CN115237244A - Control method, device, terminal device and storage medium for graphics processing unit - Google Patents

Abstract

The present application discloses a control method, device, terminal device and storage medium of a graphics processing unit, which belong to the technical field of terminal control. Applied to a terminal device, the method includes: obtaining the working cycle and working voltage of a graphics processing unit GPU; obtaining the power consumption of the GPU in the working cycle according to the working cycle and the working voltage; adjusting the working frequency of the GPU and the working voltage of the GPU, In order to make the power consumption of the GPU in the working cycle reach the minimum value, so that the power consumption of the GPU in the working cycle can reach the minimum value. The present application adjusts the working frequency of the GPU and the working voltage of the GPU based on the minimum value of the power consumption, so as to avoid the situation that the power consumption of the GPU still increases during the process of reducing the working frequency of the GPU, and improve the battery life of the terminal device. , reducing the waste of power on the terminal equipment.

Description

Control method, device, terminal device and storage medium for graphics processing unit

technical field

The present application relates to the technical field of terminal control, and in particular, to a control method, device, terminal device and storage medium for a graphics processing unit.

Background technique

With the development of information technology and Internet technology, it has become very common for users to use terminal devices in daily life. How to improve the battery life of terminal devices during use is a problem that people pay more attention to.

At present, most terminal devices use a dynamic voltage and frequency scaling (Dynamic voltage and frequency scaling, DVFS) mechanism to adjust the working frequency and working voltage of a graphics processing unit (Graphic Process Unit, GPU) when reducing power consumption. Among them, the DVFS mechanism is to adjust the working frequency of the GPU according to the current utilization of the GPU, so that the utilization of the GPU is maintained at a certain target value to meet the computing power requirements of the current graphics processing task.

For the above-mentioned existing DVFS mechanism, after the frequency is adjusted based on the power of the GPU, although the power of the GPU can be reduced, the power consumption of the GPU will still increase, and there is a problem of wasting the power of the terminal equipment.

SUMMARY OF THE INVENTION

In order to solve the problems in the prior art, improve the battery life of the terminal device, and reduce the waste of electricity by the terminal device, the embodiments of the present application provide a control method, device, terminal device and storage medium for a graphics processing unit. The technical solution is as follows:

In one aspect, the present application provides a method for controlling a graphics processing unit, which is applied to a terminal device, and the method includes:

Obtain the working cycle and working voltage of the graphics processing unit GPU;

According to the working cycle and the working voltage, obtain the power consumption of the GPU in the working cycle;

The working frequency of the GPU and the working voltage of the GPU are adjusted so that the power consumption of the GPU in the working cycle reaches a minimum value.

In one aspect, the present application provides an apparatus for controlling terminal equipment, which is applied to terminal equipment, and the apparatus includes:

a first obtaining module, configured to obtain the working cycle and working voltage of the graphics processing unit GPU;

a second obtaining module, configured to obtain the power consumption of the GPU in the working cycle according to the working cycle and the working voltage;

A parameter adjustment module, configured to adjust the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU in the working cycle reaches a minimum value.

In another aspect, the present application provides a terminal device, the terminal device includes a processor and a memory, the memory stores at least one instruction, at least a piece of program, code set or instruction set, the at least one instruction, all the The at least one piece of program, the code set or the instruction set is loaded and executed by the processor to implement the control method of the graphics processing unit according to one aspect.

In another aspect, the present application provides a computer-readable storage medium, wherein the storage medium stores at least one instruction, at least one piece of program, code set or instruction set, the at least one instruction, the at least one piece of program, all the The code set or instruction set is loaded and executed by the processor to implement the control method of the graphics processing unit according to one aspect.

On the other hand, an embodiment of the present application provides a computer program product, which, when the computer program product runs on a computer, causes the computer to execute the method for controlling a graphics processing unit according to the above-mentioned one aspect.

On the other hand, an embodiment of the present application provides an application publishing platform, the application publishing platform is used for publishing a computer program product, wherein, when the computer program product runs on a computer, the computer is made to execute the above one The control method of the graphics processing unit described in the aspect.

The beneficial effects brought by the technical solutions provided in the embodiments of the present application include at least:

The present application obtains the working cycle and working voltage of the graphics processing unit GPU; obtains the power consumption of the GPU during the working cycle according to the working cycle and working voltage; adjusts the working frequency of the GPU and the working voltage of the GPU, so that the GPU is in the working cycle The power consumption within the GPU reaches the minimum value, so that the power consumption of the GPU during the working cycle reaches the minimum value. The present application adjusts the working frequency of the GPU and the working voltage of the GPU based on the minimum value of the power consumption, so as to avoid the situation that the power consumption of the GPU still increases during the process of reducing the working frequency of the GPU, and improve the battery life of the terminal device. , reducing the waste of power on the terminal equipment.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative effort.

1 is a method flowchart of a method for controlling a graphics processing unit provided by an exemplary embodiment of the present application;

2 is a method flowchart of a method for controlling a graphics processing unit provided by an exemplary embodiment of the present application;

3 is a structural block diagram of a graphics processing unit involved in an exemplary embodiment of the present application;

FIG. 4 is a schematic structural diagram of an SOC system architecture involved in an exemplary embodiment of the present application;

5 is a method flowchart of a method for controlling a graphics processing unit provided by an exemplary embodiment of the present application;

6 is a structural block diagram of a control device of a graphics processing unit provided by an exemplary embodiment of the present application;

FIG. 7 is a schematic structural diagram of a terminal device provided by an exemplary embodiment of the present application.

Detailed ways

Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. Where the following description refers to the drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the illustrative examples below are not intended to represent all implementations consistent with this application. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present application as recited in the appended claims.

