Method and system for dynamically adjusting heat dissipation design power and storage medium

Abstract

The invention discloses a method and a system for dynamically adjusting heat dissipation design power and a non-transitory computer readable storage medium. The system will dynamically adjust the turbine mode operating parameters of the central processor regardless of the relevant settings of the turbine operating state in which the central processor is located. In one embodiment, a controller sends a power drop signal to a basic input / output system, directly intervenes in and adjusts the real-time operating power of a central processing unit.

Description

Technical Field

[0001] The present invention relates to a technology for dynamically adjusting the thermal design power, and in particular to a method, a system and a non-temporary computer-readable storage medium for dynamically adjusting the thermal design power of a processor in a turbo state. Background Art

[0002] The prior art already has methods for dynamically adjusting the system performance of the CPU based on the actual user's usage conditions, thereby optimizing the performance of the original single wattage output. In common prior art, the system can predict the workload through artificial intelligence pre-training algorithms, and can increase higher power to accelerate the processing of workload when the system needs it. This can be referred to as a turbo (Turbo Boost) mechanism. However, when the turbo mechanism is operating, the system gives top priority to the overclocking operation of the CPU, and the temperature protection mechanism of other electronic components or circuits will not affect the operation of the turbo mechanism. In other words, even if the electronic component or circuit activates the temperature protection mechanism due to overheating, the CPU will continue to run the turbo mechanism autonomously without interruption.

[0003] Please also read Figure 1 , which shows the power, temperature and time relationship of the CPU in the turbo state; as shown in the figure, when the CPU operates in the turbo mechanism, the power will quickly increase to a power limit (about 135W), and the temperature of the electronic components on the motherboard will also rise rapidly and exceed 100℃. The temperature of the CPU will slowly drop until the turbo mechanism ends. However, since the temperature of the electronic components exceeds 100℃ during the turbo mechanism process and even after the turbo mechanism ends, the computer may crash or the electronic components may burn out, and in the worst case, the entire machine may burn out.

[0004] However, the most common method for system manufacturers to solve this problem is to sacrifice system performance in the turbo mode, such as reducing the power limit of the CPU turbo mechanism or adjusting related parameters (Tau parameters). For example, as an example, suppose a system has been configured with a (fixed) turbo mode set point, such as 20W, 4.0GHz and 20 seconds of Tau parameters; in order to solve the above problem, the system manufacturer may set the power limit lower, such as 18W. Alternatively, replace electronic components with higher temperature resistance to improve the temperature resistance of the system, but this will increase the cost. Summary of the invention

[0005] In view of this, embodiments of the present invention provide a method, system and non-transitory computer-readable storage medium for dynamically adjusting thermal design power, which can dynamically adjust the operating parameters of a central processing unit when the central processing unit is in turbo mode to improve the problems of the prior art.

[0006] The method for dynamically adjusting the thermal design power of an embodiment of the present invention is used to dynamically adjust the turbo mode operating parameters of a central processing unit. The method includes (but is not limited to): a controller sends a power adjustment signal to a basic input output system; and the basic input output system executes a thermal design power setting to adjust the real-time operating power of the central processing unit.

[0007] The system for dynamically adjusting the thermal design power of an embodiment of the present invention is used to dynamically adjust the turbo mode operating parameters of a central processing unit. The system includes (but is not limited to): a controller and a basic input-output system; the basic input-output system is electrically coupled to the controller and the central processing unit; the controller is configured to execute instructions to perform the operations contained in the aforementioned method.

[0008] A non-transitory computer-readable storage medium according to an embodiment of the present invention stores instructions, which are executed by a controller when a central processing unit is in a turbo mode, so that the controller executes the aforementioned method.

