US-12625532-B2 - Electronic device for predicting chip temperature and performing pre-operation, and operation method thereof

US12625532B2US 12625532 B2US12625532 B2US 12625532B2US-12625532-B2

Abstract

A method for operating an electronic device includes predicting a temperature rise of the electronic device when the application is started, predicting a temperature of the electronic device based on the predicted temperature rise and a current temperature of the electronic device, and lowering the temperature of the electronic device when the predicted temperature of the electronic device is higher than a preset threshold temperature.

Inventors

Eun Jae Ock

Assignees

SK Hynix Inc.

Dates

Publication Date: 20260512
Application Date: 20230427
Priority Date: 20221111

Claims (19)

1 . A method for operating an electronic device, the method comprising: predicting a temperature rise of the electronic device based on a trained artificial neural network when an application is started, the temperature rise being associated with executing the application; predicting a temperature of the electronic device based on the predicted temperature rise and a current temperature of the electronic device; and lowering the temperature of the electronic device when the predicted temperature of the electronic device is higher than a preset threshold temperature, wherein the predicting of the temperature rise of the electronic device based on the trained artificial neural network comprises: obtaining data sets, each of the data sets including a plurality of input parameters that include a measured temperature rise during a period in which the application is executed and the current temperature of the electronic device when the application is started; determining whether each of the plurality of input parameters has a correlation with the measured temperature rise; selecting one or more input parameters each determined to have the correlation with the measured temperature rise among the plurality of input parameters as each of input data sets; and training the artificial neural network through supervised learning based on the input data sets each including the selected input parameters and the measured temperature rise.
2 . The method of claim 1 , wherein the predicting of the temperature rise of the electronic device based on the trained artificial neural network further comprises: testing performance of the trained artificial neural network; and finishing training of the artificial neural network when the performance of the trained artificial neural network is greater than or equal to a first threshold.
3 . The method of claim 2 , wherein the plurality of input parameters further comprises a remaining battery level, a current time, a memory used by applications running in background, a usage time of the executed application, and whether or not charging is being performed.
4 . The method of claim 2 , wherein the obtaining of the data sets comprises obtaining more than a preset number of data sets, the preset number being sufficiently great to ensure normality of data.
5 . The method of claim 4 , wherein the selecting of the input parameters comprises: obtaining a correlation coefficient by performing a linear regression analysis on each of the plurality of input parameters and the measured temperature rise; determining that one or more input parameters each having the correlation coefficient greater than a threshold value have the correlation with the measured temperature rise; and selecting the input parameters determined as each having the correlation with the measured temperature rise to generate the input data sets of the artificial neural network.
6 . The method of claim 5 , wherein the training of the artificial neural network comprises training the artificial neural network using a portion of the input data sets, and wherein the testing of the performance of the trained artificial neural network comprises testing the performance of the trained artificial neural network using a remaining portion of the input data sets.
7 . The method of claim 5 , wherein the obtaining of the data sets, the determining of whether each of the plurality of input parameters has the correlation with the measured temperature rise, the selecting of the input parameters, and the training of the artificial neural network are repeatedly performed when the performance of the trained artificial neural network is less than the first threshold.
8 . The method of claim 7 , further comprising: determining whether a number of repetitions of the training of the artificial neural network is greater than a second threshold; and finishing the training of the artificial neural network when the number of repetitions of the training of the artificial neural network is equal to or greater than the second threshold and the performance of the trained artificial neural network is greater than or equal to a third threshold.
9 . The method of claim 8 , further comprising: discarding the artificial neural network for the application, when the number of repetitions of the training of the artificial neural network is equal to or greater than the second threshold and the performance of the trained artificial neural network is less than the third threshold.
10 . The method of claim 2 , wherein the lowering of the temperature of the electronic device comprises: reducing an operating frequency of a processor that controls the electronic device, or reducing an operating frequency of a memory included in the electronic device, or both.
11 . A memory controller comprising: a first interface configured to perform data communication with a first device; a second interface configured to generate a signal for controlling an operation of a second device; a temperature predictor configured to predict a temperature rise of an electronic device based on a trained artificial neural network when an application is started, the temperature rise being associated with executing the application, generate input data sets for the artificial neural network, and train the artificial neural network; and a processor configured to predict a temperature of the electronic device based on the temperature rise predicted by the temperature predictor and a current temperature of the electronic device, and perform an operation for lowering the temperature of the electronic device based on the predicted temperature of the electronic device, wherein the temperature predictor is configured to: obtain data sets, each of the data sets including a plurality of input parameters that include a measured temperature rise during a period in which the application is executed and the current temperature of the electronic device when the application is started; determine whether each of the plurality of input parameters has a correlation with the measured temperature rise; select one or more input parameters each determined to have the correlation with the measured temperature rise among the plurality of input parameters as each of the input data sets; and train the artificial neural network through supervised learning based on the input data sets each including the selected input parameters and the measured temperature rise.
