing a computer program, wherein the computer program implements the aforementioned disclosed heat dissipation control method when executed by a processor.

[0046] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which implements the steps of the aforementioned disclosed heat dissipation control method when executed by a processor.

[0047] It can be seen from the above scheme that the present application provides a heat dissipation control system, including: a master control node and at least one cabinet node; the master control node includes a heat dissipation control module; the at least one cabinet node includes a slave management controller, at least one computing device and a plurality of fans; wherein the heat dissipation control module is network-connected to the slave management controller in any cabinet node; the heat dissipation control module is used to: obtain device operation information of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node, obtain operation risk prediction data according to the device operation information, and configure the fan control strategy corresponding to the operation risk prediction data to the slave management controller; the slave management controller is used to: control the speed of the fan around the at least one computing device according to the fan control strategy.

[0048] It can be seen that the beneficial effects of the present application are: the heat dissipation control module in the master control node is connected to the slave management controller network in any cabinet node; thereby, the heat dissipation control module obtains the device operation information of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node, obtains the operation risk prediction data according to the device operation information, and configures the fan control strategy corresponding to the operation risk prediction data to the slave management controller; the slave management controller controls the speed of the fan around at least one computing device according to the fan control strategy. In this way, the temperature risk prediction can be carried out in real time according to the actual operation of the computing device, and the fan in the area where the corresponding computing device is located can be accurately controlled according to the corresponding prediction results, so that the master control node can predict the temperature and heat dissipation of each cabinet node, and realize the heat dissipation warning and control of each cabinet.

[0049] Correspondingly, a heat dissipation control method, device, medium and program product provided by the present application also have the above-mentioned technical effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0051] Figure 1 A schematic diagram of a heat dissipation control system disclosed in this application;

[0052] Figure 2 This is a schematic diagram of a second heat dissipation control system disclosed in this application;

[0053] Figure 3 A schematic diagram of a heat dissipation control module disclosed in this application;

[0054] Figure 4 A flow chart of a heat dissipation control method disclosed in this application;

[0055] Figure 5 A schematic diagram of an electronic device disclosed in this application;

[0056] Figure 6 A server structure diagram provided for this application;

[0057] Figure 7 A terminal structure diagram provided for this application. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other examples obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0059] Currently, multiple BOX (cabinet) nodes can be connected to the same host. Each cabinet has its own management controller, such as BMC; the management controller in each cabinet manages the cooling fan of the corresponding cabinet. The host cannot predict the cooling conditions of each cabinet, and cannot jointly control the cooling of each cabinet. To this end, the present application provides a cooling control solution that can instantly predict temperature risks based on the actual operation of the computing equipment, and accurately control the cooling of the fans in the area where the corresponding computing equipment is located based on the corresponding prediction results, so that the master control node can predict the temperature and cooling conditions of each cabinet node, and realize the cooling warning and control of each cabinet.

[0060] See also Figure 1 As shown, an embodiment of the present application discloses a heat dissipation control system, comprising: a master control node and at least one cabinet node; the master control node comprises a heat dissipation control module; at least one cabinet node comprises a slave management controller, at least one computing device and a plurality of fans; wherein the heat dissipation control module is network-connected to the slave management controller in any cabinet node.

[0061] In this embodiment, the heat dissipation control module is used to obtain the device operation information of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node, obtain the operation risk prediction data according to the device operation information, and configure the fan control strategy corresponding to the operation risk prediction data to the slave management controller; accordingly, the slave management controller is used to control the speed of the fan around at least one computing device according to the fan control strategy. The identification information of the fan around any computing device can be recorded in advance, and the fan around any computing device and its identification information can be determined based on the record.

[0062] Among them, the equipment operation information may include core temperature, memory temperature and power consumption data, etc. In one embodiment, the heat dissipation control module is used to: obtain the core temperature, memory temperature and power consumption data of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node to obtain the equipment operation information; normalize the core temperature, memory temperature and power consumption data; calculate the comprehensive index data based on the normalized core temperature, normalized memory temperature and normalized power consumption data; obtain the operation risk prediction data according to the comprehensive index data. Among them, the correspondence table between the comprehensive index data and the corresponding risks can be pre-set. For example: when the comprehensive index data is 20, it corresponds to low-risk operation risk prediction data; when the comprehensive index data is 80, it corresponds to high-risk operation risk prediction data.