With the rapid development of science and technology, various terminal devices have been used in people's daily life. People need to use terminal devices in work, life and study. Take pictures, use terminal equipment to record data in work, etc. With more and more scenarios where people use terminal devices and the promotion and use of the Internet of Everything, it is more and more important to improve battery life in terminal devices.

At present, the main systems of terminal devices include Android system, iOS system, Linux system, etc. For terminal devices of various kernel systems, their respective central processing unit (Central Processing Unit, CPU), image processing unit GPU and other processors During the operation of the terminal device, it is mainly processed as a task processing unit, and accordingly, these task processing units consume the power of the terminal device during operation.

In practical applications, in order to improve the battery life of the terminal device and prolong the battery life of the mobile phone, the system-on-chip (SOC) design in the terminal adopts various schemes to reduce power consumption. Among them, the graphics processor in most terminal equipment will use the DVFS mechanism to adjust the working frequency and the working voltage, thereby reducing the power consumption level when the GPU is working. Among them, the DVFS mechanism is to adjust the working frequency of the GPU according to the current utilization of the GPU, so that the utilization of the GPU is maintained at a certain target value to meet the computing power requirements of the current graphics processing task.

That is, the frequency and voltage adjustment scheme of the above DVFS technology is to reduce the operating frequency of the graphics processor as much as possible to meet the processing requirements of the load task. The essence of this adjustment scheme is to adjust the operating frequency so that the GPU can run at the lowest power to meet the load task. processing needs. However, in real life, the usage scenarios of terminal equipment are limited by the battery capacity. Continuously reducing the working power of the GPU is equivalent to continuously reducing the working frequency of the GPU. When the working voltage of the GPU is reduced to the lowest level, if the working power continues to be reduced , on the contrary, it will lead to an increase in the power consumption of the GPU, and there is a problem of wasting the power of the terminal device.

In order to solve the problems existing in the above-mentioned related technologies, improve the battery life of the terminal device, and reduce the waste of power by the terminal device, the present application provides a control method for a graphics processing unit, which can control the working frequency and the working frequency based on the power consumption of the GPU. The voltage is adjusted, and after the operating frequency and operating voltage are adjusted, the power consumption of the GPU on the battery can be minimized.

Please refer to FIG. 1 , which shows a method flowchart of a method for controlling a graphics processing unit provided by an exemplary embodiment of the present application, and the method for controlling a graphics processing unit can be used in a terminal device. As shown in FIG. 1, the control method of the graphics processing unit may include the following steps:

Step 101: Obtain the working cycle and working voltage of the graphics processing unit GPU.

Optionally, the terminal device mentioned in this application may be a mobile phone, a tablet computer, a notebook computer, smart glasses, a smart watch, and an MP3 player (Moving Picture Experts Group Audio Layer III, moving picture expert compression standard audio layer 3), MP4 (Moving Picture Experts Group Audio Layer IV) players, desktop computers, laptop computers, smart home devices and other electronic devices with GPU.

During the running process of the GPU of the terminal device, various data processing tasks need to be processed by the GPU, for example, the data processing tasks may be rendering and displaying of images. In this solution, the terminal device can obtain the working cycle and working voltage of the GPU through the CPU, and can also obtain its own working cycle and working voltage through the GPU itself. This solution does not limit the manner of obtaining the working cycle and working voltage of the GPU.

Step 102: Obtain the power consumption of the GPU in the working cycle according to the working cycle and the working voltage.

Optionally, the terminal device may bring the obtained working cycle and working voltage of the GPU into the calculation formula of power consumption, so as to calculate the power consumption of the GPU in the working cycle. For example, the power consumption of the GPU of the terminal device can be written as the following function:

e=g(f, v, c, w);

In the above formula, f represents the operating frequency of the GPU, v represents the operating voltage of the GPU, c represents the number of computing units in the GPU, w represents the workload of the GPU, and e represents the power consumption of the GPU. Wherein, for the GPU of the terminal device, its current working cycle may be t, and the power consumption of the GPU in one working cycle t may be represented by e. c can be a default fixed number or an adjusted number of currently running arithmetic units.

Step 103: Adjust the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU during the working cycle reaches the minimum value, so that the power consumption of the GPU during the working cycle reaches the minimum value.

Optionally, in this solution, by adjusting the working frequency of the GPU and the working voltage of the GPU, the power consumption of the GPU during the working cycle can reach the minimum value, so that the power consumption of the GPU during the working cycle can reach the minimum value. . During the adjustment process, the terminal device can adjust the working frequency, and obtain the target voltage according to the adjusted working frequency and the voltage calculation formula. The target voltage is the working voltage corresponding to the adjusted working frequency. Among them, the voltage calculation formula is: v=k2*f. Among them, k2 is a constant, and k2 can be set by developers or operation and maintenance personnel according to the process parameters of the system-on-chip of the terminal device. Optionally, this step is equivalent to the process of obtaining the minimum value of the above e by adjusting f and v.

To sum up, the present application obtains the working cycle and working voltage of the graphics processing unit GPU; obtains the power consumption of the GPU during the working cycle according to the working cycle and working voltage; adjusts the working frequency of the GPU and the working voltage of the GPU to The power consumption of the GPU in the work cycle is minimized, so that the power consumption of the GPU in the work cycle is minimized. The present application adjusts the working frequency of the GPU and the working voltage of the GPU based on the minimum value of the power consumption, so as to avoid the situation that the power consumption of the GPU still increases during the process of reducing the working frequency of the GPU, and improve the battery life of the terminal device. , reducing the waste of power on the terminal equipment.

In a possible implementation manner, in this application, the number of computing units of the GPU may also be adjusted to improve the energy efficiency of the GPU of the terminal device.

