TW201527961A - System and method of detecting heat sink status of central processor - Google Patents

Abstract

A system of detecting a heat sink status of a central processor comprises: a hardware monitoring device for monitoring and obtaining a temperature signal of a temperature detector and a pulse width modulation signal of a fan of a computer; and a heat sink status detection module comprising a database for storing data obtained from the hardware monitoring device that is captured by the heat sink status detection module, operating frequency and usage frequency of the central processor and a temperature signal of the hardware device of the computer; and a temperature prediction model for receiving inputted data in the database and performing prediction based upon the data to generate a temperature prediction value corresponding to the heat sink of the central processor; wherein the heat sink status detection module compares the temperature prediction value with a captured temperature actual value so as to generate a temperature error value.

Description

System and method for detecting central processor heat sink status

The invention relates to the technical field of a computer, and in particular to a system and method for detecting the state of a central processor heat sink.

The heat sink is one of the important components of the computer. It can absorb the heat generated during the operation of the computer and then diverge into the chassis or outside the chassis to ensure the temperature of the computer and its components are normal. For example, the central processor will run during the process. The high temperature is generated and the central processor is operating normally through the central processor heat sink to lower the temperature of the central processor.

For the state of the central processor heatsink on the computer, the current software and hardware on the market only provide a detection mechanism for the fan on the central processor heatsink. For the central processor heatsink after a period of use, because of the dust cloth The heat dissipation performance of the central processor heatsink is greatly reduced. Compared with the normal clean central processor heatsink, the temperature difference between the two can reach 30 degrees, which may cause the cooling fan to burn out.

At present, most of the computer systems on the market are only for the central processor or system temperature detection, fan speed is too low or fault detection, etc., the mechanism for detecting the state of the central processor heatsink is not enough. . Therefore, after the computer system has been used for a long time, the user can only open the case and disassemble the heat sink to confirm whether the central processor heat sink needs to be cleaned. Opening the case and disassembling the heat sink is not difficult for a personal computer. Things, but if it's in a computer room with hundreds or thousands of servers, it will be a check that takes a lot of manpower and time.

In view of the above problems, the object of the present invention is to provide a system and method for detecting the state of a central processor heat sink, which uses a computational intelligent modeling tool to detect the state of the central processor heat sink so that the user can When the case is opened and the heat sink is removed, the state of the central processor heatsink can be correctly detected on one or several computer systems. In addition, an automatic abnormality elimination mechanism is provided to the central processor heatsink. When the status is abnormal, it actively alerts and eliminates abnormal conditions.

The first aspect of the present invention provides a system for detecting the state of a central processor heatsink, comprising: a hardware monitoring device that monitors all temperature detectors and fans of a computer to obtain all temperature detections. The temperature signal of the device and the pulse width modulation signal of all the fans; and a heat sink state detection module, comprising: a database obtained by the heat sink state detecting module capturing the hardware monitoring device The temperature signal of all the temperature detectors and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer are stored in the database; and a temperature Predicting the model, inputting the temperature signal of all the temperature detectors in the database and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signals of the hardware device of the computer to the Temperature prediction model, according to all The temperature signal of the temperature detector is predicted by the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer to generate a heat sink corresponding to the central processor. a central processor temperature prediction value; wherein the heat sink state detection module performs the central processor temperature prediction value and a central processor temperature actual value of the central processor heat sink in the captured temperature signal Compare to produce a central processor temperature error value.

A second aspect of the present invention provides a method for detecting a state of a central processor heatsink, comprising the steps of: monitoring all temperature detectors and fans of a computer to obtain temperature signals of all temperature detectors Pulse width modulation signal with all fans; captures the temperature signal of all temperature detectors and the pulse width modulation signal of all fans, the operating frequency and usage rate of the central processing unit, and the temperature of the hardware device of the computer Signal; and according to the temperature signal of all the temperature detectors and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit and the temperature signal of the hardware device of the computer, the output is corresponding to the output a central processor temperature predictor value of the central processor heat sink, and comparing the central processor temperature predicted value to a central processor temperature actual value of the central processor heat sink in the captured temperature signal to A central processor temperature error value is obtained.