[0063] Accordingly, the heat dissipation control module is used to: if the rising change rate of the comprehensive index data and the previous comprehensive index data is greater than the preset index threshold, the rising change rate of the normalized core temperature and the previous core temperature is greater than the preset core temperature threshold, the rising change rate of the normalized memory temperature and the previous memory temperature is greater than the preset memory temperature threshold, or the rising change rate of the normalized power consumption data and the previous power consumption data is greater than the preset power consumption data threshold, then determine that the operation risk prediction data belongs to high-risk data, and determine that the fan control strategy corresponding to the operation risk prediction data is a rapid cooling strategy. Accordingly, the slave management controller is used to: determine the identification information of the fan around at least one computing device, and control the fan with the identification information to run at full speed according to the rapid cooling strategy.

[0064] It should be noted that if the core temperature, memory temperature and / or power consumption data are high, it means that the corresponding computing device may be processing a large amount of data and its load is large. Therefore, in one embodiment, the heat dissipation control module is also used to: determine the corresponding task volume control strategy based on the operation risk prediction data, and send the task volume control strategy to the operating system of the main control node, so that the operating system controls the business volume transmission speed reaching at least one computing device according to the task volume control strategy, so as to reduce the amount of data that the corresponding computing device needs to process.

[0065] The main control node also needs to perform temperature control. Therefore, in one embodiment, the main control node includes a main management controller and several fans. Accordingly, the heat dissipation control module is used to: obtain the device operation information in the main control node through the main management controller, and update the fan control strategy in the main management controller according to the device operation information; the main management controller is used to: control the speed of the fan in the main control node according to the updated fan control strategy.

[0066] In one embodiment, the communication connection between the master control node and at least one cabinet node can refer to Figure 2 ,like Figure 2 As shown, the master control node includes a processor, a master management controller (i.e., a master BMC) and a target port (i.e., a PCIe port); the master management controller is connected to a slave management controller (i.e., a slave BMC) in any cabinet node; at least one cabinet node includes a switch module; the switch module (i.e., SWITCH) is connected to at least one computing device in the corresponding cabinet node, such as Figure 2 The GPU and FPGA in the rack are connected to the switching module in any cabinet node through the target port.

[0067] It should be noted that after the master control node is powered on, the computing devices in the cabinet node will be enumerated, so it is necessary to ensure that the power-on time of the cabinet node is earlier than the power-on time of the master control node. If the power-on time of the cabinet node is later than the power-on time of the master control node, the master control node will not be able to enumerate the computing devices in the cabinet node after it is powered on, which will cause enumeration identification, and the master control node will not be able to manage and control the computing devices in the cabinet node. Therefore, in one embodiment, the master management controller is used to: when it is confirmed that the slave management controller in any cabinet node is in an activated state and the cabinet node is in a powered-on state, control the master control node to power on and activate the master management controller. Correspondingly, the master management controller is used to: when it is confirmed that the slave management controller in any cabinet node is in an inactivated state and the cabinet node is in a powered-off state, control the master control node to be powered off and control the master management controller to be in an inactivated state. In one embodiment, the master management controller is used to: confirm the state of the slave management controller in any cabinet node and the state of the corresponding cabinet node by reading the register in the slave management controller.

[0068] In one embodiment, the master management controller is used to: after the master control node is powered on, activate the heat dissipation control module through the operating system of the master control node. After the master control node is powered on, the configuration file stored in itself is loaded, and the speed of the fan in the master control node is controlled according to the configuration file; the slave management controller in any cabinet node is used to: after the corresponding cabinet node is powered on, the configuration file stored in itself is loaded, and the speed of the fan in the corresponding cabinet node is controlled according to the configuration file. The original fan control strategy is stored in the configuration file.

[0069] In this embodiment, the management controller may be a BMC (Baseboard Management Controller) etc. The BMC can monitor and manage the status of each hardware in the corresponding node.

[0070] It can be seen that this embodiment enables the heat dissipation control module in the master control node to be connected to the slave management controller network in any cabinet node; thereby, the heat dissipation control module obtains the device operation information of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node, obtains the operation risk prediction data according to the device operation information, and configures the fan control strategy corresponding to the operation risk prediction data to the slave management controller; the slave management controller controls the speed of the fan around at least one computing device according to the fan control strategy. In this way, temperature risk prediction can be performed in real time according to the actual operation of the computing device, and accurate heat dissipation control can be performed on the fans in the area where the corresponding computing device is located according to the corresponding prediction results, so that the master control node can predict the temperature and heat dissipation of each cabinet node, and realize the heat dissipation warning and control of each cabinet.