Please refer to FIG. 2 , which shows a method flowchart of a method for controlling a graphics processing unit provided by an exemplary embodiment of the present application, and the method for controlling a graphics processing unit can be used in a terminal device. As shown in FIG. 2, the control method of the graphics processing unit may include the following steps:

Step 201, obtaining the working cycle and working voltage of the graphics processing unit GPU.

Optionally, t represents the work cycle, and the work cycle t of the GPU obtained by the terminal device can be represented by the following formula: t=k1*(w/f), where w represents the workload of the GPU, and f represents the current work of the GPU frequency. That is, the terminal device can calculate the current working cycle of the GPU according to the current working frequency of the GPU and the workload of the GPU. Among them, k1 is similar to the above k2, and is also set by developers or operation and maintenance personnel according to process parameters of the system-on-chip of the terminal device. In addition, the working voltage of the GPU can be represented by v, and the relationship between the working voltage of the GPU and the working frequency of the GPU (v=k2*f) can refer to the description in step 103 in FIG. 1 above, and will not be repeated here.

Step 202: Obtain the energy utilization rate of the top die and the sub-core energy utilization rate of the GPU, where the sub-core energy utilization rate is the average utilization rate of the energy utilization rate of each running computing unit in the GPU.

The energy utilization rate of the top mold of the GPU is the utilization rate of the battery energy by the top mold of the GPU. The energy utilization rate of each running computing unit in the GPU is the utilization rate of the battery energy by each running computing unit in the GPU.

Optionally, please refer to FIG. 3 , which shows a structural block diagram of a graphics processing unit involved in an exemplary embodiment of the present application. As shown in FIG. 3 , the graphics processing unit 300 includes a top model 301 and each arithmetic unit 302 . In the above-mentioned FIG. 3 , the graphics processing unit 300 can be divided into a top module 301 and each operation unit (core) 302, wherein the top module 301 can be responsible for memory access, geometric processing and other work in the terminal device, and each operation unit 302 may be responsible for Fragment processing and other work in the terminal device. The clock of the top model 301 is f _TOP , each operation unit 302 uses the same clock f _CORE , and the terminal equipment controls the clock of the top model 301 and the clock of each operation unit 302 independently of each other.

The GPU may include one top model and multiple computing units. In this solution, the terminal device can also obtain the energy utilization rate of the top mold and the energy utilization rate of the sub-cores. For example, the terminal device calculates the energy utilization rate of the top die 301 through the performance counter of the GPU to obtain αTOP, and calculates the utilization rate of each running computing unit 302 through the performance counter of the GPU. The average utilization rate of the operating computing unit 302 is obtained, thereby obtaining the average utilization rate αCORE of the energy utilization rate of each operating computing unit in the GPU.

Optionally, in this solution, the number of computing units of the GPU of the terminal device may be a fixed number by default, or may be flexibly adjusted. That is, if the number of operation units running in the GPU is fixed, the solution can skip step 202 and go directly to step 204; if the number of operation units running in the GPU is flexibly adjusted, step 203 can be entered. That is to say, if the number of operation units running in the GPU is a fixed number, in the above e=g(f, v, c, w), c is equivalent to a constant.

Step 203, according to the energy utilization rate of the top die and the energy utilization rate of the sub-core, adjust the number of each running computing unit in the GPU.

Optionally, if the number of computing units in the GPU can be individually controlled to turn on or off its working power, due to the different processing capabilities of different computing units and the different sub-workloads completed, a lot of cooperation is involved. In different situations, an optimal number of operation units can be obtained, so that the energy efficiency is the highest.

Optionally, the terminal device may adjust the number of running computing units in the GPU according to the energy utilization rate of the top mold and the energy utilization rate of the sub-cores as follows: according to the energy utilization rate of the top mold and the energy utilization rate of the sub-nuclei, Calculate the first proportion value; when the first proportion value is greater than the first threshold value, add a first preset number of operation units on the basis of the number of running operation units in the GPU; when the first proportion value is less than the second threshold value , on the basis of the number of each running operation unit in the GPU, the first preset number of operation units is reduced. Its code can be as follows:

Defineβ=k*α _CORE /α _TOP

Ifβ>β _U

Iecrease core number, N _core =N _core +Δ

Else _Ifβ <βD

Decrease top module frequency,N _core =N _core -Δ

Endif

Wherein, β represents a first proportional value, the first threshold is β _U , the second threshold is β _D , the first preset number is Δ, and N _core is the number of running operation units. k is a constant, which can be set by developers or operation and maintenance personnel according to the process parameters of the system-on-chip of the terminal device. That is, the terminal device brings the obtained energy utilization rate of the top mold and the energy utilization rate of the sub-nuclei into the above calculation formula of β, obtains the first proportional value β, and determines the size between β and the first threshold and the second threshold. relationship, when β>the first threshold, Δ computing units are added on the basis of the number of running computing units in the GPU. When β<the second threshold, Δ operation units are reduced based on the number of each running operation unit in the GPU.

For example, β _U can be 3, β _D can be 1/3, and Δ is 2. If N _core is 5 and the calculated β is 4, add 2 based on the number of running operation units. , then the adjusted number of running arithmetic units is 7. If N _core is 5 and the calculated β is 0.2, then the number of running computing units is reduced by 2, and the adjusted number of running computing units is 3. Optionally, when β is between the second threshold and the first threshold, the terminal device may keep the number of running computing units unchanged. That is, when the first proportional value is greater than the first threshold, it means that the energy utilization rate of the top model of the GPU is relatively high, and the number of computing units can be increased to enhance the processing capability of the GPU and improve the energy efficiency of the GPU. When the first proportional value is less than the second threshold, it means that the energy utilization rate of the top model of the GPU is low, and the number of computing units can be reduced to enhance the processing capability of the GPU and improve the energy efficiency of the GPU.