12‧‧‧Radiator Status Detection Module

14‧‧‧ hardware monitoring device

16‧‧‧Warning device / radiator cleaning device

18‧‧‧Database

20‧‧‧ Temperature prediction model

22‧‧‧Fan

24‧‧‧Central Processing Unit

26‧‧‧ hardware devices

1 is a block diagram of a system for detecting a state of a central processor heat sink according to the present invention; 2 is a flow chart of a training temperature prediction model of the present invention; and FIG. 3 is a flow chart of a method for detecting a state of a central processor heat sink according to the present invention.

The present invention will be further understood by those of ordinary skill in the art to which the present invention pertains. .

1 is a block diagram of a system for detecting a state of a central processor heat sink of the present invention. In FIG. 1, the system for detecting the state of the central processor heat sink includes a heat sink state detecting module 12, a hardware monitoring device 14, and a warning device/heatsink cleaning device 16. The radiator state detection module 12 includes a database 18 and a temperature prediction model 20.

The heatsink state detecting module 12 can be implemented by software or hardware. The warning device/heatsink cleaning device 16 can use a user interface, a warning light or a buzzer, and a combination of a stepping motor and a brush. .

The hardware monitoring device 14 monitors all of the computer's temperature detectors (not shown) and the fan 22 to obtain the temperature signals of all the temperature detectors and the pulse width modulation signals of all the fans 22. Among them, all temperature detectors include a central processor temperature detector (not shown) that detects the temperature of the central processor heat sink (not shown) and detects the internal temperature of the computer. A detector (not shown), all of which include a central processing unit fan (not shown) that dissipates heat from the central processor heat sink and a system fan (not shown) that dissipates heat to the internal environment of the computer.

The heat sink state detecting module 12 captures the temperature signals of all the temperature detectors and the pulses of all the fans obtained by the hardware monitoring device 14 via the busbar I/O Access. The width modulation signal, the temperature signal of the hardware 26 of the computer, and the operating frequency and usage rate (ie, the usage load) from the central processing unit 24 are retrieved in the database 18 via the MSR (Model Specific Register) Access. The temperature signal of the hardware device 26 is detected by a temperature detector (not shown) that detects the hardware device 26. The hardware device 26 is a hard disk, a VGA card, or the like.

The temperature detection signal of all the temperature detectors obtained by the hardware monitoring device 14 and the pulse width modulation signals of all the fans 22, the operating frequency and usage rate of the central processing unit 24, and The temperature signal of the hardware device 26 is stored in the database 18.

The temperature signals of all the temperature detectors in the database 18 and the pulse width modulation signals of all the fans 22, the operating frequency and usage rate of the central processing unit 24, and the temperature signals of the hardware device 26 are input to the temperature prediction model 20, The temperature prediction model 20 predicts based on the temperature signals of all the temperature detectors and the pulse width modulation signals of all the fans 22, the operating frequency and usage rate of the central processing unit 24, and the temperature signals of the hardware device 26 to generate a corresponding center. One of the processor heat sinks is the central processor temperature prediction.

The heatsink state detection module 12 compares the CPU processor temperature prediction value with a central processor temperature actual value of one of the central processor heatsinks in the temperature signal captured to generate a central processor temperature error. value.

2 is a flow chart of a training temperature prediction model of the present invention. Please refer to the system components of Figure 1 for assistance in explaining the operation steps of Figure 2.