[0071] See also Figure 2 , if the master node ( Figure 2 The computing node in the cabinet is connected to a BOX node in the cabinet. The specific connection includes: the computing node includes a master BMC, the BOX node includes a slave BMC, and the master and slave BMCs are connected through an I2C link. Among them, the computing node also includes a hardware CPU and a PCIe port, and the BOX node includes a hardware switching module SWITCH. The PCIe port of the computing node and the SWITCH are connected through a cable to extend the PCIe link. The SWITCH can reserve a PCIe slot to provide interconnection between GPU, FPGA, and HDD components, thereby realizing interconnection access from the CPU to key components in the BOX.

[0072] In the computing node, the main BMC is connected to the heat dissipation control unit (i.e., heat dissipation control module) through an external RJ45 network cable, and the slave BMC of the BOX node is also connected to the heat dissipation control unit through an external RJ45 network cable. At the same time, in the computing node, the OS, as a server system, performs software control and linkage with the CPU, main BMC, and heat dissipation control unit.

[0073] See also Figure 3 The heat dissipation control unit includes a control chip, a computing chip, and a SWITCH network. The computing chip is mainly used for data processing and performs higher-density calculations on heat dissipation data. The control chip controls the data exchange between the computing chip, the SWITCH network, the master BMC, and the slave BMC through firmware, and can achieve coordination between server function control and heat dissipation management control through the driver software in the OS. The SWITCH network realizes the interconnection and forwarding functions of the master BMC and the slave BMC network from the hardware level. At the same time, the SWITCH network also provides an external network interface for the computing node to achieve network connection with the master BMC and the slave BMC.

[0074] It should be noted that Figure 2The operation logic of the system shown includes: first, the computing node server and the BOX node server are powered on but not turned on. After the master BMC and the slave BMC are activated respectively, their respective heat dissipation controls are performed based on their respective default configuration files to ensure that the fan speed is controlled to the minimum when the hardware is not overheated in the shutdown state. At this time, the computing chip and the control chip of the heat dissipation control unit are in an unpowered state, the SWITCH network component is powered on, and the network interconnection and forwarding of the master BMC and the slave BMC are autonomously realized.

[0075] Furthermore, the master BMC obtains the status of the slave BMC through the I2C link. If the slave BMC is not activated, the master BMC restricts the computing node through the CPLD in the computing node so that it cannot be powered on when the BOX node is powered off. After the BOX node is powered on, the CPLD in the computing node receives the BOX node power-on signal and removes the power-on restriction on the computing node. The master BMC scans the activation status of the slave BMC and executes the computing node power-on action. Then the control chip and computing chip of the heat dissipation control unit are powered on. As the system enters the OS, the driver of the control chip is successfully loaded and the heat dissipation control unit is activated.

[0076] Furthermore, the control chip of the heat dissipation control unit periodically obtains the temperature and other sensor data in the corresponding nodes from the master BMC and the slave BMC under the control of the OS driver, and obtains the usage information of the CPU and BOX core components GPU, FPGA, and HDD under the OS, and pushes the captured data to the computing chip for data calculation and processing. The HDD can also be an SSD.

[0077] Among them, the temperature and other sensor data include: the air inlet temperature of the BOX node, the air outlet temperature, the fan speed of the BOX node, the core temperature of the GPU of the BOX node, the memory temperature of the GPU, the switch chip temperature of the BOX node, the temperature of the FPGA card installed in the BOX node, etc. For the usage information, taking the GPU as an example, the usage information includes: the core usage rate of the GPU, the memory usage rate of the GPU, the core clock frequency of the GPU, the memory clock frequency of the GPU, and the power consumption of the GPU.

[0078] The computing chip of the thermal control unit first normalizes the collected data. For example, for GPU core utilization and video memory utilization, their percentage values ​​can be directly used as normalized values. For GPU core temperature and video memory temperature, they can be converted into proportional values ​​to the maximum safe temperature. An example is as follows: If the GPU maximum core safe temperature is 88°C and the current core temperature is 80°C, the normalized value is 80 / 88=0.9. For power consumption, it can also be converted into a proportional value to the maximum power consumption.

