DVFS Control Using End-User Experience Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Dynamic Voltage/Frequency Scaling (DVFS) methods in mobile computing devices rely on load-based metrics, which often compromise user experience by inadequately adjusting processor operating points, failing to balance energy and thermal constraints with minimal impact on user experience.
Innovation Solution
A system and method that monitors end-user experience metrics to adjust processor operating points, incorporating device policies to optimize voltage and frequency, ensuring minimal impact on user experience while adhering to energy and thermal constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If load-based DVFS methods are used to adjust processor operating points, then energy consumption is reduced, but user experience deteriorates
Solution Approach 1:
The patent changes the basis for DVFS from load-based metrics to user experience-based metrics. By monitoring user experience metrics (such as application responsiveness, display smoothness, audio quality) and using these as the foundation for adjusting voltage and frequency operating points, the system achieves energy savings while maintaining perceived performance quality. This parameter change in the control basis resolves the contradiction between energy reduction and user experience maintenance.
Solution Approach 2:
The patent implements a feedback mechanism where user experience metrics are continuously monitored and fed back to the DVFS controller. This feedback loop allows the system to dynamically adjust processor operating points based on actual user perception of performance rather than raw load metrics. The feedback ensures that energy-saving adjustments do not degrade user experience, as the system responds to actual user experience conditions rather than theoretical load states.
2Duration of action of stationary object
If processor operating points are adjusted to meet energy and thermal constraints, then device longevity is improved, but performance quality deteriorates
Solution Approach 1:
The patent changes the control parameter from resource utilization to user experience metrics when determining operating points. By using user experience metrics as the primary basis for DVFS decisions, the system can aggressively manage power and thermal parameters (extending device longevity) while maintaining performance quality as defined by user perception. This parameter change allows the system to operate in a different control space where longevity and performance quality are not in direct conflict.
Solution Approach 2:
The patent implements dynamic adjustment of operating points based on real-time user experience conditions rather than static load-based thresholds. The system continuously adapts voltage and frequency settings according to monitored user experience metrics, allowing optimal performance-quality-to-longevity balance that changes with usage patterns. This dynamic approach ensures performance quality is maintained only when and where users actually perceive it, extending overall device longevity.
3Productivity
If load-based metrics are used for DVFS control, then resource utilization is optimized, but user experience metrics suffer
Solution Approach 1:
The patent inverts the traditional DVFS control approach by using user experience metrics as the primary control basis instead of load-based metrics. Rather than asking 'how much load is the processor handling?' the system asks 'how is the user experiencing the performance?' This inversion fundamentally changes the control objective from resource optimization to experience optimization, resolving the contradiction by making user experience the driving factor for resource allocation decisions.
Solution Approach 2:
The patent implements feedback based on user experience metrics to control DVFS decisions. By continuously monitoring user experience conditions and using this feedback to adjust operating points, the system ensures that resource utilization decisions are made with direct reference to their impact on user experience. This feedback mechanism prevents the decoupling of resource optimization from user experience that plagues load-based methods.
Data Source
AI summary
A system and method for dynamic voltage and frequency scaling for processors based on end-user experience metrics is disclosed. In an embodiment, a method in a data processing system for controlling a processor includes monitoring at least one end-user experience metric; obtaining at least one device policy, the device policy specifying a device constraint or measurable device condition; and adjusting at least one operating point for a processor according to at least one end-user experience metric and at least one device policy.