An adaptive performance scheduling system to improve the energy efficiency of thin and light laptops

Abstract

This invention belongs to the field of performance control technology. It discloses an adaptive performance scheduling system for improving the energy efficiency of thin and light laptops. The system includes: collecting and cleaning data to obtain laptop performance data; confirming power connection status and adjusting device cooling strategies; performing GPU idle detection and GPU load optimization; regulating fan operation status and generating fan noise reduction performance data; adjusting display status based on GPU operating mode and outputting display correction performance data; optimizing brightness control based on brightness adjustment records and outputting reasonable brightness control data; constructing a performance control log based on the reasonable brightness control data; and sending the performance control log to a preset database for storage and application. This improves the stability, energy efficiency, and user experience of the computer during operation.

Description

Technical Field

[0001] This invention relates to the field of performance control technology, and more specifically, to an adaptive performance scheduling system for improving the energy efficiency of thin and light laptops. Background Technology

[0002] With the rapid improvement of processor performance, graphics capabilities, and thermal design in thin and light laptops, their usage scenarios are expanding from traditional office work and web browsing to high-load applications such as gaming, video editing, AI computing, and multitasking, which have higher hardware requirements. At the same time, user behavior is becoming more fragmented and diversified. Therefore, establishing an adaptive performance scheduling system that can improve the energy efficiency of thin and light laptops has become an important direction. However, traditional performance control systems still have many pain points.

[0003] Traditional performance scheduling systems, when using PD power supply scenarios, typically rely on BIOS flags or driver signal synchronization data to determine power supply status. When the power supply path is unstable, this can easily lead to premature or delayed performance release, resulting in frame drops in high-frame-rate games or crashes due to sudden power consumption fluctuations. Furthermore, traditional performance scheduling systems often rely on task frame rates to determine if the output is normal when identifying and controlling GPU load. However, in reality, the graphics card may maintain high power consumption even when performing light tasks or not performing any tasks, without actively reducing its frequency, leading to significant energy waste. Additionally, traditional performance scheduling systems often perform cold starts with a fixed duty cycle for fan startup, ignoring fan response. The fan generates a high-pitched whistling noise during startup due to low inertia, which not only severely damages the user experience but also consumes more energy resources because it fails to start properly. In scenarios where the graphics card's operating mode is switched, if the display refresh rate is not synchronized with the discrete graphics card's context rendering state, the GPU may not warm up properly, resulting in insufficient frame output. This causes users to experience noticeable stuttering, screen flickering, or image tearing, preventing the discrete graphics card from performing optimally. Regarding brightness control, there is a lack of priority settings between energy-saving strategies and user behavior. Often, as soon as the user sets the brightness, the system's energy-saving strategy automatically adjusts to a lower brightness, resulting in a brightness rebound phenomenon. This creates a disconnect between the user's subjective intention and the actual experience, affecting the consistency of device interaction.

[0004] In view of this, the present invention proposes an adaptive performance scheduling system to improve the energy efficiency of thin and light laptops in order to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: an adaptive performance scheduling system for improving the energy efficiency of thin and light laptops, comprising:

[0006] The data acquisition module collects and cleans the laptop's operating data to obtain the laptop's performance data.

[0007] The power supply adjustment module confirms the power supply connection status based on the laptop's performance data and adjusts the device's heat dissipation strategy accordingly.

[0008] The load optimization module combines laptop performance data to perform GPU idle load detection, optimizes GPU load based on the GPU idle load detection results, and generates load control logs; it integrates device cooling strategies, load control logs, and laptop performance data to obtain load parameter performance data.

[0009] The fan control module regulates the fan's operating status based on load parameter performance data and generates fan noise reduction performance data.

[0010] The display correction module detects the graphics card's operating mode in real time, adjusts the display status based on the fan noise reduction performance data according to the graphics card's operating mode, and outputs display correction performance data.

[0011] The brightness control module acquires brightness adjustment records, optimizes brightness control based on display correction performance data using these records, and outputs reasonable brightness control data.

[0012] The log generation module constructs a performance control log based on reasonable brightness control data, and sends the performance control log to a preset database for storage and application; the modules are connected to each other via wired and / or wireless means.

[0013] Furthermore, the method for confirming the power supply connection status includes:

[0014] The sampling period is set to periodically sample the power supply voltage and current signal values ​​in the laptop performance data, and voltage period sequences and current period sequences are constructed respectively. A sliding window of preset size is constructed to traverse the voltage period sequence and current period sequence, and the voltage mean square error and current maximum deviation value in each sliding window of preset size are calculated. The voltage mean square error and current maximum deviation value are set as power supply feature pairs, and all power supply feature pairs are sorted to obtain a power supply feature pair sequence. The segments in the power supply feature pair sequence that show a continuous downward trend are extracted to obtain the downward trend segment. The average decline rate of each downward trend segment is calculated, and the average decline rate is matched with a preset stability threshold range to determine the stability level of the corresponding downward trend segment.

[0015] The system acquires the power supply protocol frame during the power handshake and identifies the power supply level field based on the power supply protocol frame to obtain the theoretical power supply power. It calculates the output power for each timestamp based on the values ​​of the voltage and current cycle sequences, and calculates the actual average output power for the sampling period based on the output power of all timestamps. It calculates the power ratio by comparing the actual average output power with the theoretical power supply power. It determines whether the power ratio is within a preset power deviation range; if so, it marks the power adaptation status of the corresponding sampling period as valid; otherwise, it is invalid. It combines the stability level and power adaptation status into a two-dimensional judgment condition, compares this two-dimensional judgment condition with the preset power supply level matching rules, and outputs the power supply connection status label for the corresponding sampling period.

[0016] Furthermore, the method for adjusting the device heat dissipation strategy includes:

[0017] The power supply connection status tag is matched with the historical temperature control strategy to obtain the thermal management parameters of the corresponding historical temperature control strategy, including the thermal sensitivity level setting value, temperature control tolerance threshold value, and heat dissipation response priority tag; the temperature control strategy of the sampling period corresponding to the current power supply connection status tag is obtained, and the thermal management parameters of the current temperature control strategy are adjusted based on the thermal management parameters of the matched historical temperature control strategy, while activating the corresponding cooling process; the adjusted temperature control strategy and the process of activating the cooling process are converted into an instruction sequence to obtain the device heat dissipation strategy.