In a possible implementation manner, when reducing the number of computing units, the terminal device may follow the following manner: when the first proportional value is smaller than the second threshold, the terminal device may determine the priority of the tasks of each running computing unit in the GPU level; according to the order of task priority from low to high, each running operation unit in the GPU is sequentially reduced by a first preset number of operation units. That is, when the first proportional value is smaller than the second threshold, the terminal device may acquire the task priority of each running computing unit. For example, each running operation unit is operation unit 1, operation unit 2, operation unit 3, operation unit 4, operation unit 5. The tasks processed by these five operation units are task 1, task 2, task 3, and task. Four, task five, the tasks processed by these five computing units are arranged in order of task priority from low to high: task one, task four, task five, task three, task two, when the terminal device calculates the first proportion When the value is less than the second threshold, the task priority of each running computing unit is obtained, still taking the value of Δ as 2 as an example, when reducing the number of computing units, turning off computing unit 1 and computing unit 4 in the GPU.

Step 204: Acquire the power consumption of the GPU in the working cycle according to the working cycle, the working voltage and the number of each running computing unit in the GPU.

Optionally, the above e is still used to represent the power consumption of the GPU in the working cycle. The terminal device can obtain the working cycle, the working voltage and the number of running computing units in the GPU into the above formula. Repeat.

Step 205: Adjust the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU during the working cycle reaches the minimum value, so that the power consumption of the GPU during the working cycle reaches the minimum value.

Optionally, the power consumption of the GPU of the above-mentioned terminal device can be written as the following function: e=g(f, v, c, w) as an example, since f represents the working frequency of the GPU, and v represents the working voltage of the GPU, c represents the number of computing units in the GPU, w represents the workload of the GPU, and e represents the power consumption of the GPU. Since there is a linear mapping relationship between f and v in the DVFS mechanism: v=k2*f, e=g(f, v, c, w) can be simplified to e=g(f, c, w), this The goal of the program is to obtain

For example, the working cycle of the GPU obtained by the terminal device in the above manner is as follows:

Then, the working voltage of the GPU is v=k2*f, therefore, the dynamic power consumption of the GPU can be expressed according to the following formula:

P _dyn = k ₃ v ² f;

Then, the power consumption corresponding to the dynamic power consumption of the GPU is _Edyn =P _dyn t=k ₄ f ² w;

Optionally, the static power consumption of the GPU is also considered in this solution, and the static power consumption of the GPU can be expressed according to the following formula:

P _lkg = k ₅ v ² ;

Then, the power consumption corresponding to the static power consumption of the GPU is E _lkg =P _lkg t=k ₆ fw;

The total power consumption of the GPU is e=E _dyn +E _lkg ;

上面k3，k4，k5均为常数，也都可以由开发人员或者运维人员根据终端设备的系统级芯片的工艺参数设定。In this scheme, the DVFS mechanism is adjusted, and the corresponding energy consumption is the smallest when the lowest operating frequency that satisfies the workload is selected, that is, to obtain

The above k3, k4, and k5 are all constants, and they can also be set by developers or operation and maintenance personnel according to the process parameters of the system-level chip of the terminal device.

In the above-mentioned FIG. 3 , the GPU includes a top mode and an arithmetic unit. In this step, the terminal device can be adjusted as follows: adjusting the working frequency of the GPU top mode to the first frequency and adjusting the working frequency of the GPU top mode The voltage is adjusted to the first voltage, so that the power consumption of the top die of the GPU reaches the minimum value in the working cycle; to the second voltage, so that the power consumption of the computing unit of the GPU in the working cycle reaches the minimum value.

Optionally, in this solution, the working frequency of the top die of the GPU is adjusted to the first frequency and the working voltage of the top die of the GPU is adjusted to the first voltage, so as to make the power consumption of the top die of the GPU in the working cycle. The process of reaching the minimum value can be as follows:

The terminal device can compare the obtained energy utilization rate αTOP of the top mold with the fifth threshold, and when αTOP > the fifth threshold, increase the operating frequency of the top mold, and obtain the adjusted operating frequency according to v=k2*f The working voltage of the corresponding top mold is adjusted, so as to adjust the operating voltage of the top mold. Compare the energy utilization rate αTOP of the top mold obtained above with the sixth threshold, when αTOP < the sixth threshold, reduce the operating frequency of the top mold, and obtain the adjusted operating frequency according to v=k2*f. The operating voltage of the top mold is adjusted to adjust the operating voltage of the top mold. When the operating frequency of the top mold reaches the highest frequency that can be operated at the lowest operating voltage, the operating frequency of the top mold is controlled to remain unchanged. The corresponding code can be as follows:

Ifα_TOP>α_(TOP_U)

Increase top module frequency, f_TOP, ##Increase the working frequency of the top module

AdjustTop module workingvoltage; ##Adjust the working voltage of the top module

Else Ifα_TOP<α_(TOP_D)

Decrease top module frequency,f_TOP##Reduce the working frequency of the top module

Iff_TOP reaches f_(TOP_min),then maintain f_TOPat the minimumfrequency.##When the operating frequency of the top mold reaches the highest frequency that can work when the minimum operating voltage is reached, the operating frequency of the control top mold remains unchanged

AdjustTop module workingvoltage; ##Adjust the working voltage of the top module

Endif

Wherein, α_(TOP_U) is the fifth threshold, and α_(TOP_D) is the sixth threshold. Optionally, the fifth threshold and the sixth threshold may be preset in the terminal device by the developer. For example, the fifth threshold may be 80 %, the sixth threshold may be 30%.

