Temperature based optimization of data storage operations
Systems, methods and apparatus of intelligent optimization of data storage operations based on operating temperature. For example, a data storage device includes: one or more storage media components; a controller; at least one temperature sensor configured to generate temperature measurements; and an artificial neural network configured to predict an operating parameter of an operation of the one or more storage media components at a temperature determined by the temperature sensor. The storage media component can use the operating parameter predicted by the artificial neural network in performing the operation, such as writing data into a memory unit or reading data from the memory unit. For example, the operating parameter can be a voltage for programming multiple bits into a memory cell, or reading data from a memory cell.
1. A data storage device, comprising:
one or more storage media components;
at least one temperature sensor configured to generate temperature measurements;
a controller comprising a processing device configured to:
receive the temperature measurements from the at least one temperature sensor;
predict, by utilizing an artificial neural network, an operating parameter of an operation of the one or more storage media components at a temperature associated with the temperature measurements determined by the temperature sensor, wherein the operating parameter is predicted based on the temperature measurements being indicative of an operating temperature for the operation and based on a temperature history and a memory usage history of the data storage device;
perform the operation based on the operating parameter predicted by utilizing the artificial neural network;
determine a difference between a measured performance level of a result of the operation performed using the operating parameter and a predicted performance level of the result of the operation to determine whether to further train the artificial neural network, wherein the predicted performance level of the result of the operation is predicted by the artificial neural network based on the temperature associated with the temperature measurements;
apply at least one random variation to the operating parameter predicted by the artificial neural network to generate at least one randomized operating parameter; and
generate, based on the difference between the measured performance level and the predicted performance level exceeding a threshold for the temperature and based on utilizing the at least one randomized operating parameter, training data to further train the artificial neural network for enhancing at least one future prediction associated with the one or more storage media components.
2. The data storage device of claim 1 , wherein the artificial neural network is trained based on unsupervised machine learning or supervised machine learning to recognize patterns in the temperature measurements.
3. The data storage device of claim 2 , wherein the operation of the data storage device includes writing data into a memory unit in a storage media component of the data storage device.
4. The data storage device of claim 3 , wherein the operating parameters include a voltage to write the data into the memory unit.
5. The data storage device of claim 2 , wherein the operation of the data storage device includes reading data from a memory unit in a storage media component of the data storage device.
6. The data storage device of claim 5 , wherein the operating parameters include a threshold voltage to read the data from the memory unit.
7. The data storage device of claim 2 , further comprising:
a neural network accelerator configured to predict, at the temperature, the operating parameter using model data of the artificial neural network stored in the data storage device.
8. The data storage device of claim 7 , wherein the neural network accelerator is further configured to predict, at the temperature, a predicted performance level of the result of the operation of the data storage device when the operation of the data storage device is performed using the operating parameter.
9. The data storage device of claim 8 , wherein the neural network accelerator configured to train the artificial neural network using performance data generated in the data storage device; and wherein the performance data include:
temperatures measured by the temperature sensor;
operating parameters used in operations of the data storage device performed at the temperatures measured by the temperature sensor; and
performance levels of results of the operations performed at the temperatures using the operating parameters.
10. The data storage device of claim 9 , wherein the performance levels identify an error rate in the result of the operations.
11. The data storage device of claim 10 , wherein one or more storage media components include an integrated circuit chip having multiple memory units; and the temperature is formed in the integrated circuit chip.
12. A method, comprising:
receiving, from a temperature sensor and at a processing device of a controller of a data storage device, a temperature measurement associated with the data storage device;
predicting, by the processing device utilizing an artificial neural network configured in the data storage device and based on the temperature, an operating parameter of an operation of the storage media component, wherein the operating parameter is predicted based on the temperature measurement being indicative of a temperature for the operation and based on a temperature history and a memory usage history of the data storage device;
performing, by a storage media component of the data storage device, the operation based on the operating parameter predicted using the artificial neural network;
determine a difference between a measured performance level of a result of the operation performed using the operating parameter and a predicted performance level of the result of the operation to determine whether to further train the artificial neural network, wherein the predicted performance level of the result of the operation is predicted by the artificial neural network based on the temperature associated with the temperature measurement;
apply at least one random variation to the operating parameter predicted by the artificial neural network to generate at least one randomized operating parameter; and
generate, based on the difference between the measured performance level and the predicted performance level exceeding a threshold for the temperature and based on utilizing the at least one randomized operating parameter, training data to further train the artificial neural network for enhancing at least one future prediction associated with the storage media component.
13. The method of claim 12 , further comprising:
measuring, by the data storage device, a measured performance level of the result of the operation performed by the storage media component based on the operating parameter predicted using the artificial neural network.
14. The method of claim 13 , wherein the method further comprises:
training, in the data storage device, the artificial neural network using:
the temperature measurement;
an operating parameter used in performance of the operation by the storage media component; and
the measured performance level measured by the data storage device.
15. The method of claim 14 , further comprising:
predicting, by the artificial neural network configured in the data storage device and based on the temperature, the predicted performance level of the result of the operation performed using the operating parameter predicted using the artificial neural network.
16. The method of claim 14 ,
wherein the storage media component performs the operation using the at least one randomized operating parameter, based on the operating parameter predicted using the artificial neural network.
17. A vehicle, comprising:
a data storage device; and
sensors configured to generate sensor data, including a temperature sensor configured to measure an environmental temperature of the data storage device;
wherein the data storage device includes a storage media component and an artificial neural network configured to predict an operating parameter of an operation of the storage media component at the environmental temperature, wherein the operating parameter is predicted based on the environmental temperature being indicative of an operating temperature for the operation and based on a temperature history of the data storage device, and wherein the artificial neural network is configured to predict an optimized refresh frequency for refreshing the data storage device based on the operating parameter and the temperature history and a memory usage history of the data storage device;
wherein the storage media component is configured to perform the operation based on the operating parameter predicted using the artificial neural network;
wherein the data storage device is configured to:
determine a difference between a measured performance level of a result of the operation performed using the operating parameter to a predicted performance level of the result of the operation to determine whether to further train the artificial neural network, wherein the predicted performance level of the result of the operation is predicted by the artificial neural network based on the temperature associated with the temperature measurements;
apply at least one random variation to the operating parameter predicted by the artificial neural network to generate at least one randomized operating parameter; and
generate, based on the difference between the measured performance level and the predicted performance level exceeding a threshold for the temperature and based on utilizing the at least one randomized operating parameter, training data to further train the artificial neural network for enhancing at least one future prediction associated with the one or more storage media components.
18. The vehicle of claim 17 , wherein the operation includes programming data into memory cells, or reading data from memory cells, or any combination thereof.
19. The vehicle of claim 18 , wherein the operating parameter includes a voltage to be applied on a memory cell during the operation.
20. The vehicle of claim 19 , wherein the data storage device is further configured to train the artificial neural network using performance data including:
temperatures measured by the temperature sensors;
operating parameters used in operations performed by the storage media component; and
performance levels of the operations performed using the operating parameters at the temperatures.