[0018] Furthermore, the method for performing graphics card idle detection includes:

[0019] Extract GPU power consumption, GPU frequency, and memory frequency from continuous running cycles of laptop performance data, and construct GPU running vectors for each running cycle at a single timestamp. Sort the GPU running vectors to obtain the GPU running vector sequence for the corresponding running cycle. Calculate the deflection angle and magnitude difference between any two adjacent GPU running vectors in the GPU running vector sequence. Based on the deflection angle and magnitude difference, construct an activity evaluation function to calculate the power consumption activity score for the time period corresponding to the adjacent GPU running vectors. Select time periods with power consumption activity scores higher than a preset activity score threshold as stable active segments.

[0020] Obtain GPU scheduling task records for stable and active segments, and calculate the task density of each thread based on the GPU scheduling task records; calculate the GPU task utilization for the corresponding stable and active segments based on the task density; integrate the GPU task utilization for consecutive stable and active segments to obtain the GPU conflict accumulation value for the corresponding time segment; if the GPU conflict accumulation value is higher than the preset conflict deviation threshold, and the GPU task utilization is not higher than the preset utilization threshold, then the corresponding time segment is determined to be a non-low load idle segment.

[0021] Furthermore, the methods for optimizing GPU load include:

[0022] Extract GPU operating status parameters from non-low load idle periods as a set of load optimization target parameters; set a GPU load operating frequency threshold, and gradually decrease the GPU operating frequency in the load optimization target parameter set until the GPU load operating frequency threshold is reached; set a memory bandwidth threshold, and simultaneously reduce the memory frequency in the current load optimization target parameter set until the memory bandwidth threshold is reached; collect GPU power consumption change feedback data within the control period, convert the control process into a structured control record, and integrate the GPU power consumption change feedback data into a load control log.

[0023] Furthermore, the method for regulating the fan's operating status includes:

[0024] The fan speed is extracted from the load parameter performance data in real time. The time interval between the initial zero time stamp of each fan speed and the time stamp of the first increase in fan speed is identified to obtain the fan start time interval. The duty cycle value corresponding to the pulse control command received at each time stamp within the fan start time interval is extracted. Based on the duty cycle value, the pulse signal change curve of the corresponding fan start time interval is constructed. The duty cycle change rate between consecutive time stamps in the pulse signal change curve is calculated. If there is a time segment where the duty cycle change rate is continuously higher than the preset control rate threshold, and the time segment is located at the beginning of the corresponding fan start time interval, then the fan start time interval is determined to be a cold start time interval.

[0025] Based on historical fan start-up records, a fan start-up gradual increase signal variation curve is constructed. This curve is then divided into segments to obtain continuous time segments, with the pulse control value increasing within each segment. Simultaneously, the pulse control value increment between adjacent time segments does not exceed a preset gradual increase amplitude threshold. The pulse control values ​​of the continuous time segments are sequentially sent to the fan control interface to execute a gradual increase operation. Fan response data during the gradual increase process is collected in real time and noise reduction processing is performed to obtain denoised operation data. This denoised operation data is then integrated with load parameter performance data to obtain fan noise reduction performance data.

[0026] Furthermore, the noise reduction process includes:

[0027] The instantaneous fan speed is extracted based on the fan response data based on the preset data acquisition cycle, and the instantaneous fan speed is limited to below the preset upper limit of speed. The fan acoustic feature data is collected in real time and matched with the preset stable start decibel range. If a match is found, the current corresponding fan response data is output as noise-reduced operation data. Otherwise, the pulse control value increment during the gradual speed increase process is adjusted until the fan acoustic feature data during the gradual speed increase process meets the preset stable start decibel range condition.

[0028] Furthermore, the method for adjusting the display state includes:

[0029] The system identifies the graphics output paths of different tasks in the current fan noise reduction performance data and determines whether a change in the graphics output path causes the graphics card to switch to dedicated graphics card direct connection mode. If a graphics card operating mode switch occurs, the system obtains the maximum refresh rate of the dedicated graphics card. If this maximum refresh rate is higher than the refresh rate configured in the current computer display, the refresh rate is adjusted to match the current computer display configuration. A refresh rate safety switching window is constructed simultaneously, and it is determined whether the dedicated graphics card is in a stable output state within this window. If the graphics output context has not been fully loaded at this time, the refresh rate update operation is temporarily suspended. The system compares the adjusted fan start time period in the fan noise reduction performance data with the timestamp of the refresh rate change request. If the timestamp is within the adjusted fan start time period, the current refresh rate is locked and the refresh rate update operation is delayed. When the safety switching window ends, it determines whether the fan start is complete. If it is, the refresh rate change operation is executed, and the fan noise reduction data and the instruction sequence corresponding to the refresh rate change process are integrated into display correction performance data.

[0030] Furthermore, the method for optimizing brightness control includes:

[0031] The system identifies manually adjusted brightness records and extracts the target brightness value and corresponding timestamp. It then determines if there is a brightness adjustment value lower than the target brightness value at times adjacent to the manually adjusted record; if so, it identifies a power-saving strategy control conflict at that time. In power-saving strategy control conflict scenarios, it sets user-led priority and locks the manually set brightness value as the target brightness control value for the corresponding scenario. Simultaneously, it suspends the execution time of the power-saving strategy, presets a brightness adjustment window after the timestamp of manual adjustment completion, and sets the power-saving strategy execution time until the end of the preset brightness adjustment window. Finally, it integrates the control command sequence corresponding to the brightness adjustment process with display correction performance data into reasonable brightness adjustment data.

[0032] Furthermore, the method for constructing the performance tuning log includes:

[0033] Extract the device heat dissipation strategy and load control log from the reasonable brightness control data, and obtain the parameter change process of fan operation status control and display status adjustment in the reasonable brightness control data. At the same time, integrate the control command sequence corresponding to the brightness control process to form a performance control log.