[0009] In summary, the method, system, and non-transitory computer-readable storage medium for dynamically adjusting the thermal design power according to some embodiments are intended to dynamically adjust the real-time operating power of the CPU when the CPU is in turbo operation to increase additional performance. In some embodiments, even if the CPU is in turbo operation, the system can still directly intervene and dynamically control and adjust the operating state of the CPU, thereby reducing the system temperature, thereby protecting the electronic components in the system, effectively solving the overheating problem, and maintaining the performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a graph showing the relationship between power, temperature and time when the CPU is in turbo operation;

[0011] Figure 2 is a system architecture diagram of an embodiment of the present invention;

[0012] Figure 3 A flowchart of an embodiment of the present invention;

[0013] Figure 4 A flowchart of another embodiment of the present invention;

[0014] Figure 5A A diagram showing the relationship between power, temperature and time when a CPU is operated in accordance with an embodiment of the present invention;

[0015] Figure 5B FIG. 1 is a diagram showing the relationship between power, temperature and time when the CPU is not involved in the operation of an embodiment of the present invention.

[0016] Explanation of symbols

[0017] 1: CPU

[0018] 2: Electronic components

[0019] 3: Controller

[0020] 4: Temperature sensor

[0021] 5: Chipset

[0022] 6: Power module

[0023] 7: Basic Input and Output System

[0024] 11: Model-specific buffer

[0025] 21: Clock pulse generation module

[0026] 22: Voltage modulation module

[0027] 31: Temperature reading unit

[0028] 32: Comparison unit

[0029] 33: Control unit DETAILED DESCRIPTION

[0030] The following will use the central processing unit (CPU) in turbo operation as some examples to illustrate, for example, the system manufacturer (OEM) uses the system software driver to dynamically adjust the CPU operation. It can predict the workload based on real-time usage conditions through artificial intelligence (AI) and machine learning (ML) algorithms to dynamically adjust system performance. In simpler terms, when the system needs it, the CPU is given a higher power (operating frequency) to speed up the processing of workloads, which is different from the power of normal operation, including but not limited to The company's Dynamic Tuning Technology (DTT).

[0031] The Thermal Design Power (TDP) mentioned in this article refers to the power consumption of the CPU under the maximum theoretical load. It is also the upper limit of the power consumption that the CPU must not exceed to prevent the cooling system from being overloaded and causing system overheating and damage. However, as mentioned in the previous paragraph, when the CPU is in turbo mode or certain workload types (such as During Advanced Vector Extensions, the real-time operating power of the CPU can exceed the maximum thermal design power to achieve better system performance. At the same time, it is possible that the CPU and other electronic components or circuits in the system may overheat. All embodiments of the present invention provide solutions to this problem.

[0032] In some embodiments, the CPU is in a turbo (Turbo Boost) operation state, i.e., a second power level is used as the upper limit of the operating power, such as 135W, which can be defined as PL2 (Power Limit), which is defined according to the specifications of different models of different manufacturers. On the other hand, the CPU is in a constant operation state, i.e., it operates at the first power level, such as 65W, which can also be called the average power of long-term operation, which can be defined as PL1, which is defined according to the specifications of different models of different manufacturers, and is usually set to not exceed the thermal design power (TDP) or directly set to the thermal design power (TDP).

[0033] The turbo mode operating parameters mentioned herein include the aforementioned second power level (PL2), clock rate, and Tau parameter (Turbo Time Parameter). The so-called Tau parameter refers to the duration of turbo mode operation, which is the time constant of the exponentially weighted moving-average (EWMA). As an example, assume that a system has been configured with (fixed) turbo mode operating parameters (e.g., at PL2) (defined by an OEM) with 135W, 5.2GHz, and Tau parameters of 20 seconds (sec). In some embodiments of the present invention, these parameters, such as the second power level (PL2), may be directly and dynamically adjusted in response to temperature parameters without being limited by the aforementioned operating parameters.

[0034] See also Figure 2 , which is a system architecture diagram of an embodiment of the present invention; this embodiment can be a desktop computer, a laptop computer, a tablet computer, a server, a workstation, a game console, an embedded system, and other electronic devices with a central processing unit. Figure 2 The embodiment shown in the figure includes a central processing unit 1, a plurality of electronic components 2, a controller 3, a temperature sensor 4, a chipset 5, a power module 6 and a basic input output system (BIOS) 7.