[0079] In one example, the core temperature normalization value: Tcore_norm=Tcore / Tcore_max; the memory temperature normalization value: Tmem_norm=Tmem / Tmem_max; the power consumption normalization value: Pnorm=P / Pmax; wherein Tcore and Tmem are the current core temperature and memory temperature respectively, Tcore_max and Tmem_max are the maximum safe temperatures of the core and memory respectively; P is the current power consumption, and Pmax is the maximum power consumption.

[0080] After that, you can calculate the comprehensive index by setting the weight of each parameter. For example: core usage weight w1, memory usage weight w2, core temperature weight w3, memory temperature weight w4, power consumption weight w5, then the comprehensive index S = w1 × core usage + w2 × memory usage + w3 × core temperature normalized value + w4 × memory temperature normalized value + w5 × power consumption normalized value. ‌

[0081] The next step is to adjust the fan speed of the corresponding area according to the comprehensive index. First, set the threshold of the fan speed, for example: when S<0.5, the fan speed is low speed; when 0.5≤S<0.8, the fan speed is medium speed; when S≥0.8, the fan speed is high speed. Alternatively, a more sophisticated linear or nonlinear function can be used to map the comprehensive index to the fan speed.

[0082] In an example, assuming that the weights of each parameter are: w1=0.1, w2=0.1, w3=0.3, w4=0.3, w5=0.2, then the comprehensive index S=0.1×core usage+0.1×memory usage+0.3×core maximum safe temperature+0.3×memory maximum safe temperature+0.2×maximum power consumption. For fan speed adjustment: When S<0.5S<0.5, fan speed=30%. When 0.5≤S<0.8, fan speed=30%+(S-0.5)×150% (linear mapping). When S≥0.8S≥0.8, fan speed=100%; where the fan speed range is 0-100%.

[0083] In order to cope with sudden changes, the rise rates of core temperature, memory temperature, and power consumption can be calculated separately. The rise rate refers to the ratio of the change of a value within a period of time to the time interval. For temperature and power consumption values, their instantaneous rise rates can be calculated. For example: 5 samples are taken per second, and then the change rate between adjacent sampling points is calculated. Take the instantaneous rise rate of the core temperature value as an example: the instantaneous rise rate of the core temperature value = ΔTcore / Δt; ΔTcore is the change of the core temperature value within the time interval Δt. When the rise rate exceeds a certain predetermined value, an alarm is triggered or other actions are taken, such as increasing the fan speed, reducing the workload, etc. For example: when the instantaneous rise rate of the core temperature value is <5, the abnormal fan speed adjustment is not triggered, and the default control conditions are maintained; when the instantaneous rise rate of the core temperature value is >5 and less than 10, the corresponding fan speed is adjusted to 100%; when the instantaneous rise rate of the core temperature is >10, feedback adjustment is made to reduce the workload.

[0084] It should be noted that after the computing chip of the cooling unit returns the data calculation results, the driver triggers the control instruction to the control chip according to the preset conditions. The control chip distributes the control of the BMC to the computing node or BOX node through the SWITCH network, and the master BMC or the slave BMC realizes fine control of the fan speed. At the same time, the control chip can also transmit the analysis data back to the server room control center through the external network interface of the SWITCH network to adjust the status and proportion of the business flow.

[0085] In one implementation, a heterogeneous server can be deployed, including computing nodes and BOX nodes, each of which is equipped with an independent BMC. The computing node selects a 2U dual-core server with a CPU of 128 cores; 64G memory, a total of 32; 6 2U fans for heat dissipation. The BOX node selects a PCIe Gen4 SWITCH board with 8 GPUs, and the computing node and BOX node are interconnected through PCIe links through cables.

[0086] A heat dissipation control unit is installed on the computing node in the form of a PCIe card. Four RJ45 1G network ports are provided externally. The control chip and computing chip are integrated into the PCB board. The PCB board provides a gen4 connection rate and a x16 connection bandwidth.

[0087] Connect the main BMC control network port of the computing node to RJ45 network port 1 and RJ45 network port 2 of the control unit card from the BMC control network port, and then use a network cable to connect the RJ45 network port 3 of the control unit to the laptop. The status indicators of RJ45 network ports 1, 2, and 3 are all off.

[0088] After the computing nodes and BOX nodes are powered on, you can observe that the BMC status indicator of each node starts to flicker after a while, and the RJ45 network ports 1, 2, and 3 of the cooling control unit all start to flash, indicating that the BMC of each node is activated, and the SWITCH network part of the control unit starts to work, and data processing and forwarding of each network can be successfully performed.