[0034] The technical effects and advantages of the adaptive performance scheduling system for improving the energy efficiency of thin and light laptops according to the present invention are as follows:

[0035] By constructing a multi-dimensional module encompassing power supply status recognition, performance load adjustment, fan control, display refresh rate correction, and reasonable brightness adjustment, this system achieves real-time perception and adaptive performance regulation of thin and light laptops under varying operating environments. Compared to existing experience, it determines the stability of the power supply path by analyzing voltage and current changes and integrating protocol information, ensuring that the device maintains controllable resource configuration before the power supply is determined, avoiding sudden risks such as frame drops or thermal crashes caused by high-frequency releases when the power supply is unstable. By judging whether the GPU utilization is reasonable, it adjusts the GPU load to ensure that GPU resources are used effectively, while improving energy consumption control. By adjusting the fan start speed and combining acoustic data feedback, it avoids unreasonable cold starts of fans and sudden auditory interference to users. By setting the user's brightness adjustment action as the highest priority, it eliminates the user experience disconnect caused by the conflict between system energy-saving strategies and user preferences. In summary, this adaptive performance scheduling system for thin and light laptops improves the stability, energy efficiency, and user experience during computer operation, and has strong real-time controllability and scalability. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of an adaptive performance scheduling system for improving the energy efficiency of thin and light laptops according to the present invention;

[0037] Figure 2 This is a schematic diagram of an adaptive performance scheduling method for improving the energy efficiency of thin and light laptops according to the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Example 1

[0040] Please see Figure 1As shown in this embodiment, an adaptive performance scheduling system for improving the energy efficiency of thin and light laptops includes:

[0041] The data acquisition module collects and cleans the laptop's operating data to obtain the laptop's performance data.

[0042] The power supply adjustment module confirms the power supply connection status based on the laptop's performance data and adjusts the device's heat dissipation strategy accordingly.

[0043] The load optimization module combines laptop performance data to perform GPU idle load detection, optimizes GPU load based on the GPU idle load detection results, and generates load control logs; it integrates device cooling strategies, load control logs, and laptop performance data to obtain load parameter performance data.

[0044] The fan control module regulates the fan's operating status based on load parameter performance data and generates fan noise reduction performance data.

[0045] The display correction module detects the graphics card's operating mode in real time, adjusts the display status based on the fan noise reduction performance data according to the graphics card's operating mode, and outputs display correction performance data.

[0046] The brightness control module acquires brightness adjustment records, optimizes brightness control based on display correction performance data using these records, and outputs reasonable brightness control data.

[0047] The log generation module constructs a performance control log based on reasonable brightness control data, and sends the performance control log to a preset database for storage and application; the modules are connected to each other via wired and / or wireless means.

[0048] The laptop in this embodiment is a thin and light gaming laptop equipped with a relatively high-end CPU and a 50-series discrete graphics card. The laptop's operating data is collected by the computer's built-in sensors, including multi-dimensional parameters reflecting the laptop's operation process, such as temperature, voltage, current, resource utilization, fan speed, and detailed task process information. Data cleaning is performed by timestamp alignment and missing value filling to obtain higher quality laptop performance data.

[0049] Methods to confirm the power supply connection status include:

[0050] The sampling period is set to periodically sample the power supply voltage and current signal values ​​in the laptop performance data, and voltage period sequence and current period sequence are constructed respectively. The sampling period length is set based on historical sampling experience. The voltage and current signal values ​​in each sampling period are obtained and sorted according to the timestamp order in the period to obtain the voltage period sequence and current period sequence of that sampling period.

[0051] A preset-size sliding window is constructed to traverse the voltage cycle sequence and the current cycle sequence. The root mean square error of voltage and the maximum deviation of current are calculated in each preset-size sliding window. The window size of the preset-size sliding window is set based on historical experience. The voltage cycle sequence and the current cycle sequence belonging to the same sampling period are traversed synchronously using preset-size sliding windows of the same size. The root mean square error of voltage is used to reflect the stability level of power supply. The maximum deviation of current refers to the difference between the maximum current signal value and the minimum current signal value in a single preset-size sliding window, which is used to reflect the sudden change behavior of current.

[0052] The voltage mean square error and the maximum current deviation are set as power supply feature pairs. All power supply feature pairs are sorted to obtain a power supply feature pair sequence. The segment in the power supply feature pair sequence that shows a continuous downward trend is extracted to obtain the downward trend segment. Specifically, the voltage mean square error and the maximum current deviation in each preset-sized sliding window are combined into power supply feature pairs and sorted in chronological order to obtain the power supply feature pair sequence for the corresponding sampling period. The segment in the power supply feature pair sequence that consists of several adjacent power supply feature pairs with decreasing values ​​is extracted as the downward trend segment. It should be noted that the number of adjacent power supply feature pairs corresponding to "continuous" here is shorter than the total length of the power supply feature pair sequence.

[0053] The average rate of decline for each downward trend segment is calculated and matched with a preset stability threshold range to determine the stability level of the corresponding downward trend segment. Specifically, the average rate of decline is obtained by calculating the quotient of the difference between adjacent power supply characteristic pairs in each downward trend segment and the quotient per unit time, and then summing all the quotients in the downward trend segment and taking the average. The average rate of decline is then matched with a stability threshold range set based on historical judgment experience to output the stability level of the corresponding downward trend segment. For example, if the average rate of decline falls into a certain range, it is determined to be "rapidly stabilizing".

[0054] The power supply protocol frame during the power supply handshake is obtained, and the power supply level field is identified based on the power supply protocol frame to obtain the theoretical power supply power. The power supply handshake refers to the power negotiation and interaction process between the power supply device and the power consumption device in the power supply protocol. The power supply protocol frame refers to the power supply capability information transmitted during the power supply handshake process, including fields such as power level, voltage level and current limit. In this embodiment, the theoretical maximum power supply power of the corresponding laptop is determined by reading the corresponding power supply level field in the power supply protocol frame.

[0055] The output power of each time point is calculated based on the value of each time point in the voltage period sequence and the current period sequence. The actual average output power of the sampling period is calculated based on the output power of all time points. The output power is the product of voltage and current. The output power of each time point is obtained by multiplying the specific values ​​of voltage and current. The actual average output power is obtained by summing all the output powers belonging to the same sampling period and taking the average value.