[0035] The central processing unit 1 can be a complex instruction set computer (CISC) microprocessor, a reduced instruction set computer (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, an x86 instruction set compatible processor, a processor implementing a combination of instruction sets, a multi-core processor such as a dual-core processor or a dual-core mobile processor, or any other microprocessor.

[0036] according to Figure 2 , the electronic component 2 includes a clock pulse generating module (Clock Generator Module) 21 and a voltage modulation module (Voltage Regulator Module, VRM) 22; in some embodiments, not shown in the drawings, the electronic component 2 also includes a PWM controller, a fan controller or other similar controllers, a hardware monitoring chip, other similar chips, active components, passive components and other temperature-sensitive circuits or electronic components.

[0037] The clock pulse generating module 21 provides operating frequencies for various components with different functions such as the CPU 1, memory, and bus. Furthermore, the voltage modulation module 22 is used to provide an operating voltage (Vcore) to the CPU 1 and can also adjust the operating voltage (Vcore) so that the CPU 1 can operate at a higher or lower operating frequency.

[0038] In some embodiments, the controller 3 may be a baseboard management controller, which may be an embedded controller (EC); in other embodiments, the controller 3 may also be a Super I / O, or other similar baseboard management controllers (Baseboard Manager Controller, BMC). In one embodiment, the controller 3 may include a temperature reading unit 31, a comparison unit 32, and a control unit 33. The temperature reading unit 31 is configured to measure the temperature of the electronic component 2 through the temperature sensor 4, that is, it can receive the signal transmitted by the temperature sensor 4 and interpret it as a temperature parameter. The comparison unit 32 is configured to compare the temperature measured by the temperature reading unit 31 with a pre-set critical temperature. The control unit 33 is configured to send a power reduction signal or a power recovery signal to the basic input and output system 7 in response to the result output by the comparison unit 32.

[0039] The temperature sensor 4 may be a chip temperature sensor, a thermistor or other similar temperature sensing device disposed on the computer motherboard, and its main function is to measure the temperature of a specific object (electronic component 2) or its surrounding components. In addition, the chipset 5 in some embodiments may be a south bridge chip, a north bridge chip, a system-on-chip (SoC), or other equivalent chipsets, which is mainly used for signal processing and transmission between the central processing unit 1 and other chips, circuits or I / O devices on the computer motherboard. In some embodiments, the chipset 5 is also one of the objects measured by the temperature sensor 4.

[0040] In one embodiment, the power module 6 may include but is not limited to a power supply and other power supply or power control related components, such as a power connector, a power timing module, a power phase, and a power management circuit, etc. In one embodiment, the basic input and output system 7 may be composed of a FLASH ROM or EEPROM, which can store program codes and setting values, and it provides some basic software functions and hardware initialization, such as hardware initialization during the startup phase, and firmware that provides runtime services for the operating system.

[0041] Please also read Figure 2 and Figure 3 , Figure 3 is a flow chart of an embodiment of the present invention; the following describes the operation process of some embodiments: first, when the CPU 1 is in a turbo operation state, that is, operating at or close to the second power level (PL2), the second power level (PL2) can be 135W or higher, and of course can be lower than 135W. At this time, when the temperature sensor 4 measures that the temperature of the electronic component 2 is equal to or greater than a critical temperature, such as 90°C, and lasts for a specific time, such as 3 seconds; the controller 3 uses a query scheduling and completion event (Q event) instruction to send a power adjustment (down) signal to the basic input and output system (BIOS) 7.