In FIG. 2, the heatsink state detecting module 12 needs to capture multiple temperature signals of all temperature detectors (for example, 1500 pens or more) and pulses of all the fans 22. The width modulation signal, the operating frequency and usage rate of the central processing unit 24, and the temperature signal of the hardware device 26 of the computer are used as training materials. The heat sink state detecting module 12 stores the plurality of captured data in the database 18 or forms a data file (step S30), and a part of the plurality of data (for example, 1000 data) is used as training. The temperature prediction model 20 uses the training data, and the remaining data (for example, 500 data) is used as verification data for verifying the temperature of the central processor heat sink by the temperature prediction model 20. The heat sink state detecting module 12 captures temperature signals from all the temperature detectors of the hardware monitoring device 14 and the temperature signals from the hardware device 26 via the busbar I/O Access, and via the bus bar MSR. Access captures the operating frequency and usage from the central processor 24.

The temperature signal of all the temperature detectors includes a temperature signal generated by a central processing unit temperature detector (located at the central processor heat sink) for detecting the temperature of the central processor heatsink, and detecting the computer. A temperature signal generated by an ambient temperature detector of the internal ambient temperature (provided inside the hardware monitoring device 14); a pulse width modulation signal of all the fans 22 includes a central processing unit fan that dissipates heat from the central processor heat sink The pulse width modulation signal and the pulse width modulation signal of the system fan for dissipating heat to the internal environment of the computer; the temperature signal of the hardware device 26 includes a temperature detector for detecting the hard disk (set at the hard disk) The temperature signal and the temperature signal generated by the temperature detector (set at the VGA card) of the VGA card.

Then, the heat sink state detecting module 12 determines the temperature signals of all the temperature detectors captured and the pulse width modulation signals of all the fans 22, the operating frequency and usage rate of the central processing unit 24, and the hardware device. Whether the number of temperature signals of 26 reaches a predetermined number (for example, 1500 pieces of data) (step S32).

If the heat sink state detecting module 12 determines that the captured data has not reached the preset number, the process returns to step S30, and the heatsink state detecting module 12 continues to retrieve the required data; The group 12 judges that the captured data has reached the preset number, and uses the database 18 or the data file as the temperature data of all the temperature detectors of the training data and the pulse width modulation signal of all the fans 22, and the central Each of the operating frequency and usage rate of the processor 24 and the temperature signal of the hardware device 26 is input to the temperature prediction model 20, and the pulse width modulation signal and central processing are performed according to the temperature signals of all the temperature detectors and all the fans 22. The operating frequency and usage rate of the device 24 and the temperature signal of the hardware device 26 are used to train the parameters of the temperature prediction model 20 (step S34). Among them, the training temperature prediction model 20 establishes a model by using a neural network, a fuzzy theory or a regression model.

Then, the temperature signals of all the temperature detectors in the database 18 or the data file as the training data and the verification data and the pulse width modulation signal of all the fans 22, the operating frequency of the central processing unit 24, and the use thereof. Each of the rate and temperature signals of the hardware device 26 is input to the temperature prediction model 20, and the temperature prediction operation of the temperature prediction model 20 is performed to output a plurality of central processor temperature verification values corresponding to the central processor heat sink (central processing unit temperature) The number of verification values is the same as the sum of the training data and the verification data. The heat sink state detection module 12 compares the plurality of central processor temperature verification values with the actual value of the central processor temperature of the plurality of central processor heat sinks (the actual value of the central processor temperature is the same as the central processing) Comparing the number of temperature verification values to obtain multiple CPU temperature verification error values (the number of CPU temperature verification error values is the same as the actual value of the CPU temperature) and multiple CPUs The temperature verification error value is averaged to obtain an average error value (step S36).

The radiator state detecting module 12 determines whether the average error value is greater than or equal to a training verification value (set by the user) to know whether the temperature prediction model 20 is completed (step S38).

If the heatsink state detection module 12 determines that the average error value is less than the training verification value, indicating that the temperature prediction model 20 predicts that the accuracy of the central processor temperature prediction value corresponding to the central processor heatsink does not meet the requirements, then returning to step S34, The temperature prediction model 20 is trained again by performing the operation of step S34.

If the radiator state detecting module 12 determines that the average error value is greater than or equal to the training verification value, indicating that the temperature prediction model 20 predicts that the accuracy of the central processor temperature prediction value corresponding to the central processor heatsink meets the requirements, the temperature prediction model ends. Training of 20 (step S40).