[0056] The power ratio is obtained by calculating the ratio of the actual average output power to the theoretical power supply. The calculated power ratio is used as one of the characteristics to determine whether the output is stable.

[0057] Determine whether the power ratio is within the preset power deviation range. If it is, mark the power adaptation status of the corresponding sampling period as valid; otherwise, it is invalid. The preset power deviation range is set based on the computer configuration information. The power ratio is matched with the preset power deviation range. If the power ratio is within the preset power deviation range, the power adaptation status of the corresponding sampling period is determined to be valid.

[0058] The stability level and power adaptation status are combined into a two-dimensional judgment condition. This two-dimensional judgment condition is compared with the preset power supply level matching rule, and the power supply connection status label corresponding to the sampling period is output. The stability level and power adaptation status are used as two-dimensional judgment conditions. Only when the power adaptation status is valid can the stability level be matched with the power supply level matching rule set based on the computer configuration information, and the power supply connection status label corresponding to the sampling period of the stability level is output. In this embodiment, the power supply connection status label is used to reflect the specific situation of the power supply connection, such as label information such as "handshake in progress", "power supply stabilizing", and "power supply abnormal".

[0059] Methods for adjusting equipment heat dissipation strategies include:

[0060] Matching the power supply connection status tag with historical temperature control strategies yields the corresponding thermal management parameters, including the heat sensitivity level setpoint, temperature control tolerance threshold, and heat dissipation response priority tag. The historical temperature control strategy is a sequence of temperature control adjustment instructions set based on historical temperature control records, used to regulate relevant parameters to achieve temperature control. Matching the power supply connection status tag with the historical temperature control strategy yields three types of parameters for temperature regulation: the heat sensitivity level setpoint, the temperature control tolerance threshold, and the heat dissipation response priority tag. The heat sensitivity level setpoint is used to classify the device's sensitivity to changes in heat generation; the temperature control tolerance threshold is used to define the safe operating range of various components; and the heat dissipation response priority tag is used to identify the scheduling tendency of thermal management resources under the current operating conditions.

[0061] The system obtains the temperature control strategy corresponding to the sampling period of the current power connection status tag, and adjusts the thermal management parameters of the current temperature control strategy based on the thermal management parameters of the matching historical temperature control strategy. At the same time, it triggers the corresponding cooling process. Specifically, it obtains the parameters related to the temperature control strategy corresponding to the sampling period of the current power connection status tag in real time, and adjusts these parameters in the direction of the thermal management parameters of the historical temperature control strategy. It also triggers cooling processes such as increasing fan speed and switching air ducts.

[0062] The adjusted temperature control strategy and the process of triggering the cooling process are transformed into an instruction sequence to obtain the device heat dissipation strategy. The executable instruction sequence generated based on the adjusted temperature control strategy and the cooling process execution process is the device heat dissipation strategy, which is received and executed by the device terminal.

[0063] Methods for performing graphics card idle load testing include:

[0064] The GPU power consumption, GPU frequency, and memory frequency for consecutive running cycles are extracted from the laptop performance data. A GPU running vector for each time stamp in each running cycle is constructed. The GPU running vectors are sorted to obtain the GPU running vector sequence for the corresponding running cycle. The running cycle is divided based on the different task processes executed by the laptop. The time to complete any task process is one running cycle. The GPU power consumption, GPU frequency, and memory frequency for adjacent consecutive running cycles are obtained. The three types of parameters corresponding to each time stamp are used to construct a GPU running vector for the corresponding time stamp. The GPU running vectors in the same running cycle are sorted to obtain the GPU running vector sequence for that running cycle.

[0065] Calculate the deflection angle and magnitude difference between any two adjacent GPU running vectors in the GPU running vector sequence. Based on the deflection angle and magnitude difference, construct an activity evaluation function to calculate the power consumption activity score for the corresponding time period of the adjacent GPU running vectors. The formula for calculating the activity evaluation function is as follows: ;in, Indicates a power consumption activity score; This represents an angle, which in this embodiment is taken as a value. ; This represents the deflection angle between any two adjacent GPU running vectors in the GPU running vector sequence; It represents the difference in magnitude between any two adjacent GPU running vectors in the GPU running vector sequence; This represents the standard module length value set based on system configuration. and These represent the weights of the deflection angle and the difference in magnitude, respectively, and their values ​​in this embodiment range from [value range missing]. Furthermore, since the laptop in this embodiment is a thin and light gaming laptop with frequent application switching, it should be more inclined to consider the changes in activity caused by task switching. Therefore, it is set... .

[0066] The time periods in which the power consumption activity score is higher than the preset activity score threshold are selected as stable active periods. The preset activity score threshold is set based on historical activity judgment experience. The time periods consisting of consecutive timestamps in which the power consumption activity score is higher than the preset activity score threshold are determined as stable active periods, which represent time periods of continuous activity.

[0067] Obtain GPU scheduling task records for stable and active segments, and calculate the task density of each thread based on the GPU scheduling task records. The task density refers to the ratio of the number of times each thread is triggered to the total number of times all threads are executed. This value is used to reflect the thread's call frequency and continuous activity level.

[0068] The GPU task utilization is calculated based on the task density of the corresponding stable and active segment. A density threshold is set based on historical GPU scheduling task records. The number of threads with task density higher than the density threshold is counted and divided with the total number of threads. The result is the GPU task utilization. The smaller the value, the more likely that most GPU logic units are in standby mode, while the power consumption may still be high but the response is lacking.

[0069] The GPU task utilization of a continuous, stable, and active period is integrated to obtain the GPU conflict accumulation value for the corresponding time period. If the GPU conflict accumulation value is higher than a preset conflict deviation threshold, and the GPU task utilization is not higher than a preset utilization threshold, then the corresponding time period is determined to be a non-low-load idle period. The GPU conflict accumulation value refers to the integral value of GPU task utilization in different continuous stable and active periods at the time level, which is used to reflect the accumulation of the mismatch between power consumption and task utilization. If the GPU conflict accumulation value is higher than a preset conflict deviation threshold set based on historical judgment experience, and the GPU task utilization is not higher than the preset utilization threshold, it indicates that the current situation is characterized by high power consumption but low task response. Therefore, the corresponding time period is determined to be a non-low-load idle period.