[0042] Furthermore, the basic input and output system 7 causes the central processing unit 1 to reduce the real-time operation power according to the first power level (PL1) through a memory mapped input and output (MMIO) write operation; for example, the power control unit (Power Control Unit) in the central processing unit 1 is driven to adjust the real-time operation power. In one embodiment of the above operation, the program code for defining the memory mapped input and output stored in the basic input and output system (BIOS) 7 is shown in the following Table 1:

[0043] Table 1

[0044]

[0045] In addition, in one embodiment of the aforementioned operation, the program code stored in the basic input and output system (BIOS) 7 to drive the power control unit of the central processor 1 to reduce the real-time operation power is shown in Table 2 below:

[0046] Table 2

[0047]

[0048] In one embodiment, the first power level (PL1) is 65W; however, in some embodiments of the present invention, it may be other higher or lower powers, which can be changed in the program code provided in Table 2.

[0049] It should be particularly noted that, in another preferred embodiment, when the basic input and output system (BIOS) 7 receives a power reduction signal, the basic input and output system 7 can reduce the real-time operating power of the central processing unit 1 to another operating power through a memory mapped input and output (MMIO) write operation, which can be higher than the first power level (PL1), such as 75W, to improve system performance, and this other operating power can also be changed in the program code provided in Table 2.

[0050] At the same time, the temperature sensor 4 continues to measure the temperature of the electronic component 2; and when the temperature of the electronic component 2 is lower than the critical temperature, and the central processing unit 1 operates according to the first power level (PL1) for a specific time (3 seconds); the controller 3 also uses the query scheduling and completion event (Q event) instruction to send a power adjustment (recovery) signal to the basic input and output system (BIOS) 7 to increase the real-time operating power of the central processing unit 1 to be close to or equal to the second power level (PL2), that is, to restore the central processing unit 1 to the preset power (135W) when it is in the turbo operation state.

[0051] In one embodiment of the above operation, the program code stored in the basic input and output system (BIOS) 7 to drive the power control unit of the central processing unit 1 to increase the real-time operating power is shown in Table 3 below:

[0052] Table 3

[0053]

[0054] In addition, in the aforementioned embodiment, the specific time is set to 3 seconds, but it is not limited thereto and can be shorter or longer. However, it should be noted that if the specific time is set too short, it may cause the central processing unit 1 to be frequently intervened in the operation, causing the protection mechanism to be too sensitive and making the system performance unstable. On the other hand, if the specific time is set too long, in addition to possibly affecting the operating performance of the central processing unit 1, it may also cause the temperature of the electronic component 2 to be too high due to insufficient response, making the present invention ineffective. Therefore, the setting value of the specific time can be adjusted according to the hardware specifications and performance requirements of the system. In addition, the critical temperature of 90°C set in some embodiments can also be set to other values, such as a higher or lower critical temperature.

[0055] Alternatively, in some embodiments, the BIOS 7 may also write the operation of reducing or increasing the real-time operating power via a model specified register (MSR) 11 in the CPU 1, and the model specified register (MSR) 11 can be seen at Figure 2 .

[0056] Please also read Figure 4 , which is a flow chart of another embodiment of the present invention; the main process of using the model specified register (MSR) 11 to write the operation is described as follows; when the controller 3 uses the query scheduling and completion event (Q event) instruction to send a power reduction signal to the basic input and output system 7, the basic input and output system 7 directly writes the model specified register (MSR) 11 in the central processor 1 to make the central processor 1 reduce the real-time operating power according to the first power level (PL1); similarly, when the controller 3 uses the query scheduling and completion event (Q event) instruction to send a power recovery signal to the basic input and output system 7, the basic input and output system 7 directly writes the model specified register (MSR) 11 in the central processor 1 to increase the real-time operating power of the central processor 1 to be close to or equal to the second power level (PL2). As for, Figure 4 The operation process of the CPU 1, the temperature sensor 4, and the controller 3 shown in the embodiment is similar to the above Figure 3 The embodiments shown are similar and will not be described in detail here.

[0057] See also Figure 5A , which is a diagram showing the relationship between power, temperature and time when the CPU 1 is operated according to an embodiment of the present invention. Figure 5A It can be seen that in one embodiment, when the CPU 1 is in the turbo operation state, the real-time operating power of the CPU 1 is maintained at 130W; at about the 5th second, the electronic component 2 (see Figure 2) exceeds 90°C. After a specific time (3 seconds), at about the 8th second, the real-time operating power of the CPU 1 is reduced to about 65W according to the first power level (PL1), and the temperature begins to decrease accordingly.

