DVFS Module Frequency Scaling Using Application-Specific Vsync Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Dynamic Voltage/Frequency Scaling (DVFS) algorithms in mobile computing devices often make counterproductive adjustments, leading to inefficient energy consumption and frame drops during image rendering, due to their reliance on application-agnostic performance data and failure to distinguish between different reasons for processor idleness.
Innovation Solution
A method that utilizes application-based performance data, obtained through a windowing system, to dynamically adjust the processor's operating point, focusing on the time increment between Vsync events to determine whether to scale up or maintain frequency, thereby avoiding unnecessary frequency changes that could lead to further frame drops.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If DVFS algorithms scale down processor frequency based on application-agnostic performance data, then energy consumption is reduced, but frame drops occur during image rendering
Solution Approach 1:
The patent segments performance data into application-specific categories, particularly distinguishing graphics-intensive applications from other workloads. The DVFS module separately monitors frame rendering metrics, processor utilization, and memory usage patterns to make informed frequency scaling decisions that preserve graphics performance while reducing energy consumption during non-graphics workloads.
Solution Approach 2:
The system implements a feedback mechanism where the DVFS module continuously monitors frame drop events and adjusts processor frequency accordingly. When frame drops are detected, the system increases frequency to restore rendering performance. This closed-loop control prevents frame drops while maintaining energy efficiency during normal operation.
2Reliability
If DVFS algorithms increase processor frequency to prevent frame drops, then image rendering quality is maintained, but energy consumption increases
Solution Approach 1:
The system dynamically adjusts processor frequency based on real-time detection of frame drop conditions. Instead of maintaining a constantly high frequency, the DVFS module transitions between frequency states only when frame rendering issues are detected, creating a dynamic response that balances quality and energy consumption.
Solution Approach 2:
The patent changes the operating parameters of the processor dynamically by adjusting frequency based on detected frame drop patterns. The system monitors frame timing metrics and modifies frequency parameters selectively, increasing frequency only when rendering quality deteriorates and reducing it when quality is maintained.
3Device complexity
If DVFS algorithms use application-agnostic performance data for frequency scaling, then device complexity is reduced, but performance optimization precision deteriorates
Solution Approach 1:
The patent segments performance monitoring into distinct application categories with specific metrics for each. Graphics applications receive dedicated monitoring for frame timing and rendering performance, while other applications use general performance counters. This segmentation provides precise measurements without requiring complex unified monitoring for all workloads.
Data Source
AI summary
Methods are provided for monitoring and adjusting performance of a mobile computing device having a windowing system. The windowing system is advantageously employed for both purposes. It generates performance data related to an application executed by a processor of the device and provides the performance data to a DFVS module of the device, which in turn determines the operating point of the device based on that data. As a consequence, the DFVS refrains from scaling-down the operating frequency of the processor in certain cases wherein idleness of the processor might provoke such a scale-down by a DVFS module that was informed only of application-agnostic performance data. The avoidance of inappropriate scale-downs may, for instance, improve the perceived smoothness of a progression of images presented on a display unit of the device.