Methods and apparatus to manage energy usage and compute performance
Methods, apparatus, systems, and articles of manufacture are disclosed. In one example, a compute device to manage energy usage and compute performance includes at least one memory, instructions, and processor circuitry. The processor circuitry executes the instructions to determine a system power mode based on first telemetry data associated with the compute device. The processor circuitry executes the instructions to provide user activity data and second telemetry data associated with the compute device to a classification system. The processor circuitry executes the instructions to configure a plurality of parameters to manage power consumption and performance of the compute device based on a classification by the classification system.
1. A compute device to manage energy usage and compute performance comprising:
at least one memory;
instructions; and
processor circuitry to execute the instructions to:
determine a system power mode based on first telemetry data associated with the compute device;
determine user activity data and second telemetry data associated with the compute device to a classification system;
obtain a classification of a workload for the system based on the system power mode, the second telemetry data, and the user activity data; and
configure a plurality of parameters to manage performance of the compute device based on the classification of the workload by the classification system and a normalized foreground ratio per processing unit of the compute device.
2. The compute device of claim 1 , wherein the classification system is a trained machine learning model.
3. The compute device of claim 1 , wherein the processor circuitry is to determine the system power mode from a set of predefined power modes.
4. The compute device of claim 3 , wherein the set of predefined power modes consists of first, second, and third power modes, the first power mode to configure the compute device to consume less power than the second power mode, the second power mode to configure the compute device to consume less power than the third power mode.
5. The compute device of claim 1 , wherein the first telemetry data includes a battery charge state and a determination that the compute device is coupled to an alternating current source.
6. The compute device of claim 5 , wherein the first telemetry data further includes ambient environment data, system configuration data, and a customer preference.
7. The compute device of claim 1 , wherein the second telemetry data includes system on a chip workload information and a user-defined configuration setting.
8. The compute device of claim 1 , wherein the plurality of parameters includes processor frequency, wireless communication enablement, graphics processor enablement, and secondary storage enablement.
9. The compute device of claim 1 , wherein ones of the plurality of parameters can be configured to a set of parameter values, each set of parameter values corresponding to a performance state of the compute device.
10. The compute device of claim 2 , wherein the classification is a first classification associated with a burst workload, wherein the plurality of parameters is a first plurality of parameters, and wherein the processor circuitry is to execute the instructions to:
configure the first plurality of parameters to increase power consumption by the compute device; and
after the first plurality of parameters is configured, in response to a change in a primary application associated with the burst workload, configure a second plurality of parameters to reduce power consumption by the compute device.
11. A non-transitory computer readable storage medium comprising instructions which, when executed, cause processor circuitry to:
determine a system power mode based on first telemetry data associated with a compute device;
determine user activity data and second telemetry data associated with the compute device to a classification system;
obtain a classification of a workload for the system based on the system power mode, the second telemetry data, and the user activity data; and
configure a plurality of parameters to manage performance of the compute device based on the classification of the workload by the classification system and a normalized foreground ratio per processing unit of the compute device.
12. The non-transitory computer readable storage medium of claim 11 , wherein the classification system is a trained machine learning model.
13. The non-transitory computer readable storage medium of claim 11 , wherein the instructions, when executed, cause the processor circuitry to determine the system power mode from a set of predefined power modes.
14. The non-transitory computer readable storage medium of claim 13 , wherein the set of predefined power modes consists of first, second, and third power modes, the first power mode to configure the compute device to consume less power than the second power mode, the second power mode to configure the compute device to consume less power than the third power mode.
15. The non-transitory computer readable storage medium of claim 11 , wherein the first telemetry data includes a battery charge state and a determination that the compute device is coupled to an alternating current source.
16. The non-transitory computer readable storage medium of claim 15 , wherein the first telemetry data further includes ambient environment data, system configuration data, and a customer preference.
17. The non-transitory computer readable storage medium of claim 12 , wherein the classification is a first classification associated with a burst workload, wherein the plurality of parameters is a first plurality of parameters, and wherein the instructions, when executed, cause the processor circuitry to:
configure the first plurality of parameters to increase power consumption by the compute device; and
after the first plurality of parameters is configured, in response to a change in a primary application associated with the burst workload, configure a second plurality of parameters to reduce power consumption by the compute device.
18. A method for managing energy usage and performance of a compute device, the method comprising:
determining, by executing an instruction with processor circuitry, a system power mode based on first telemetry data associated with the compute device;
determining, by executing an instruction with the processor circuitry, user activity data and second telemetry data associated with the compute device to a classification system;
obtaining, by executing an instruction with the processor circuitry, a classification of a workload for the system based on the system power mode, the second telemetry data, and the user activity data; and
configuring, by executing an instruction with the processor circuitry, a plurality of parameters to manage performance of the compute device based on the classification of the workload by the classification system and a normalized foreground ratio per processing unit of the compute device.
19. The method of claim 18 , wherein the classification system is a trained machine learning model.
20. The method of claim 18 , further including determining the system power mode from a set of predefined power modes.
21. The method of claim 20 , wherein the set of predefined power modes consists of first, second, and third power modes, the first power mode to configure the compute device to consume less power than the second power mode, the second power mode to configure the compute device to consume less power than the third power mode.