3 is a flow chart of a method for detecting a state of a central processor heat sink according to the present invention. Please refer to the system components of Figure 1 for assistance in explaining the operation steps of Figure 3.

The user can detect the shape of the central processor heat sink periodically or irregularly. State, and no need to detect at any time. In FIG. 3, all the temperature detectors and fans 22 of the computer are monitored by the hardware monitoring device 14 to obtain the temperature signals of all the temperature detectors and the pulse width modulation signals of all the fans 22 as described above (steps) S50).

The heat sink state detecting module 12 captures the temperature signals of all the temperature detectors and the pulse width modulation signals of all the fans 22, the operating frequency and usage rate of the central processing unit 24, and the temperature of the hardware device 26. The signals are stored in the database 18 (step S52).

The temperature signals of all the temperature detectors in the database 18 will be The pulse width modulation signal of the fan 22, the operating frequency and usage rate of the central processing unit 24, and the temperature signal of the hardware device 26 are input to the temperature prediction model 20, and the temperature prediction model 20 is based on the temperature signals of all the temperature detectors. The pulse width modulation signal of all the fans 22, the operating frequency and usage rate of the central processing unit 24, and the temperature signal of the hardware device 26 are subjected to temperature prediction operations to output a CPU processor temperature prediction value corresponding to the central processor heatsink. The heatsink state detection module 12 compares the CPU processor temperature prediction value with a central processor temperature actual value of one of the central processor heatsinks in the temperature signal captured to obtain a central processor temperature error value (step S54).

Then, the heat sink state detecting module 12 determines whether the temperature error value of the central processing unit is greater than or equal to a warning value (step S56), and if the heat sink state detecting module 12 determines that the temperature error value of the central processing unit is less than the warning value, Ending the detection of the state of the central processor heatsink; if the heatsink state detection module 12 determines that the temperature error value of the central processor is greater than or equal to the warning value, the heatsink state detection module 12 is connected via the busbar The I/O Access outputs a warning signal and a cleaning signal to the warning device/heatsink cleaning device 16.

When the warning device/radiator cleaning device 16 receives the warning signal and the cleaning signal, for example, the user interface, the warning light or the buzzer warning device/radiator cleaning device 16 generates an action such as a warning message, a light or a sound (step S58). And the warning device/heatsink cleaning device 16 such as a combination of a stepping motor and a brush cleans the dust on the central processor heat sink so that the central processor heat sink has a good heat dissipation effect on the central processing unit 24 (step S60) ).

The present invention provides a system and method for detecting the state of a central processor heat sink, which has the advantage of using a computational intelligent modeling tool to detect a central processor heat sink. The state allows the user to correctly detect the state of the central processor heatsink on one or several computer systems without opening the case and disassembling the heatsink. In addition, an automatic abnormality exclusion is provided. Mechanism to proactively alert and eliminate abnormal conditions when the state of the central processor's heat sink is abnormal.

The present invention has been described above with reference to the preferred embodiments and the accompanying drawings, and should not be considered as limiting. Various modifications, omissions and changes may be made without departing from the scope of the invention.

12‧‧‧Radiator Status Detection Module

14‧‧‧ hardware monitoring device

16‧‧‧Warning device / radiator cleaning device

18‧‧‧Database

20‧‧‧ Temperature prediction model

22‧‧‧Fan

24‧‧‧Central Processing Unit

26‧‧‧ hardware devices

Claims (14)