[0089] At this time, the default cooling configuration files are loaded on the master BMC and slave BMC respectively to ensure that the fan speed is controlled at the lowest level when the server is not turned on to prevent hardware overheating.

[0090] At this time, press the compute node power key. Due to CPLD control, the BOX node is not powered on, and the compute node cannot be powered on. Further, press the BOX node power, and the power status indicator changes from yellow to green. After a while, the compute node BMC scans that the BOX node is powered on, and it automatically powers on. The compute node power status indicator changes from yellow to green. At this time, the heat dissipation control unit status indicator lights up, and the control chip starts to initialize the firmware.

[0091] Furthermore, as the computing node is turned on, the system enters the operating system OS, the thermal control unit driver under the OS is successfully loaded, and the thermal control unit is activated. At this time, the control network IP of the control unit and the master BMC IP and slave BMC IP managed separately can be viewed through the relevant commands of the driver under the OS. At this time, through the control software of the laptop, enter the control network IP, user name and password of the control unit to view the operating status of the thermal control unit.

[0092] Under the control of the OS driver, the control chip of the heat dissipation unit regularly obtains sensor data of key nodes from the main BMC and the slave BMC. Under the default configuration, data is obtained every 1 minute. The heat dissipation related data in this embodiment includes the temperature of each core component, the air inlet and outlet temperature of each node, the fan speed of each node, etc. At the same time, the control chip obtains the CPU and GPU usage data under the OS, CPU / GPU occupancy, temperature, power consumption, frequency, and currently running task or process information.

[0093] In the OS, CPU0 is loaded with the stress command, GPU6 and 7 are stressed with gpu-burn, and the auxiliary adjustment function of the heat dissipation control unit is disabled. At this time, the heat dissipation control is implemented by each BMC. After a period of time, the fans of the computing nodes and BOX nodes rotate at high speeds, exceeding 80% of the maximum speed, and generating high noise. The auxiliary adjustment function of the heat dissipation control unit is enabled in the control software, and the heat dissipation control unit returns the corresponding fan speed control plan based on the continuously collected data.

[0094] Calculate the comprehensive indicators corresponding to each GPU, and obtain the fan speed setting parameters of the corresponding area based on the comprehensive indicators. Distribute the speed ratio control instructions of each fan to the slave BMC, which further issues control instructions. For example: the fan speed in the GPU0 area is maintained, the fan speed in the GPU1 area is lowered, the fan speed in the GPU6 and 7 areas is maintained, and the fan speed in other areas is reduced accordingly, so that the fan noise is effectively reduced.

[0095] In this embodiment, the auxiliary adjustment function provided by the heat dissipation control unit can be turned on or off. When the auxiliary adjustment function of the heat dissipation control unit is disabled, the heat dissipation control is implemented by each BMC; if the auxiliary adjustment function of the heat dissipation control unit is turned on, the heat dissipation control unit collects the actual operation data of each device and sends the fan speed control instruction to the slave BMC accordingly.

[0096] It can be seen that this embodiment adjusts the fan speed with lower energy consumption based on the actual business flow situation; based on the previous data analysis, it can find the heat dissipation abnormal points that have not generated alarms, and provide risk warnings and avoidance in time.

[0097] A heat dissipation control method provided in an embodiment of the present application is introduced below. The heat dissipation control method described below can be referenced to other embodiments described in this document.

[0098] An embodiment of the present application discloses a heat dissipation control method, which is applied to a heat dissipation control module. The heat dissipation control module is arranged in a main control node, and the main control node includes the heat dissipation control module; the heat dissipation control module is network-connected to a slave management controller in at least one cabinet node; at least one cabinet node includes a slave management controller, at least one computing device and a plurality of fans.

[0099] See also Figure 4 As shown, a heat dissipation control method disclosed in an embodiment of the present application includes:

[0100] S401. Obtain device operation information of at least one computing device in a corresponding cabinet node through a slave management controller in any cabinet node.

[0101] S402. Obtain operation risk prediction data based on equipment operation information.

[0102] S403: configuring a fan control strategy corresponding to the operation risk prediction data in a slave management controller, so that the slave management controller controls the rotation speed of a fan around at least one computing device according to the fan control strategy.