[0070] Methods for optimizing GPU load include:

[0071] Extract the GPU running status parameters from the non-low load idle period as the set of target parameters for load optimization. The GPU running status parameters include two types of parameters: GPU running frequency and memory frequency in the non-low load idle period. These parameters are used as the set of target parameters for load optimization.

[0072] A GPU load operating frequency threshold is set, and the GPU operating frequency in the load optimization target parameter set is gradually reduced until the GPU load operating frequency threshold is reached. The GPU load operating frequency threshold is set based on the configuration information of the graphics card. In this embodiment, the lowest stable operating frequency of the graphics card is used as the value of the GPU load operating frequency threshold. The GPU operating frequency in the non-low load idle segment is gradually reduced with a fixed step size until it is reduced to the GPU load operating frequency threshold.

[0073] Set the memory bandwidth threshold, and at the same time reduce the memory frequency in the current load optimization target parameter set until the memory bandwidth threshold is reached. The memory frequency during normal continuous operation is set based on the graphics card configuration information as the memory bandwidth threshold at this time. By gradually reducing the memory frequency until the memory bandwidth threshold is reached, the power consumption of invalid memory refresh behavior is reduced.

[0074] The system collects GPU power consumption change feedback data within the control period, transforms the control process into a structured control record, and integrates the GPU power consumption change feedback data into a load control log. The control period refers to the time interval for adjusting the set of target parameters for load optimization. GPU power consumption change feedback data is collected within this time interval to reflect the power consumption fluctuations caused by the control. The control process for adjusting the set of target parameters for load optimization is encoded into a structured control record, and the GPU power consumption change feedback data is integrated to form a load control log. This load control log records the changes in parameters of various dimensions during the control process and the power consumption changes as a result of the control.

[0075] Methods for regulating fan operation include:

[0076] The fan speed is extracted from the load parameter performance data in real time. The time period between the initial zero time stamp of the fan speed and the time stamp of the first increase in fan speed is identified to obtain the fan start time period. Each start time period starts from the time stamp of zero speed and ends at the time stamp after which the first increase in speed is detected. This is used to reflect the process of the fan from the stopped state to the start state, and serves as the data basis for subsequent operations.

[0077] Extract the duty cycle value corresponding to the pulse control command received at each timestamp during the fan start-up period. Based on the duty cycle value, construct the pulse signal change curve for the corresponding fan start-up period. In this embodiment, the fan is driven by PWM pulse signal control. The duty cycle value represents the proportion of the signal that remains at a high level per unit time. Construct the PWM curve for that stage based on the duty cycle value of each timestamp, which is the pulse signal change curve, to reflect the changing trend of the fan's electronic control logic.

[0078] The duty cycle change rate between consecutive timestamps in the pulse signal change curve is calculated. If there is a time segment where the duty cycle change rate is consistently higher than a preset control rate threshold, and this time segment is located at the beginning of the corresponding fan start-up time period, then the fan start-up time period is determined to be a cold start time period. The preset control rate threshold is set based on the relevant theory of PWM pulse signal control. The slope of the duty cycle value between consecutive timestamps is used as the duty cycle change rate. Time segments where the duty cycle change rate is consistently higher than the preset control rate threshold are selected. If this time segment is located at the beginning of the fan start-up time period, then the fan start-up time period is determined to be a cold start time period. The time length corresponding to the beginning of the fan start-up time period is set based on historical fan control records.

[0079] A fan start-up gradual rise signal variation curve is constructed based on historical fan start-up records. This curve is then divided into continuous time segments, with the pulse control value increasing within each segment. Simultaneously, the increase in pulse control value between adjacent time segments does not exceed a preset gradual rise amplitude threshold. Based on PWM variation data from multiple historical fan start-up records, an ideal fan start-up gradual rise signal variation curve is constructed and evenly divided into several continuous time segments. The segment length is adjusted to ensure equal time length within each segment, while maintaining an increasing pulse control value within each segment. This pulse control value represents the PWM duty cycle. It should be noted that the overall trend between adjacent time segments is also increasing, and the increase in pulse control value cannot exceed the preset gradual rise amplitude threshold set based on the PWM pulse signal control, ensuring a gradual fan start rather than a sudden cold start.

[0080] The pulse control values ​​of continuous time segments are sequentially sent to the fan control interface and a gradual speed increase operation is performed. The fan control interface performs the operation of gradually increasing the fan speed by receiving the pulse control values ​​of continuous time segments.

[0081] The fan response data during the gradual speed increase process is collected in real time and noise reduction processing is performed to obtain noise-reduced operation data. The noise-reduced operation data is then integrated with the load parameter performance data to obtain fan noise reduction performance data. The fan response data includes parameters such as fan speed and power consumption, which reflect the changes in the fan after adjustment.

[0082] Noise reduction methods include:

[0083] The instantaneous fan speed value is extracted based on the fan response data based on the preset data acquisition cycle, and the instantaneous fan speed value is limited to below the preset speed limit. The fan data acquisition cycle is set based on historical sampling experience, and the instantaneous fan speed value of each time stamp in each data acquisition cycle is obtained from the fan response data collected during the gradual speed increase process. The preset speed limit is set based on the computer configuration to keep the instantaneous fan speed at startup below the preset speed limit.

[0084] The system collects real-time acoustic characteristic data of the fan and matches it with a preset stable start-up decibel range. If a match is found, the corresponding fan response data is output as noise-reducing operation data. Otherwise, the system continues to adjust the pulse control value increment during the gradual speed-up process until the fan acoustic characteristic data during the gradual speed-up process meets the preset stable start-up decibel range condition. Here, the fan acoustic characteristic data refers to the instantaneous decibel value. Based on acoustic knowledge, a preset stable start-up decibel range is set. The fan acoustic characteristic data is matched with this preset stable start-up decibel range. If a match is found, the fan response data at this time is used as noise-reducing operation data. Otherwise, it is considered that the PWM pulse signal control process at this time is unstable. Therefore, by continuously adjusting the pulse control value increment, the system ensures that the fan acoustic characteristic data meets the preset stable start-up decibel range, thus ensuring that the fan can start stably and will not produce excessive noise.