Optionally, in this solution, the operating frequency of the computing unit of the GPU is adjusted to the second frequency and the operating voltage of the computing unit of the GPU is adjusted to the second voltage, so that the power consumption of the computing unit of the GPU in the working cycle is adjusted. The process of reaching the minimum value can be as follows:

Optionally, the terminal device may also compare the obtained sub-core energy utilization αCORE with the fifth threshold, and when αCORE > the seventh threshold, increase the operating frequency of the computing unit, and obtain the adjustment according to v=k2*f. The operating voltage of the operation unit corresponding to the later operating frequency is adjusted, so as to adjust the operation voltage of the operation unit. Compare the obtained sub-core energy utilization αCORE with the eighth threshold, when αCORE<the eighth threshold, reduce the operating frequency of the arithmetic unit, and obtain the arithmetic unit corresponding to the adjusted operating frequency according to v=k2*f The operating voltage of the arithmetic unit is adjusted, and when the operating frequency of the arithmetic unit reaches the highest frequency that can work when the minimum operating voltage is reached, the operating frequency of the arithmetic unit is controlled to remain unchanged. The corresponding code can be as follows:

Ifα_CORE>α_(CORE_U)

Increase core module frequency,f_CORE##Increase the operating frequency of the operation unit

Adjust core workingvoltage; ##Adjust the working voltage of the operation unit

Else Ifα_CORE<α_(CORE_D)

Decrease core module frequency,f_CORE##Reduce the operating frequency of the operation unit

Iff_CORE reaches f_(CORE_min), then maintain f_CORE at the minimumfrequency. When the operating frequency of the operation unit reaches the highest frequency that can work when the minimum operating voltage is reached, the operating frequency of the control operation unit remains unchanged

Adjustcoreworkingvoltage; ##Adjust the working voltage of the operation unit

Endif

Wherein, α_(CORE_U) is the seventh threshold, and α_(CORE_D) is the eighth threshold. Optionally, the seventh threshold and the eighth threshold can be preset in the terminal device by the developer. For example, the seventh threshold can be 80 %, the eighth threshold may be 30%.

In the above content, since the actual GPU is limited by the minimum working voltage, when the working voltage of the top die in the GPU reaches the minimum working voltage and the working voltage of the computing unit in the GPU reaches the minimum working voltage, the above v can also be seen. Do is a constant, so the power consumption corresponding to the dynamic power consumption of the GPU can be obtained by derivation as E _dyn =P _dyn t=k ₄ w;

The power consumption corresponding to the static power consumption of the GPU is

Therefore, this solution keeps the working frequency unchanged at the lowest working voltage to prevent the problem of increasing static power consumption caused by the constant reduction of the working frequency. When the power consumption of the GPU reaches the minimum value, when the working voltage of the GPU reaches the minimum working voltage, the working frequency of the GPU is controlled to remain unchanged.

In a possible implementation manner, the terminal device may further adjust the working frequency of the top die of the GPU to the first frequency and the working voltage of the top die of the GPU to the first voltage, so that the top die of the GPU is working After the power consumption in the cycle reaches the minimum value, obtain the working frequency of the adjusted GPU's top mode and the pre-adjustment GPU's top mode's operating frequency; When the ratio between the operating frequencies of the top mode of the GPU is greater than the third threshold, the operating frequency of the double-rate DDR storage system is increased; When the ratio between the two is less than the fourth threshold, the operating frequency of the DDR storage system is reduced; when the ratio between the operating frequency of the adjusted GPU's top mode and the pre-adjusted GPU's top mode operating frequency is between the fourth threshold and the third When the threshold value is within the range, the operating frequency of the DDR memory system is kept unchanged; wherein, the third threshold value is greater than the fourth threshold value.

Optionally, in this application, the GPU may be in the SOC system architecture, and the tasks processed by the GPU may also involve other subsystems in the SOC system to be completed together. When the load processed by the GPU changes, the frequencies of other subsystems also need to be adjusted. to improve the energy efficiency of the SOC. Subsystems related to GPU include Double Data Rate (DDR) storage system, CPU system and Display display system.

Please refer to FIG. 4 , which shows a schematic structural diagram of an SOC system architecture involved in an exemplary embodiment of the present application. As shown in FIG. 4 , the SOC architecture 400 includes a CPU 401 , a GPU 402 , a Display system 403 and a DDR storage system 404 . The terminal device can also adjust the working frequency of the top die of the GPU to the first frequency and the working voltage of the top die of the GPU to the first voltage, so that the power consumption of the top die of the GPU in the working cycle reaches the minimum value. After that, the operating frequency of the DDR memory system can be continuously adjusted to improve the energy efficiency of the SOC.

Optionally, the code for adjusting the operating frequency of the DDR storage system can be as follows:

If f_(TOP_curr)/f_(TOP_prev)>f_(TOP_ratio_u)

IncreaseDDRfrequency,f_(DDR_freq);##Increase the working frequency of the DDR storage system

Adjust DDR workingvoltage; ##Adjust the working voltage of the DDR storage system

Else if f_(TOP_curr)/f_(TOP_prev)<f_(TOP_ratio_d)

DecreaseDDRfrequency,f_(DDR_freq);##Decrease the operating frequency of the DDR memory system

Adjust DDR workingvoltage; ##Adjust the working voltage of the DDR storage system

Else

Maintain DDR frequency.##Maintain the operating frequency of the DDR storage system

Endif

Among them, f_(TOP_curr) is the operating frequency of the top mode of the GPU after adjustment, f_(TOP_prev) is the operating frequency of the top mode of the GPU before adjustment, f_(TOP_ratio_u) is the third threshold, and f_(TOP_ratio_d) is the fourth The threshold, optionally, the third threshold and the fourth threshold may be preset in the terminal device by the developer, for example, the third threshold may be 2, and the fourth threshold may be 1/2. Optionally, the operating frequency of the DDR memory system that is increased or decreased each time may be half of the current operating frequency of the DDR memory system.