[0103] In one embodiment, the device operation information of at least one computing device in the corresponding cabinet node is obtained through the slave management controller in any cabinet node, including: obtaining the core temperature, memory temperature and power consumption data of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node to obtain the device operation information.

[0104] In one embodiment, operation risk prediction data is obtained based on device operation information, including: normalizing core temperature, memory temperature and power consumption data; calculating comprehensive indicator data based on the normalized core temperature, normalized memory temperature and normalized power consumption data; and obtaining operation risk prediction data based on the comprehensive indicator data.

[0105] In one embodiment, operation risk prediction data is obtained based on comprehensive indicator data, including: if the rising change rate of the comprehensive indicator data and the previous comprehensive indicator data is greater than a preset indicator threshold, it is determined that the operation risk prediction data belongs to high-risk data, and the fan control strategy corresponding to the operation risk prediction data is determined to be a rapid cooling strategy.

[0106] In one embodiment, if the rising rate of change of the normalized core temperature and the previous core temperature is greater than the preset core temperature threshold, the rising rate of change of the normalized memory temperature and the previous memory temperature is greater than the preset memory temperature threshold, or the rising rate of change of the normalized power consumption data and the previous power consumption data is greater than the preset power consumption data threshold, then it is determined that the operation risk prediction data is high-risk data, and the fan control strategy corresponding to the operation risk prediction data is determined to be a rapid cooling strategy.

[0107] In one embodiment, the speed of fans around at least one computing device is controlled from a management controller according to a fan regulation strategy, including: determining identification information of fans around at least one computing device from the management controller, and controlling the fans with the identification information to run at full speed according to a rapid cooling strategy.

[0108] In one embodiment, it also includes: the heat dissipation control module determines the corresponding task volume control strategy based on the operation risk prediction data, and sends the task volume control strategy to the operating system of the main control node, so that the operating system controls the business volume transmission speed reaching at least one computing device according to the task volume control strategy.

[0109] In one embodiment, the main control node includes a main management controller and several fans; accordingly, the heat dissipation control module obtains the device operation information in the main control node through the main management controller, and updates the fan control strategy in the main management controller according to the device operation information, so that the main management controller controls the speed of the fans in the main control node according to the updated fan control strategy.

[0110] In one embodiment, the master control node includes a processor, a master management controller and a target port; the master management controller is connected to a slave management controller in any cabinet node; at least one cabinet node includes a switching module; the switching module is connected to at least one computing device in the corresponding cabinet node; the processor is connected to the switching module in any cabinet node through the target port.

[0111] In one implementation, the master management controller is used to: when confirming that the slave management controller in any cabinet node is in an activated state and the cabinet node is in a powered-on state, control the master control node to power on and activate the master management controller.

[0112] In one embodiment, the master management controller is used to: when confirming that the slave management controller in any cabinet node is in an inactive state and the cabinet node is in a powered-off state, control the master control node to be powered-off and control the master management controller to be in an inactive state.

[0113] In one implementation, the master management controller is used to confirm the state of the slave management controller in any cabinet node and the state of the corresponding cabinet node by reading a register in the slave management controller.

[0114] In one implementation, the main management controller is used to: after the main control node is powered on, activate the heat dissipation control module through the operating system of the main control node.

[0115] In one embodiment, the main management controller is used to: after the main control node is powered on, load the configuration file stored in itself, and control the speed of the fan in the main control node according to the configuration file;

[0116] The slave management controller in any cabinet node is used to: after the corresponding cabinet node is powered on, load the configuration file stored in itself, and control the speed of the fan in the corresponding cabinet node according to the configuration file.

[0117] Among them, for more specific working processes of each module and unit in this embodiment, reference can be made to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.

[0118] It can be seen that this embodiment provides a heat dissipation control device, which can perform temperature risk prediction in real time according to the actual operation of the computing equipment, and perform precise heat dissipation control on the fans in the area where the corresponding computing equipment is located according to the corresponding prediction results, so that the main control node can predict the temperature and heat dissipation conditions of each cabinet node, thereby realizing heat dissipation early warning and control of each cabinet.

[0119] An electronic device provided in an embodiment of the present application is introduced below, and the electronic device described below can be cross-referenced with other embodiments described herein. The electronic device in this embodiment can be a main management controller, a slave management controller, and a heat dissipation control module.

[0120] See also Figure 5 As shown, the embodiment of the present application discloses an electronic device, including:

[0121] Memory 501, used for storing computer programs;

[0122] The processor 502 is used to execute the computer program to implement the method disclosed in any of the above embodiments.