[0058] Then, at about 13 seconds, the electronic component 2 (see Figure 2 ) drops back to below 90°C; at this time, after a specific time (3 seconds), at about the 16th second, the real-time operating power of the CPU 1 is increased to about 130W according to the second power level (PL2); at this time, the temperature of the electronic component 2 begins to reverse and rise again. The subsequent operation is repeated in this mode until the turbo operation state of the CPU is cancelled.

[0059] See also Figure 5B , which is a graph showing the relationship between power, temperature and time when the CPU 1 is not involved in the operation of an embodiment of the present invention. As a comparative example, Figure 5B It shows that at the 55th second, CPU 1 enters turbo mode, while electronic component 2 (see Figure 2 ) began to rise, and soon broke through 90°C at the 71st second, and then rose to about 110°C and continued to maintain this high temperature.

[0060] Accordingly, by Figure 5A It can be seen that when the CPU 1 is in turbo operation, changing the system overclocking operation to a dynamic control adjustment mode can control the system temperature below 100°C, thereby protecting the electronic components 2 in the system (see Figure 2 ) or circuit, effectively solving the overheating problem while maintaining the system's performance.

[0061] For implementations involving firmware and / or software, the methods may be implemented by executing the functions or steps described herein. Any machine-readable medium that can implement instructions for the methods described herein may be used to implement the methods herein. For example, software program codes may be stored in a memory and executed by a processor. If implemented in firmware or software, the functions may be stored as one or more instructions or program codes on a computer-readable storage medium, which may include RAM, ROM, electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage, semiconductor memory or other storage device, or any other medium that can be used to store the desired program code in the form of instructions or data structures and can be accessed by a computer.

[0062] In summary, in order to provide over-temperature protection for all components on the motherboard, in an embodiment of the present invention, the Turbo Boost mechanism of the CPU 1 can be ignored, and the operating state of the CPU 1 can be forcibly intervened to reduce the real-time operating power of the CPU 1 without being interfered by the Turbo operating power upper limit (second power level) and Tau parameters, so as to reduce the temperature of the electronic components.

[0063] On the other hand, if the CPU 1 is still in the turbo operation state and the temperature of the electronic component 2 is lower than the critical temperature, the system can restore the CPU 1 to the turbo operation power (second power level) to further improve the system performance. Thus, the embodiment of the present invention can dynamically adjust the operating power of the CPU 1 in the turbo operation state while taking into account the system performance and stability, and can avoid the entire system crash or failure caused by high temperature of the electronic component, thereby achieving the requirements of maximizing system performance and reducing costs.

[0064] Although the present invention is disclosed in conjunction with the above embodiments, they are not intended to limit the present invention. Any ordinary technician in the technical field may make slight changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be based on the definition of the attached claims.

Claims

1. A method for dynamically adjusting thermal design power, which is used to dynamically adjust the turbo mode operating parameters of a central processing unit, the method comprising: The controller sends a power adjustment signal to a basic input output system (BIOS); and The basic input and output system executes the thermal design power setting to adjust the real-time operation power of the central processing unit.

2. The method for dynamically adjusting thermal design power according to claim 1, wherein: The thermal design power setting includes a first power level and a second power level; in response to the temperature of the electronic component being equal to or greater than a critical temperature, the central processing unit reduces the real-time operating power according to the first power level.

3. The method for dynamically adjusting thermal design power according to claim 2, wherein: In response to the temperature of the electronic component being lower than the critical temperature and the central processor operating according to the first power level for a specific period of time, the central processor increases the real-time operating power according to the second power level.

4. The method for dynamically adjusting thermal design power according to claim 2, wherein: The basic input and output system enables the central processing unit to reduce the real-time operation power according to the first power level through a memory mapped input and output (MMIO) write operation.