It should be noted that this solution can also adjust the operating voltage and operating frequency of the display system and the CPU in the SOC architecture, thereby improving the energy efficiency of the SOC. The method can refer to the adjustment method for the DDR storage system, which is not repeated here. Repeat.

To sum up, the present application obtains the working cycle and working voltage of the graphics processing unit GPU; obtains the power consumption of the GPU during the working cycle according to the working cycle and working voltage; adjusts the working frequency of the GPU and the working voltage of the GPU to The power consumption of the GPU in the work cycle is minimized, so that the power consumption of the GPU in the work cycle is minimized. The present application adjusts the working frequency of the GPU and the working voltage of the GPU based on the minimum value of the power consumption, so as to avoid the situation that the power consumption of the GPU still increases during the process of reducing the working frequency of the GPU, and improve the battery life of the terminal device. , reducing the waste of power on the terminal equipment.

In addition, by adjusting the number of computing units in the GPU, the application can flexibly change the execution strength of the GPU on the load, and improve the energy efficiency of the GPU of the terminal device. The present application can also adjust the operating frequency and operating voltage of other subsystems on the SOC system architecture, thereby improving the energy efficiency of the SOC.

Taking the terminal device as a mobile phone as an example, when the mobile phone runs the game application program, the working voltage and operating frequency of the GPU are adjusted through the improved DVFS mechanism in this solution, thereby prolonging the duration of the mobile phone and improving the battery life.

Please refer to FIG. 5 , which shows a method flowchart of a method for controlling a graphics processing unit provided by an exemplary embodiment of the present application, and the method for controlling a graphics processing unit can be used in a terminal device. As shown in FIG. 5 , the control method of the graphics processing unit may include the following steps:

Step 501, check whether α_TOP is greater than α_(TOP_U).

If it is greater than, go to step 502, otherwise go to step 503.

Step 502 , increasing the operating frequency of the top mold, and adjusting the operating voltage of the top mold.

Step 503, check whether α_TOP is smaller than α_(TOP_D).

If less than, go to step 504, otherwise go to step 505.

In step 504, the operating frequency of the top mold is reduced, and the operating voltage of the top mold is adjusted.

Step 505, check whether α_CORE is greater than α_(CORE_U).

If it is greater than, go to step 506, otherwise go to step 507.

Step 506 , increase the operating frequency of the arithmetic unit, and adjust the operating voltage of the arithmetic unit.

Step 507, check whether α_CORE is less than α_(CORE_D).

If it is less than, go to step 508 , otherwise go to step 509 .

Step 508 , reduce the operating frequency of the arithmetic unit, and adjust the operating voltage of the arithmetic unit.

Step 509, calculate β.

Step 510, check whether β is greater than β _U .

If yes, go to step 511, otherwise go to step 512.

Step 511, increase the number of running operation units.

Step 512, check whether β is less than β _D .

If yes, go to step 513, otherwise keep the number of running operation units unchanged.

Step 513, reducing the number of running operation units.

Optionally, the above parameters have the same meanings as those in the embodiment of FIG. 2 , and details are not repeated here. It should be noted that the execution of steps 501 to 504 and the execution of steps 505 to 508 may be performed at the same time, or the sequence may be adjusted, which will not be repeated here.

Step 514: Obtain the power consumption of the GPU during the working cycle; adjust the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU during the working cycle reaches a minimum value.

To sum up, the present application obtains the working cycle and working voltage of the graphics processing unit GPU; obtains the power consumption of the GPU during the working cycle according to the working cycle and working voltage; adjusts the working frequency of the GPU and the working voltage of the GPU to The power consumption of the GPU in the work cycle is minimized, so that the power consumption of the GPU in the work cycle is minimized. The present application adjusts the working frequency of the GPU and the working voltage of the GPU based on the minimum value of the power consumption, so as to avoid the situation that the power consumption of the GPU still increases during the process of reducing the working frequency of the GPU, and improve the battery life of the terminal device. , reducing the waste of power on the terminal equipment.

The following are apparatus embodiments of the present application, which can be used to execute the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.

Please refer to FIG. 6 , which shows a structural block diagram of a control apparatus of a graphics processing unit provided by an exemplary embodiment of the present application. The control apparatus 600 of the graphics processing unit can be used in a terminal device to perform all or part of the steps performed by the terminal device in the method provided by the embodiment shown in FIG. 1 , FIG. 2 or FIG. 5 . The control device 600 of the graphics processing unit includes:

a first obtaining module 601, configured to obtain the working cycle and working voltage of the graphics processing unit GPU;

A second obtaining module 602, configured to obtain the power consumption of the GPU in the working cycle according to the working cycle and the working voltage;

The parameter adjustment module 603 is configured to adjust the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU in the working cycle reaches a minimum value.

To sum up, the present application obtains the working cycle and working voltage of the graphics processing unit GPU; obtains the power consumption of the GPU during the working cycle according to the working cycle and working voltage; adjusts the working frequency of the GPU and the working voltage of the GPU to The power consumption of the GPU in the work cycle is minimized, so that the power consumption of the GPU in the work cycle is minimized. The present application adjusts the working frequency of the GPU and the working voltage of the GPU based on the minimum value of the power consumption, which can avoid the situation that the power consumption of the GPU still increases during the process of reducing the working frequency of the GPU, and improves the battery life of the terminal device. , reducing the waste of power on the terminal equipment.