[0123] Furthermore, the present application also provides an electronic device. The electronic device can be Figure 6 The server shown can also be Figure 7 Terminal shown. Figure 6 and Figure 7 All of them are structural diagrams of electronic devices according to an exemplary embodiment, and the contents in the diagrams cannot be regarded as any limitation on the scope of use of the present application.

[0124] Figure 6 A schematic diagram of the structure of a server provided in an embodiment of the present application. The server may specifically include: at least one processor, at least one memory, a power supply, a communication interface, an input / output interface, and a communication bus. The memory is used to store a computer program, and the computer program is loaded and executed by the processor to implement the relevant steps in the heat dissipation control disclosed in any of the aforementioned embodiments.

[0125] In this embodiment, the power supply is used to provide working voltage for each hardware device on the server; the communication interface can create a data transmission channel between the server and external devices, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited here; the input and output interface is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0126] In addition, the memory as a carrier for resource storage can be a read-only memory, random access memory, disk or CD, etc. The resources stored thereon include operating system, computer programs and data, etc. The storage method can be temporary storage or permanent storage.

[0127] The operating system is used to manage and control the hardware devices and computer programs on the server to realize the operation and processing of the data in the memory by the processor, and it can be Windows Server, Netware, Unix, Linux, etc. In addition to the computer program that can be used to complete the heat dissipation control method disclosed in any of the above embodiments, the computer program can also further include computer programs that can be used to complete other specific tasks. In addition to data such as application update information, the data can also include data such as application developer information.

[0128] Figure 7 A schematic diagram of the structure of a terminal provided in an embodiment of the present application, the terminal may specifically include but is not limited to a smart phone, a tablet computer, a laptop computer or a desktop computer, etc.

[0129] Generally, the terminal in this embodiment includes: a processor and a memory.

[0130] Among them, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0131] The memory may include one or more computer non-volatile storage media, which may be non-transitory. The memory may also include high-speed random access memory, and non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In this embodiment, the memory is at least used to store the following computer program, wherein, after the computer program is loaded and executed by the processor, it can implement the relevant steps in the heat dissipation control method performed by the terminal side disclosed in any of the aforementioned embodiments. In addition, the resources stored in the memory may also include an operating system and data, etc., and the storage method may be temporary storage or permanent storage. Among them, the operating system may include Windows, Unix, Linux, etc. The data may include, but is not limited to, update information of the application.

[0132] In some embodiments, the terminal may also include a display screen, an input and output interface, a communication interface, a sensor, a power supply, and a communication bus.

[0133] Those skilled in the art will understand that Figure 7 The structure shown in the figure does not constitute a limitation on the terminal, and may include more or fewer components than those shown in the figure.

[0134] A non-volatile storage medium provided in an embodiment of the present application is introduced below. The non-volatile storage medium described below can be cross-referenced with other embodiments described herein.

[0135] A non-volatile storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the heat dissipation control method disclosed in the above-mentioned embodiment. The non-volatile storage medium is a computer-readable non-volatile storage medium, which, as a carrier for storing resources, may be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon include an operating system, a computer program and data, etc., and the storage method may be temporary storage or permanent storage.

[0136] A computer program product provided in an embodiment of the present application is introduced below. The computer program product described below can be cross-referenced with other embodiments described in this document.

[0137] A computer program product includes a computer program / instruction, which implements the steps of the above-disclosed heat dissipation control method when executed by a processor.

[0138] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0139] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of non-volatile storage medium known in the art.

[0140] Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A heat dissipation control system, characterized in that: include: A master control node and at least one cabinet node; the master control node includes a heat dissipation control module; the at least one cabinet node includes a slave management controller, at least one computing device and a plurality of fans; Wherein, the heat dissipation control module is connected to the slave management controller network in any cabinet node; The heat dissipation control module is used to: obtain device operation information of at least one computing device in the corresponding cabinet node through a slave management controller in any cabinet node, obtain operation risk prediction data according to the device operation information, and configure the fan control strategy corresponding to the operation risk prediction data to the slave management controller; The slave management controller is used to control the rotation speed of the fans around the at least one computing device according to the fan regulation strategy.