5. The method for dynamically adjusting thermal design power according to claim 2, wherein: The basic input and output system writes a model specified register (MSR) to enable the central processing unit to reduce the real-time operation power according to the first power level.

6. The method for dynamically adjusting thermal design power according to claim 1, wherein: The controller sends the power adjustment signal to the basic input and output system using query scheduling and completion event (Q event) instructions.

7. The method for dynamically adjusting thermal design power according to claim 1, wherein: The controller is an embedded controller (Embedded Controller), which is configured to obtain the temperature of the electronic component through a temperature sensor.

8. A system for dynamically adjusting thermal design power, which is used to dynamically adjust the turbo mode operating parameters of a central processing unit, the system comprising: Controller; as well as A basic input-output system, electrically coupled to the controller and the central processing unit; In response to the controller sending a power adjustment signal to the basic input output system, the basic input output system executes the thermal design power setting to adjust the real-time operation power of the central processing unit.

9. The system for dynamically adjusting thermal design power as described in claim 8 further comprises a temperature sensor electrically coupled to the controller; the controller measures the temperature of the electronic component through the temperature sensor; and in response to the temperature of the electronic component, the controller sends the power adjustment signal to the basic input and output system.

10. The system for dynamically adjusting thermal design power according to claim 9, wherein: The controller is an embedded controller, which includes a temperature reading unit, a comparison unit, and a control unit; the temperature reading unit is configured to measure the temperature of the electronic component through the temperature sensor; the comparison unit is configured to compare the critical temperature with the temperature of the electronic component; the control unit is configured to send the power adjustment signal to the basic input and output system.

11. The system for dynamically adjusting thermal design power according to claim 9, wherein: The thermal design power setting includes a first power level and a second power level; in response to the temperature of the electronic component being equal to or greater than a critical temperature, the central processing unit reduces the real-time operating power according to the first power level.

12. The system for dynamically adjusting thermal design power according to claim 11, wherein: In response to the temperature of the electronic component being lower than the critical temperature and the central processor operating at the first power level for a specific period of time, the central processor increases the real-time operating power according to the second power level.

13. The system for dynamically adjusting thermal design power according to claim 11, wherein: The basic input and output system enables the central processing unit to reduce the real-time operation power according to the first power level through a memory-mapped input and output write operation.

14. The system for dynamically adjusting thermal design power according to claim 11, wherein: The basic input and output system uses a model-specified register write operation to enable the central processing unit to reduce the real-time operation power according to the first power level.

15. A non-transitory computer-readable storage medium storing instructions, the instructions being executed by a controller when a central processing unit is in a turbo mode, causing the controller to perform operations comprising: A power adjustment signal is sent to a basic input output system; and the basic input output system executes a thermal design power setting to adjust the real-time operating power of the central processing unit.

16. The non-transitory computer-readable storage medium of claim 15, wherein: The thermal design power setting includes a first power level and a second power level; in response to the temperature of the electronic component being equal to or greater than a critical temperature, the central processing unit reduces the real-time operating power according to the first power level.

17. The non-transitory computer-readable storage medium of claim 16, wherein: In response to the temperature of the electronic component being lower than the critical temperature and the central processor operating at the first power level for a specific period of time, the central processor increases the real-time operating power according to the second power level.

18. The non-transitory computer-readable storage medium of claim 16, wherein: The basic input and output system enables the central processing unit to reduce the real-time operation power according to the first power level through a memory-mapped input and output write operation.

19. The non-transitory computer-readable storage medium of claim 16, wherein: The basic input and output system uses a model-specified register write operation to enable the central processing unit to reduce the real-time operation power according to the first power level.

20. The non-transitory computer-readable storage medium of claim 15, wherein: The controller is an embedded controller, which is configured to obtain the temperature of the electronic component through a temperature sensor; and the controller uses query scheduling and completion event instructions to send the power adjustment signal to the basic input and output system.

Images