[0085] The methods for adjusting the display status include:

[0086] The system identifies the graphics output paths of different tasks in the current fan noise reduction performance data and determines whether a change in the graphics output path causes the graphics card operating mode to switch to dedicated graphics direct connection mode. Specifically, by identifying the graphics processing scheduling information of the dedicated graphics card included in the current fan noise reduction performance data, and analyzing the graphics output path of each executed graphics task, it determines whether a graphics card operating mode switch occurs when the graphics output path changes due to task switching. In this embodiment, the laptop has three graphics card operating modes, including dedicated graphics direct connection, hybrid mode, and integrated graphics mode. Here, it only identifies whether there is a switch to dedicated graphics direct connection mode, because only dedicated graphics direct connection mode directly utilizes the performance of the dedicated graphics card, resulting in a significant increase in frame rate. The other two modes involve integrated graphics, and the frame rate will not change drastically.

[0087] If there is a graphics card operating mode switch, obtain the maximum refresh rate of the dedicated graphics card. If this maximum refresh rate is higher than the refresh rate of the current computer display configuration, adjust the refresh rate to be the same as the current computer display configuration. When the graphics card operating mode is switched to dedicated graphics card direct connection mode, the laptop will try to enable the highest supported refresh rate to increase the frame rate. However, if the current device's monitor or the current system's image processing configuration process still maintains a lower refresh rate, frame rate stuttering will occur. Therefore, adjust the refresh rate of the dedicated graphics card to be the same as the configuration to ensure that the graphics card and display configuration are synchronized.

[0088] A refresh rate safety switching window is constructed synchronously. It is determined whether the discrete graphics card is in a stable output state within the refresh rate safety switching window. If the graphics output context has not been fully loaded at this time, the refresh rate update operation is suspended. The size of the refresh rate safety switching window is set based on historical refresh rate control experience. Within the refresh rate safety switching window, the stability of the GPU's operating power is detected to determine whether the discrete graphics card is in a stable output state. If the graphics output context has not been fully loaded at this time, it is assumed that the graphics card cannot output enough frame content. Therefore, the refresh rate update needs to be suspended and the graphics output context configuration needs to be loaded. The graphics output context refers to a series of runtime resources that the discrete graphics card must build and initialize to complete the graphics output path and correctly output the rendered frames to the display when it takes over the graphics output task.

[0089] By comparing the adjusted fan start-up time period in the fan noise reduction performance data with the timestamp of the refresh rate change request, if the timestamp falls within the adjusted fan start-up time period, the current refresh rate is locked and the refresh rate update operation is delayed. Specifically, by comparing the time interval of the adjusted fan start-up time period with the timestamp of the refresh rate change request, if it is found that the refresh rate change request is within the fan's slow-start process, there may be a temperature control loading delay, which could easily interfere with the display control flow. Therefore, to avoid such hardware resource usage conflicts that cause abnormal refresh rate switching, the current refresh rate should be temporarily locked, and the refresh rate update should be performed after the fan slow-start process is completed, ensuring smooth communication between resources in different control domains.

[0090] When the safe switching window ends, it checks whether the fan startup is complete. If it is, the refresh rate change operation is performed. The fan noise reduction data and the instruction sequence corresponding to the refresh rate change process are integrated into display correction performance data. First, it checks whether the safe switching window has ended, and then it checks whether the fan startup is complete. The refresh rate change can only be performed if the safe switching window has ended and the fan startup is complete. If the safe switching window has ended but the fan startup is not complete, it is necessary to wait for the fan startup to complete before the refresh rate change can be performed. The original fan noise reduction data, the parameter changes during the refresh rate change process, and the application instructions are integrated into display correction performance data.

[0091] Methods for optimizing brightness control include:

[0092] The system identifies manually adjusted brightness records by the user, extracts the target brightness value and corresponding timestamp of each record, and collects these records. This is done by collecting brightness adjustment records and identifying those manually adjusted by the user. The method for identifying manually adjusted records is as follows: the system identifies brightness change events directly triggered by the user using interface controls; the final brightness value adjusted in the manually adjusted record is used as the target brightness value, and the timestamp corresponding to the adjustment time is used as the basis for subsequent data.

[0093] The system determines whether there is a situation where the brightness adjustment value is lower than the target brightness value at a time adjacent to the manually adjusted record. If so, it determines that there is a conflict in the energy-saving strategy control at the corresponding time. Specifically, it identifies whether there is a situation where the final brightness adjustment value is lower than the target brightness value at a time adjacent to the brightness control interface of the device after obtaining the target brightness value manually adjusted by the user. If such a situation exists where the actual adjusted brightness value is lower than the user's target brightness value, it indicates that there is a conflict in the energy-saving strategy control caused by the energy-saving strategy of the brightness control interface and the user's intention.

[0094] In scenarios where energy-saving strategy control conflicts occur, user-led priority is set, and manually set brightness values ​​are locked as the target brightness control values ​​for the corresponding scenarios. When an energy-saving strategy control conflict is confirmed, the user's manual adjustment is given the highest priority, and the target brightness value is locked as the only effective target brightness control value for the current scenario.

[0095] Simultaneously, the execution time of the energy-saving strategy is temporarily suspended. After the timestamp of manual adjustment completion, a preset brightness adjustment window is set, and the execution time of the energy-saving strategy is set until the end of the preset brightness adjustment window. Specifically, a preset brightness adjustment window is set after the timestamp corresponding to the brightness control target value is manually adjusted. The size of the preset brightness adjustment window is set based on historical human-computer conflict adjustment experience to ensure that the energy-saving strategy can only be executed again after a certain length of brightness adjustment window. This window serves as a coordination time interval between the energy-saving strategy and user perception, ensuring the user experience.

[0096] The control command sequence corresponding to the brightness adjustment process is integrated with the display correction performance data into reasonable brightness adjustment data.