2. The heat dissipation control system according to claim 1, characterized in that: The heat dissipation control module is used to: obtain the core temperature, memory temperature and power consumption data of at least one computing device in the corresponding cabinet node through the slave management controller in any cabinet node to obtain the device operation information; Normalizing the core temperature, the memory temperature, and the power consumption data; Comprehensive indicator data is calculated based on the normalized core temperature, the normalized memory temperature and the normalized power consumption data; and the operation risk prediction data is obtained according to the comprehensive indicator data.

3. The heat dissipation control system according to claim 2, characterized in that: The heat dissipation control module is used for: if the rising change rate of the comprehensive index data and the previous comprehensive index data is greater than the preset index threshold, the rising change rate of the normalized core temperature and the previous core temperature is greater than the preset core temperature threshold, the rising change rate of the normalized memory temperature and the previous memory temperature is greater than the preset memory temperature threshold, or the rising change rate of the normalized power consumption data and the previous power consumption data is greater than the preset power consumption data threshold, then it is determined that the operation risk prediction data is high-risk data, and the fan control strategy corresponding to the operation risk prediction data is determined to be a rapid cooling strategy.

4. The heat dissipation control system according to claim 3, characterized in that: The slave management controller is used to determine identification information of fans around the at least one computing device, and control the fans with the identification information to run at full speed according to the rapid cooling strategy.

5. The heat dissipation control system according to claim 2, characterized in that: The heat dissipation control module is used to determine the corresponding task volume control strategy based on the operation risk prediction data, and send the task volume control strategy to the operating system of the main control node so that the operating system controls the business volume transmission speed reaching the at least one computing device according to the task volume control strategy.

6. The heat dissipation control system according to claim 1, characterized in that: The master control node includes a master management controller and a plurality of fans; Accordingly, the heat dissipation control module is used to: obtain the device operation information in the main control node through the main management controller, and update the fan control strategy in the main management controller according to the device operation information; The main management controller is used to control the rotation speed of the fan in the main control node according to the updated fan control strategy.

7. The heat dissipation control system according to any one of claims 1 to 6, characterized in that: The master control node includes a processor, a master management controller and a target port; the master management controller is connected to a slave management controller in any cabinet node; The at least one cabinet node comprises a switching module; the switching module is connected to at least one computing device in the corresponding cabinet node; The processor is connected to a switch module in any cabinet node through the target port.

8. The heat dissipation control system according to claim 7, characterized in that: The master management controller is used to control the master control node to start up and activate the master management controller when confirming that the slave management controller in any cabinet node is in an activated state and the cabinet node is in a powered-on state.

9. The heat dissipation control system according to claim 8, characterized in that: The master management controller is used to control the master control node to be shut down and the master management controller to be in an inactive state when confirming that the slave management controller in any cabinet node is in an inactive state and the cabinet node is in a shut down state.

10. The heat dissipation control system according to claim 9, characterized in that: The master management controller is used to confirm the state of the slave management controller in any cabinet node and the state of the corresponding cabinet node by reading the register in the slave management controller.

11. The heat dissipation control system according to claim 8, characterized in that: The main management controller is used to activate the heat dissipation control module through the operating system of the main control node after the main control node is powered on.

12. The heat dissipation control system according to claim 8, characterized in that: The main management controller is used to: after the main control node is powered on, load the configuration file stored in itself, and control the speed of the fan in the main control node according to the configuration file; The slave management controller in any cabinet node is used to: after the corresponding cabinet node is powered on, load the configuration file stored in itself, and control the speed of the fan in the corresponding cabinet node according to the configuration file.

13. A heat dissipation control method, characterized in that: Applied to a heat dissipation control module, the heat dissipation control module is arranged in a master control node, the master control node includes the heat dissipation control module; the heat dissipation control module is connected to a slave management controller network in at least one cabinet node; The at least one cabinet node comprises a slave management controller, at least one computing device and a plurality of fans; The method includes: Obtaining, through a slave management controller in any cabinet node, device operation information of at least one computing device in the corresponding cabinet node; Obtaining operation risk prediction data according to the equipment operation information; The fan regulation policy corresponding to the operation risk prediction data is configured in the slave management controller, so that the slave management controller controls the rotation speed of the fan around the at least one computing device according to the fan regulation policy.

14. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to execute the computer program to implement the method according to claim 13.

15. A non-volatile storage medium, characterized in that: Used to store a computer program, wherein the computer program implements the method according to claim 13 when executed by a processor.

16. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method of claim 13 is implemented.

Images