A system for detecting the state of a central processor heatsink includes: a hardware monitoring device that monitors all temperature detectors and fans of a computer to obtain temperature signals of all temperature detectors and pulses of all fans a width modulation signal; and a heat sink state detection module, comprising: a database, wherein the heat sink state detection module captures temperature signals of all temperature detectors obtained by the hardware monitoring device a pulse width modulation signal with all fans, an operating frequency and usage rate of the central processor, and a temperature signal of the hardware device of the computer are stored in the database; and a temperature prediction model, all in the database The temperature signal of the temperature detector and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer are input to the temperature prediction model, according to all temperature detections. The temperature signal of the detector and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature of the hardware device of the computer Predicting to generate a central processor temperature prediction value corresponding to one of the central processor heat sinks; wherein the heat sink state detection module determines the central processor temperature prediction value and the extracted temperature signal The central processor temperature actual value of one of the central processor heatsinks is compared to produce a central processor temperature error value. For example, in the system of claim 1, wherein all of the temperature detectors comprise a central processor temperature detector and detection for detecting the temperature of the central processor heat sink. One of the computer's internal ambient temperature ambient temperature detectors, all fans contain one central processor fan that dissipates heat to the central processor heatsink and one of the system fans that dissipate heat to the internal environment of the computer. The system of claim 1, wherein the heat sink state detecting module determines that the temperature error value of the central processing unit is greater than or equal to a warning value, and outputs a warning signal and a cleaning signal. The system of claim 3, further comprising a warning device that receives the warning signal output by the heat sink state detection module to generate an alert action. The system of claim 4, wherein the alerting device is one of a user interface, a warning light, and a buzzer. The system of claim 3, further comprising a heat sink cleaning device that receives the cleaning signal generated by the heat sink state detecting module to clean the central processor heat sink. The system of claim 6, wherein the radiator cleaning device comprises a stepping motor and a brush. For example, the system of claim 1 is characterized in that the temperature prediction model is modeled by one of a neural network, a fuzzy theory, and a regression model. A method for detecting a state of a central processor heat sink includes the steps of: monitoring all temperature detectors and fans of a computer to obtain temperature signals of all temperature detectors and pulse width modulation signals of all fans Capture the temperature signal of all temperature detectors and the pulse width modulation signal of all fans, the operating frequency and usage of the central processing unit, and the temperature of the hardware of the computer Degree signal; and based on the temperature signal of all the temperature detectors and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer, to output the pair A central processor temperature prediction value of one of the central processor heat sinks is to be compared, and the central processor temperature prediction value is compared to a central processor temperature actual value of the central processor heat sink in the captured temperature signal, A central processor temperature error value is obtained. For example, in the method of claim 9, after the step of obtaining the temperature error value of the central processing unit, further comprising determining that the temperature error value of the central processing unit is greater than or equal to a warning value, an alert signal and a clearing signal are output. The method of claim 10, wherein the warning signal is generated according to the warning signal. The method of claim 10, wherein the central processor heat sink is cleaned according to the cleaning signal. The method of claim 9, further comprising the steps of: capturing a temperature signal of all the temperature detectors of the plurality of pens and a pulse width modulation signal of all the fans, an operating frequency and usage rate of the central processing unit, and the The temperature signal of the hardware device of the computer is stored in the database or in the form of a data file; determining the temperature signals of all the temperature detectors and the pulse width modulation signals of all the fans, Whether the operating frequency and usage rate of the central processing unit and the number of temperature signals of the hardware device of the computer reach a preset amount; the temperature of all the temperature detectors in the database or part of the data file letter And the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer are input to the temperature prediction model, according to the temperature of some of the temperature detectors Training the temperature prediction model with the signal and the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer; and the temperature signals of all the remaining temperature detectors And the pulse width modulation signal of all the fans, the operating frequency and usage rate of the central processing unit, and the temperature signal of the hardware device of the computer are input to the temperature prediction model to output a plurality of central processing corresponding to the central processor heat sink The temperature verification value, the central processor temperature verification value is compared with a plurality of central processor temperature actual values of the central processor heatsink to obtain a plurality of central processor temperature verification error values, and the central The processor temperature verification error value is averaged to obtain an average error value; and the average error is determined Is greater than or equal to a training verification value, to know whether the temperature prediction model training is completed. For example, in the method of claim 13, wherein the temperature prediction model is trained to establish a model by using one of a neural network, a fuzzy theory, and a regression model.