12 . The memory controller of claim 11 , wherein the temperature predictor is configured to: test performance of the trained artificial neural network; and finish training of the artificial neural network when the performance of the trained artificial neural network is greater than or equal to a first threshold.
13 . The memory controller of claim 12 , wherein the temperature predictor is configured to: obtain more than a preset number of the data sets, the preset number being sufficiently great to ensure normality of data; obtain a correlation coefficient by performing a linear regression analysis on each of the plurality of input parameters and the measured temperature rise; determine that one or more input parameters each having the correlation coefficient greater than a threshold value have the correlation with the measured temperature rise; select the input parameters each determined as having the correlation with the measured temperature rise to generate the input data sets of the artificial neural network; train the artificial neural network using a portion of the input data sets; and test the performance of the trained artificial neural network using a remaining portion of the input data sets.
14 . The memory controller of claim 13 , wherein the temperature predictor is configured to: repeat the obtaining of the data sets, the determining of whether each of the plurality of input parameters has the correlation with the measured temperature rise, the selecting of the input parameters, and the training of the artificial neural network when the performance of the trained artificial neural network is less than the first threshold; finish the training of the artificial neural network when a number of repetitions of the training of the artificial neural network is equal to or greater than a second threshold, and the performance of the trained artificial neural network is greater than or equal to a third threshold; and discard the artificial neural network for the application, when the number of repetitions of the training of the artificial neural network is equal to or greater than the second threshold and the performance of the trained artificial neural network is less than the third threshold.
15 . The memory controller of claim 11 , wherein the processor is configured to reduce the temperature of the electronic device when the predicted temperature of the electronic device is higher than a threshold temperature, by performing one or more of reducing an operating frequency of the processor, reducing an operating frequency of the second device, and reducing a number of dies activated in the second device.
16 . An electronic device comprising: a temperature predictor configured to predict a temperature rise of the electronic device based on a trained artificial neural network when an application is started, the temperature rise being associated with executing the application, generate input data sets for the artificial neural network, and train the artificial neural network; and a processor configured to predict a temperature of the electronic device based on the temperature rise predicted by the temperature predictor and a current temperature of the electronic device, and perform an operation for lowering the temperature of the electronic device based on the predicted temperature of the electronic device, wherein the temperature predictor is configured to: obtain data sets, each of the data sets including a plurality of input parameters that include a measured temperature rise during a period in which the application is executed and the current temperature of the electronic device when the application is started; determine whether each of the plurality of input parameters has a correlation with the measured temperature rise; select one or more input parameters each determined to have the correlation with the measured temperature rise among the plurality of input parameters as each of the input data sets; and train the artificial neural network through supervised learning based on the input data sets each including the selected input parameters and the measured temperature rise.
17 . The electronic device of claim 16 , wherein the temperature predictor is configured to: test performance of the trained artificial neural network; and finish training of the artificial neural network when the performance of the trained artificial neural network is greater than or equal to a first threshold, and wherein the processor is configured to perform the operation for lowering the temperature of the electronic device when the predicted temperature of the electronic device is higher than a threshold temperature.
18 . The electronic device of claim 17 , wherein the temperature predictor is configured to: obtain more than a preset number of the data sets, the preset number being sufficiently great to ensure normality of data; obtain a correlation coefficient by performing a linear regression analysis on each of the plurality of input parameters and the measured temperature rise; determine that one or more input parameters each having the correlation coefficient greater than a threshold value have the correlation with the temperature rise; select the input parameters each determined as having the correlation with the measured temperature rise to generate the input data sets of the artificial neural network; train the artificial neural network using a portion of the input data sets; and test the performance of the trained artificial neural network using a remaining portion of the input data sets.
19 . The electronic device of claim 18 , wherein the temperature predictor is configured to: repeat the obtaining of the data sets, the determining of whether each of the plurality of input parameters has the correlation with the measured temperature rise, the selecting of the input parameters, and the training of the artificial neural network when the performance of the trained artificial neural network is less than the first threshold; finish training of the artificial neural network when a number of repetitions of the training of the artificial neural network is equal to or greater than a second threshold and the performance of the trained artificial neural network is greater than or equal to a third threshold; and discard the artificial neural network for the application, when the number of repetitions of the training of the artificial neural network is equal to or greater than the second threshold and the performance of the trained artificial neural network is less than the third threshold.