[0097] The methods for constructing performance tuning logs include:

[0098] The device heat dissipation strategy and load control log are extracted from the reasonable brightness control data. The parameter change process of fan operation status control and display status adjustment in the reasonable brightness control data is obtained. At the same time, the control command sequence corresponding to the brightness control process is integrated to form a performance control log. The reasonable brightness control data includes the data or command sequence generated by each module in this embodiment. By storing the change process of these data and the command sequence in the data storage unit, and timestamping each parameter change and command sequence, a performance control log covering the entire process of control is formed. The log is sent to the preset database of the laptop for storage, so as to facilitate direct query and retrieval.

[0099] This embodiment achieves real-time perception and adaptive performance regulation of thin and light laptops under varying operating environments by constructing a multi-dimensional module that includes power supply status recognition, performance load adjustment, fan control, display refresh rate correction, and reasonable brightness adjustment. Compared with existing experience, it determines the stability of the power supply path by analyzing voltage and current changes and integrating protocol information, ensuring that the device maintains controllable resource configuration before the power supply is determined, avoiding sudden risks such as frame drops or thermal crashes caused by high-frequency release when the power supply is unstable. By judging whether the GPU utilization is reasonable, the GPU load is adjusted to ensure that GPU resources are used effectively, while improving the ability to control energy consumption. By adjusting the fan start speed and combining acoustic data feedback, unreasonable cold starts of fans and sudden auditory interference to users are avoided. By setting the user's brightness adjustment action as the highest priority, the user experience disconnect caused by the conflict between the system's energy-saving strategy and the user's wishes is eliminated. In summary, this adaptive performance scheduling system for thin and light laptops improves the stability, energy efficiency, and user experience during computer operation, and has strong real-time controllability and scalability.

[0100] Example 2

[0101] Please see Figure 2 As shown, for parts not described in detail in this embodiment, please refer to the description in Embodiment 1. An adaptive performance scheduling method for improving the energy efficiency of thin and light laptops is provided, including:

[0102] S1. Collect laptop operating data and perform data cleaning to obtain laptop performance data;

[0103] S2. Confirm the power supply connection status based on the laptop's performance data, and adjust the device's heat dissipation strategy based on the power supply connection status;

[0104] S3. Combine laptop performance data to perform GPU idle load detection, optimize GPU load based on GPU idle load detection results, and generate load control logs; integrate device cooling strategies, load control logs, and laptop performance data to obtain load parameter performance data;

[0105] S4. Adjust the fan operating status based on load parameter performance data and generate fan noise reduction performance data;

[0106] S5. Real-time detection of graphics card operating mode, adjustment of fan noise reduction performance data based on graphics card operating mode, and output of display correction performance data;

[0107] S6. Obtain brightness adjustment records, optimize brightness control based on display correction performance data based on brightness adjustment records, and output reasonable brightness control data;

[0108] S7. Construct a performance control log based on reasonable brightness control data, and send the performance control log to a preset database for storage and application.

[0109] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0110] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0111] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. An adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer, the system comprising: The application relates to a notebook computer performance data collection and adjustment method. The method comprises the following steps: A data collection module collects notebook computer running data and performs data cleaning to obtain notebook computer performance data; A power supply adjustment module confirms a power supply connection state based on the notebook computer performance data and adjusts a device cooling strategy based on the power supply connection state; A load optimization module performs a graphics card idle load detection in combination with the notebook computer performance data, performs GPU load optimization based on the graphics card idle load detection result, and generates a load control log; the device cooling strategy, the load control log and the notebook computer performance data are integrated to obtain load parameter performance data; A fan control module controls a fan running state based on the load parameter performance data and generates fan noise reduction performance data; A display correction module detects a graphics card running mode in real time, adjusts a display state of the fan noise reduction performance data based on the graphics card running mode, and outputs display correction performance data; A brightness control module obtains a brightness adjustment record, optimizes brightness control of the display correction performance data based on the brightness adjustment record, and outputs reasonable brightness control data; A log generation module constructs a performance control log based on the reasonable brightness control data, and sends the performance control log to a preset database for storage and application; 2. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 1, wherein The modules are connected through wired and / or wireless modes. The power supply connection state confirming mode comprises the following steps: A sampling period is set to periodically sample power voltage values and current signal values in the notebook computer performance data, and voltage period sequences and current period sequences are constructed; a preset size sliding window is constructed to traverse the voltage period sequences and the current period sequences, and voltage mean square deviations and current maximum deviation values in each preset size sliding window are calculated; the voltage mean square deviations and the current maximum deviation values are set as power supply feature pairs, all the power supply feature pairs are sorted to obtain a power supply feature pair sequence, a section with a continuous descending trend in the power supply feature pair sequence is extracted, and a descending trend section is obtained; an average descending rate of each descending trend section is calculated, and the average descending rate is matched with a preset stability threshold interval to determine a stability level of the corresponding descending trend section; 3. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 2, wherein A power supply protocol frame at a power supply handshake is obtained, a power supply level field is identified based on the power supply protocol frame, and a theoretical power supply power is obtained; the output power of each timestamp is calculated based on the values of the voltage period sequences and the current period sequences, and the actual average output power of the sampling period is calculated based on the output powers of all the timestamps; a power ratio is obtained by calculating the ratio of the actual average output power to the theoretical power supply power; whether the power ratio is in a preset power deviation interval is judged, if yes, the power adaptation state of the corresponding sampling period is marked as valid, otherwise, the power adaptation state is invalid; the stability level and the power adaptation state are combined into a two-dimensional judgment condition, the two-dimensional judgment condition is compared with a preset power supply level matching rule, and a power supply connection state label of the corresponding sampling period is output. The device cooling strategy adjusting mode comprises the following steps: The power supply connection state label is matched with the historical temperature control strategy, the thermal management parameters of the corresponding historical temperature control strategy are obtained, including the thermal sensitive level setting value, the temperature control tolerance threshold value and the heat dissipation response priority label; the temperature control strategy of the sampling period corresponding to the current power supply connection state label is obtained, and the thermal management parameters of the matched historical temperature control strategy are used as the basis to adjust the thermal management parameters of the current temperature control strategy, and the corresponding cooling process is triggered; the adjusted temperature control strategy and the process of triggering the cooling process are converted into an instruction sequence, and the device heat dissipation strategy is obtained.

4. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 3, wherein The method for detecting the idle state of the graphics card comprises the following steps: GPU running energy consumption, GPU running frequency and video memory frequency in a continuous running period are extracted from notebook performance data, and a graphics card running vector of each time stamp in each running period is constructed, the graphics card running vector is sorted to obtain a graphics card running vector sequence corresponding to the running period; a deflection angle and a module length difference value of any two adjacent graphics card running vectors in the graphics card running vector sequence are calculated, and an activity evaluation function is constructed based on the deflection angle and the module length difference value to calculate the power consumption activity score of the time period corresponding to the adjacent graphics card running vectors; time periods with power consumption activity scores higher than a preset activity score threshold are selected as stable active sections; GPU scheduling task records of the stable active sections are obtained, and the task density of each thread is calculated based on the GPU scheduling task records; the GPU task utilization degree of the corresponding stable active section is calculated based on the task density; the GPU task utilization degrees of the continuous stable active sections are integrated to obtain the GPU contradiction accumulation value of the corresponding time section, and if the GPU contradiction accumulation value is higher than a preset contradiction deviation threshold and the GPU task utilization degree is not higher than a preset utilization degree threshold, the corresponding time section is determined as a non-low-load idle section.

5. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 4, wherein The method for optimizing the GPU load comprises the following steps: GPU running state parameters of the non-low-load idle section are extracted as a load optimization target parameter set; a GPU load running frequency threshold is set, and the GPU running frequency in the load optimization target parameter set is gradually decreased until the GPU load running frequency threshold is reached; a video memory bandwidth threshold is set, and the video memory frequency in the current load optimization target parameter set is lowered until the video memory bandwidth threshold is reached; GPU power consumption change feedback data in the regulation period is collected, the regulation process is converted into a structured regulation record, and the GPU power consumption change feedback data is encoded into a load regulation log.

6. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 5, wherein The method for regulating the fan running state comprises the following steps: The fan rotation speed in the real-time extracted load parameter performance data is identified, a time period between a time stamp at which the fan rotation speed is initially zero and a time stamp at which the fan rotation speed is first increased is identified, and a fan start-up time period is obtained; a duty cycle value corresponding to a pulse control instruction received at each time stamp in the fan start-up time period is extracted, and a pulse signal change curve corresponding to the fan start-up time period is constructed based on the duty cycle value; a duty cycle change rate between consecutive time stamps in the pulse signal change curve is calculated, and if there is a time sub-period in which the duty cycle change rate is continuously higher than a preset control rate threshold and the time sub-period is located at the beginning of the corresponding fan start-up time period, the fan start-up time period is determined as a cold start-up time period; A fan start-up slow-rising signal change curve is constructed based on historical fan start-up records, and the fan start-up slow-rising signal change curve is divided into continuous time segments, and the pulse control value in each time segment is increased, and the pulse control value increment between adjacent time segments is not greater than a preset slow-rising amplitude threshold; the pulse control values of the continuous time segments are sequentially sent to a fan control interface and a step-by-step speed-up operation is performed; fan response data of the step-by-step speed-up process is collected in real time and denoising processing is performed to obtain denoised running data, and the denoised running data and the load parameter performance data are integrated to obtain fan denoising performance data.

7. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 6, wherein The denoising processing includes: Fan instantaneous rotation speed values of the fan response data are extracted based on a preset data collection period, and the fan instantaneous rotation speed values are limited to be below a preset upper rotation speed limit; fan acoustic characteristic data are collected in real time, and the fan acoustic characteristic data are matched with a preset stable start-up decibel interval, if the matching is successful, the current corresponding fan response data is output as denoised running data, otherwise, the pulse control value increment in the step-by-step speed-up process is continuously adjusted until the fan acoustic characteristic data in the step-by-step speed-up process meet the preset stable start-up decibel interval condition.

8. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 7, wherein The display state adjustment includes: Different task graphic output paths in the current fan denoising performance data are identified, and whether the graphic output path transformation causes the graphic card running mode to be switched to the independent display direct connection mode is judged; if the graphic card running mode is switched, the maximum refresh rate of the independent graphic card is obtained, and if the maximum refresh rate is higher than the refresh rate of the current computer display configuration, the refresh rate is adjusted to be the same as the current computer display configuration; a refresh rate safe switching window is constructed synchronously, and whether the independent graphic card is in a stable output state in the refresh rate safe switching window is judged, and if the graphic output context is not loaded at this time, the refresh rate updating operation is temporarily suspended; the time stamp at which the adjusted fan start-up time period in the fan denoising performance data and the refresh rate change request are compared, and if the time stamp is in the adjusted fan start-up time period, the current refresh rate is locked and the refresh rate updating operation is delayed; whether the fan start-up is completed is judged when the safe switching window ends, and if the fan start-up is completed, the refresh rate changing operation is performed, and the fan denoising data and the instruction sequence corresponding to the refresh rate changing process are integrated into display correction performance data.

9. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 8, wherein The brightness regulation optimization includes: The manual control record in the brightness adjustment record manually adjusted by the user is identified, the target brightness value and the corresponding timestamp of the manual control record are extracted, it is judged whether there is a case that the brightness adjustment value is lower than the target brightness value at the time of the timestamp adjacent to the manual control record, if there is, it is determined that there is a conflict of energy-saving strategy control at the corresponding time, the user's priority is set in the conflict scene of the energy-saving strategy control, and the manually set brightness value is locked as the brightness control target value of the corresponding scene, the execution time of the energy-saving strategy is suspended at the same time, the preset brightness regulation window is set after the timestamp of the completion of the manual control, the execution time of the energy-saving strategy is set to the end of the preset brightness regulation window, and the control instruction sequence corresponding to the brightness regulation process and the display correction performance data are integrated into reasonable brightness regulation data.

10. The adaptive performance scheduling system for improving energy efficiency of a thin and light notebook computer according to claim 9, wherein, The construction performance regulation log mode comprises: The device heat dissipation strategy and the load regulation log in the reasonable brightness regulation data are extracted, the parameter change process of the fan running state regulation and the display state adjustment in the reasonable brightness regulation data is obtained, and the control instruction sequence corresponding to the brightness regulation process is integrated to form the performance regulation log.

Images