Optionally, the device further includes: a third acquisition module and a first adjustment module

The third obtaining module is configured to obtain the energy utilization rate of the top die of the GPU before obtaining the power consumption of the GPU in the working cycle according to the working cycle and the working voltage And the sub-core energy utilization rate, the sub-core energy utilization rate is the average utilization rate of the energy utilization rate of each running computing unit in the GPU;

The first adjustment module is configured to adjust the number of each running computing unit in the GPU according to the energy utilization rate of the top die and the sub-core energy utilization rate;

The second obtaining module is configured to obtain the power consumption of the GPU in the working cycle according to the working cycle, the working voltage and the number of each running operation unit in the GPU.

Optionally, the first adjustment module includes: a first calculation unit, a first increase unit and a first decrease unit;

the first calculation unit, configured to calculate a first proportional value according to the energy utilization rate of the top mold and the energy utilization rate of the sub-nuclei;

The first adding unit is configured to add a first preset number of computing units on the basis of the number of operating computing units in the GPU when the first proportional value is greater than the first threshold;

The first reducing unit is configured to reduce a first preset number of operation units on the basis of the number of each running operation unit in the GPU when the first proportional value is smaller than the second threshold.

Optionally, the first reduction unit includes: a first determination subunit and a first reduction subunit;

The first determination subunit is configured to determine the task priority of each running computing unit in the GPU when the first proportional value is less than the second threshold;

The first reduction sub-unit is configured to sequentially reduce each running operation unit in the GPU by a first preset number of operation units according to the order of the task priority from low to high.

Optionally, the parameter adjustment module includes: a second adjustment module and a third adjustment module,

The second adjustment module is used to adjust the working frequency of the top die of the GPU to the first frequency and adjust the working voltage of the top die of the GPU to the first voltage, so that the top die of the GPU is at The power consumption in the working cycle reaches a minimum value;

The third adjustment module is used to adjust the operating frequency of the computing unit of the GPU to the second frequency and adjust the operating voltage of the computing unit of the GPU to the second voltage, so that the computing unit of the GPU is The power consumption during the working cycle reaches a minimum value.

Optionally, the device further includes:

a fourth acquisition module, configured to adjust the operating frequency of the top mold of the GPU to the first frequency and the operating voltage of the top mold of the GPU to the first voltage, so that the top mold of the GPU After the power consumption in the working cycle reaches the minimum value, obtain the adjusted operating frequency of the GPU's top mold and the pre-adjusted operating frequency of the GPU's top mold;

A second adding module, configured to add a double-rate DDR storage system when the ratio between the adjusted operating frequency of the GPU's top mode and the pre-adjusted operating frequency of the GPU's top mode is greater than a third threshold operating frequency;

A second reducing module, configured to reduce the operating frequency of the DDR memory system when the ratio between the adjusted operating frequency of the GPU's top mold and the pre-adjusted operating frequency of the GPU's top mold is less than a fourth threshold working frequency;

The first holding module is used for when the ratio between the adjusted operating frequency of the GPU's top mold and the pre-adjusted operating frequency of the GPU's top mold is in the range between the fourth threshold and the third threshold when the operating frequency of the DDR storage system is kept unchanged; wherein the third threshold is greater than the fourth threshold.

Optionally, in the process of adjusting the working frequency of the GPU to the target frequency and adjusting the working voltage of the GPU to the target voltage, when the working voltage of the GPU reaches a minimum working voltage, control the The operating frequency of the GPU remains the same. .

FIG. 7 is a schematic structural diagram of a terminal device provided by an exemplary embodiment of the present application. As shown in FIG. 7 , the terminal device 700 includes a central processing unit (Central Processing Unit, CPU) 701 , a system memory including a random access memory (Random Access Memory, RAM) 702 and a read only memory (Read Only Memory, ROM) 703 704 , and a system bus 705 connecting the system memory 704 and the central processing unit 701 . The terminal device 700 also includes a basic transmission/output system (Input/Output System, I/O system) 706 that helps to transmit information between various devices in the computer, and is used to store the operating system 712, application programs 713 and other programs. Mass storage device 707 for module 714 .

The basic transmission/output system 706 includes a display 706 for displaying information and a transmission device 709 such as a mouse, keyboard, etc., for the user to transmit information. The display 706 and the transmission device 709 are both connected to the central processing unit 701 through a transmission output controller 710 connected to the system bus 705 . The basic transmission/output system 706 may also include a transmission output controller 710 for receiving and processing transmissions from various other devices such as a keyboard, mouse, or electronic stylus. Similarly, transmit output controller 710 also provides output to a display screen, printer, or other type of output device.

The mass storage device 707 is connected to the central processing unit 701 through a mass storage controller (not shown) connected to the system bus 705 . The mass storage device 707 and its associated computer-readable media provide non-volatile storage for the terminal device 700 . That is, the mass storage device 707 may include a computer-readable medium (not shown) such as a hard disk or a CD-ROM (Compact Disc Read-Only Memory) drive.

The computer-readable media may include computer storage media and communication media. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media include RAM, ROM, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other solid state Storage of its technology, CD-ROM, DVD (Digital Video Disc, high-density digital video disc) or other optical storage, cassettes, magnetic tape, disk storage or other magnetic storage devices. Of course, those skilled in the art know that the computer storage medium is not limited to the above-mentioned ones. The system memory 704 and the mass storage device 707 described above may be collectively referred to as memory.

The terminal device 700 can be connected to the Internet or other network devices through the network interface unit 711 connected to the system bus 705 . The memory also includes one or more programs stored in the memory.

Embodiments of the present application further provide a computer-readable medium, where the computer-readable medium stores at least one instruction, and the at least one instruction is loaded and executed by the processor to implement the graphics processing unit described in the above embodiments In the control method, all or part of the steps performed by the terminal device.

Embodiments of the present application further provide a computer program product, where at least one instruction is stored in the computer program product, and the at least one instruction is loaded and executed by the processor to implement the control of the graphics processing unit according to the above embodiments A method, all or part of the steps performed by a terminal device or a server.

