Authors:
Saturn Tsai, Hawaii Xuan, Edward Lee, Merrick Wang
Abstract:
This disclosure proposes a novel proactive power-saving mechanism that dynamically adjusts CPU frequency based on acoustic sensing. By leveraging the built-in microphone of POS systems, the proposed method can accurately detect user interactions and adjust the CPU's P-state accordingly, optimizing performance and energy consumption.
Background:
Currently, CPU frequency scaling is either fixed or dynamically adjusted based on application load. Existing power-saving mechanisms primarily focus on internal CPU optimizations, lacking external environmental sensing and response. While some devices are equipped with proximity sensors, they can be susceptible to environmental interference, leading to inaccurate detection.
Description:
We propose an acoustic-based CPU frequency scaling solution. By analyzing audio signals captured by the POS system's microphone, we can accurately determine whether a user is interacting with the system. When no voice activity is detected, the system will lower the CPU's P-state through ACPI P-States, reducing power consumption.
ACPI P-States: P-States is a CPU power management feature that allows dynamic adjustment of frequency and voltage to match the workload. P0 is the highest performance state, while P1-Pn are lower power states.
Acoustic Sensing Trigger: When voice activity is detected, the system will increase the P-state to ensure a quick response to user input. A configurable threshold can be set to accurately distinguish between user interactions and background noise.
TGCS Reference 00259