Description

CROSS REFERENCE TO RELATED APPLICATION The present application claims the benefit 35 U.S.C. 119(a) of Korea Patent Application No. 10-2022-0150837, filed Nov. 11, 2022, the contents of which is incorporated herein for all purposes by reference in its entirety. FIELD Various embodiments relate to an electronic device which predicts a temperature of a chip according to execution of an application (e.g., application program) and performs a pre-operation to substantially prevent an excessive increase in temperature of the chip, and an operation method of the electronic device. BACKGROUND A storage device is a device capable of storing data at a request from an external device such as a computer, a mobile terminal such as a smart phone or tablet, or various electronic devices. The storage device may include a memory and a memory controller for controlling the memory. The memory controller may receive a command from the external device and perform or control operations for reading data from the memory, writing/programming data in the memory, or erasing data from the memory based on the received command. Some memory devices may require a higher level of reliability than general-purpose products. In particular, these memory devices are desirable to ensure operation at a higher temperature than general-purpose products. When the memory devices operate at a high temperature, a policy such as clock frequency reduction may be performed to lower a product temperature, but the policy may be performed in a situation such as sudden power off (SPO), resulting in undesirable problems in terms of device operation. SUMMARY In order to solve the above problem, it is necessary to predict in advance when and how much the temperature of a device is to rise. Various embodiments of the present disclosure may provide a method of predicting in advance how much the temperature of a device is to rise based on artificial intelligence using supervised learning based on machine learning. The technical tasks to be achieved in the present disclosure are not limited to the technical tasks mentioned above, and other technical tasks not mentioned can be clearly understood by those of ordinary skill in the art to which the present disclosure belongs from the description below. According to various embodiments of the present disclosure, a method for operating an electronic device may include predicting a temperature rise of the electronic device when an application is started, the temperature rise being associated with executing the application, predicting a temperature of the electronic device based on the predicted temperature rise and a current temperature of the electronic device and lowering the temperature of the electronic device when the predicted temperature of the electronic device is higher than a preset threshold temperature. According to various embodiments of the present disclosure, the predicting of the temperature rise of the electronic device may include predicting the temperature rise of the electronic device caused by the application based on a trained artificial neural network. According to various embodiments of the present disclosure, the predicting of the temperature rise of the electronic device based on the trained artificial neural network may include obtaining data sets, each of the data sets including a plurality of input parameters that include a measured temperature rise during a period in which the application is executed and the current temperature of the electronic device when the application is started, determining whether each of the plurality of input parameters has a correlation with the measured temperature rise, selecting one or more input parameters each determined to have the correlation with the measured temperature rise among the plurality of input parameters as each of input data sets, training the artificial neural network through supervised learning based on the input data sets each including the selected input parameters and the measured temperature rise, testing performance of the trained artificial neural network and finishing training of the artificial neural network when the performance of the trained artificial neural network is greater than or equal to a first threshold. According to various embodiments of the present disclosure, the plurality of input parameters may further include a remaining battery level, a current time, a memory used by applications running in background, a usage time of the application, and whether or not charging is being performed. According to various embodiments of the present disclosure, the obtaining of the data of may include obtaining more than a preset number of data sets, the preset number being sufficiently great to ensure normality of data. According to various embodiments of the present disclosure, the selecting of input parameters may include obtaining a correlation coefficient by performing a linear regression analysis on each of the plurality of input para