It should be noted that: when the apparatus provided in the above-mentioned embodiment executes the control of the terminal equipment, only the division of the above-mentioned functional modules is used as an example for illustration. That is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process thereof is detailed in the method embodiments, which will not be repeated here.

The above-mentioned serial numbers of the embodiments of the present application are only for description, and do not represent the advantages or disadvantages of the embodiments.

Those of ordinary skill in the art can understand that all or part of the steps of implementing the above embodiments can be completed by hardware, or can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. The storage medium mentioned may be a read-only memory, a magnetic disk or an optical disk, etc.

The above descriptions are only optional embodiments of the present application, and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the protection of the present application. within the range.

Claims (10)

1. A control method for a graphics processing unit, characterized in that, applied to a terminal device, the method comprising: Obtain the working cycle and working voltage of the graphics processing unit GPU; According to the working cycle and the working voltage, obtain the power consumption of the GPU in the working cycle; The working frequency of the GPU and the working voltage of the GPU are adjusted so that the power consumption of the GPU in the working cycle reaches a minimum value. 2 . The method according to claim 1 , wherein before acquiring the power consumption of the GPU in the working cycle according to the working cycle and the working voltage, the method further comprises: 2 . Acquiring the energy utilization rate of the top die of the GPU and the energy utilization rate of the sub-core, where the energy utilization rate of the sub-core is the average utilization rate of the energy utilization rate of each running computing unit in the GPU; According to the energy utilization rate of the top die and the energy utilization rate of the sub-core, adjust the number of each running computing unit in the GPU; The obtaining the power consumption of the GPU in the working cycle according to the working cycle and the working voltage includes: According to the working cycle, the working voltage and the number of each running operation unit in the GPU, the power consumption of the GPU in the working cycle is obtained. 3. The method according to claim 2, wherein, according to the energy utilization rate of the top mold and the utilization rate of the arithmetic unit, adjusting the number of each running arithmetic unit in the GPU, comprising: Calculate the first proportional value according to the energy utilization rate of the top mold and the energy utilization rate of the sub-nuclei; When the first proportional value is greater than the first threshold, adding a first preset number of computing units on the basis of the number of running computing units in the GPU; When the first proportional value is smaller than the second threshold, a first preset number of operation units is reduced based on the number of each running operation unit in the GPU. 4. The method according to claim 3, wherein when the first proportional value is smaller than the second threshold, the first preset is reduced on the basis of the number of each running operation unit in the GPU Number of arithmetic units, including: When the first proportional value is less than the second threshold, determining the task priority of each running computing unit in the GPU; According to the order of the task priority from low to high, each running operation unit in the GPU is sequentially reduced by a first preset number of operation units. 5 . The method according to claim 2 , wherein the adjusting the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU in the working cycle reaches a minimum value. 6 . ,include: Adjusting the working frequency of the top die of the GPU to the first frequency and adjusting the working voltage of the top die of the GPU to the first voltage, so as to make the power consumption of the top die of the GPU in the working cycle reaches the minimum value; and, Adjusting the operating frequency of the computing unit of the GPU to the second frequency and adjusting the operating voltage of the computing unit of the GPU to the second voltage, so as to make the power consumption of the computing unit of the GPU in the working cycle reach the minimum value. 6 . The method according to claim 5 , wherein, in the adjustment of the operating frequency of the top die of the GPU to the first frequency and the adjustment of the operating voltage of the top die of the GPU to the first voltage, 6 . After the power consumption of the top die of the GPU in the working cycle reaches a minimum value, the method further includes: Acquiring the adjusted operating frequency of the GPU's top mode and the pre-adjusted operating frequency of the GPU's top mode; When the ratio between the adjusted operating frequency of the GPU's top mode and the pre-adjusted operating frequency of the GPU's top mode is greater than a third threshold, increasing the operating frequency of the double-rate DDR storage system; When the ratio between the adjusted operating frequency of the GPU's top mode and the pre-adjusted operating frequency of the GPU's top mode is less than a fourth threshold, reducing the operating frequency of the DDR storage system; When the ratio between the adjusted operating frequency of the GPU's top mode and the pre-adjusted operating frequency of the GPU's top mode is within the range of the fourth threshold and the third threshold, keep the The operating frequency of the DDR memory system remains unchanged; wherein, the third threshold is greater than the fourth threshold. 7. The method according to any one of claims 1 to 6, wherein in the process of adjusting the working frequency of the GPU to the target frequency and adjusting the working voltage of the GPU to the target voltage, when When the working voltage of the GPU reaches the minimum working voltage, the working frequency of the GPU is controlled to remain unchanged. 8. A control device for a graphics processing unit, characterized in that, when applied to a terminal device, the device comprises: a first obtaining module, configured to obtain the working cycle and working voltage of the graphics processing unit GPU; a second obtaining module, configured to obtain the power consumption of the GPU in the working cycle according to the working cycle and the working voltage; A parameter adjustment module, configured to adjust the working frequency of the GPU and the working voltage of the GPU, so that the power consumption of the GPU in the working cycle reaches a minimum value. 9. A terminal device, characterized in that the terminal device comprises a processor and a memory, and the memory stores at least one instruction, at least a section of program, code set or instruction set, the at least one instruction, the at least one A piece of program, the code set or the instruction set is loaded and executed by the processor to implement the control method of the graphics processing unit according to any one of claims 1 to 7. 10. A computer-readable storage medium, wherein the storage medium stores at least one instruction, at least one piece of program, code set or instruction set, the at least one instruction, the at least one piece of program, the code The set or instruction set is loaded and executed by the processor to implement the control method of the graphics processing unit according to any one of claims 1 to 7 .

Images