Adaptive Low-Power Policy Selection for Thermal Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional thermal throttling methods in mobile and wireless devices often result in performance degradation due to inadequate temperature control, leading to reduced efficiency and potential device damage from overheating, as they rely on a single control policy that does not adapt to varying scenarios and user preferences.
Innovation Solution
The implementation of adaptive optimization strategies that monitor thermal-performance parameters to determine operation scenarios and select corresponding low-power policies, such as Dynamic Voltage and Frequency Scaling (DVFS), CPU hot-plug, and task migration, based on long-term and short-term averages, and user-defined preferences to dynamically adjust power usage and core operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Temperature
If traditional thermal throttling is used to control temperature, then temperature can be maintained within target range, but performance degradation occurs due to unnecessary power reduction
Solution Approach 1:
The patent implements dynamic thermal throttling by transitioning from a static single-threshold policy to a dynamic multi-threshold policy that adapts to different operation scenarios. The system dynamically adjusts power reduction strategies based on real-time temperature monitoring and operational context, allowing performance to be maintained when temperatures are acceptable while applying targeted power reduction only when necessary, thus resolving the contradiction between temperature control and performance maintenance.
Solution Approach 2:
The patent changes the parameter structure by introducing multiple temperature thresholds (first threshold, second threshold) instead of a single threshold, and by adding operation scenario as a new parameter. This allows the system to differentiate between mild thermal conditions and critical thermal conditions, applying appropriate power reduction strategies for each scenario, thereby avoiding unnecessary performance degradation while still achieving temperature control.
2Speed
If power reduction is applied quickly to control temperature, then temperature can be reduced faster, but noticeable performance degradation occurs
Solution Approach 1:
The patent applies partial power reduction instead of excessive power reduction by using differentiated threshold-based strategies. When temperature exceeds the first threshold but remains below the second threshold, the system applies mild power reduction strategies that partially address the thermal issue without causing significant performance degradation. Only when temperature exceeds the second threshold does the system apply more aggressive power reduction, thus achieving temperature reduction speed while minimizing performance impact.
Solution Approach 2:
The system dynamically adjusts the aggressiveness of power reduction based on the operation scenario and temperature level. By monitoring both temperature and operation scenario, the system can modulate the speed of temperature reduction to match the actual thermal urgency, avoiding unnecessarily fast power reduction that would cause performance degradation while still achieving adequate temperature control.
3Device complexity
If a single control policy is used for all temperature control methods, then system complexity is reduced, but control efficiency decreases
Solution Approach 1:
The patent segments the control policy into operation scenario-specific sub-policies. Instead of using a single monolithic control policy, the system divides the control space into multiple operation scenarios, each with its own optimized control strategy. This segmentation allows each sub-policy to be tailored for maximum efficiency in its specific scenario while keeping the overall system manageable through automated scenario detection and policy selection.
Solution Approach 2:
The patent creates a universal control framework that handles multiple temperature control methods (DVFS, CPU hot-plug, task migration) through a unified multi-threshold policy structure. This universal framework maintains reduced complexity by using a consistent decision-making approach across different control methods, while achieving high control efficiency by selecting the appropriate control method and policy parameters based on the operation scenario.
4Productivity
If power reduction is applied slowly to maintain performance, then performance is maintained better, but temperature continues to rise causing overheating
Solution Approach 1:
The patent applies preliminary action by implementing preventive thermal management through multiple thresholds. The first threshold serves as a warning level that triggers mild power reduction strategies before the temperature reaches critical levels. This preliminary intervention prevents the need for aggressive power reduction later, allowing the system to maintain performance better while still preventing temperature from rising to dangerous levels. The second threshold provides a safety net for when preventive measures are insufficient.
Solution Approach 2:
The system continuously monitors temperature and operation scenario, providing real-time feedback to the control decision-making process. This feedback mechanism allows the system to adjust power reduction strategies dynamically, maintaining performance when temperatures are stable while applying necessary power reduction when temperature trends indicate potential overheating, thus balancing performance maintenance with temperature prevention.
Data Source
AI summary
Methods and apparatus are provided for adaptive optimization of low-power strategies. In one novel aspect, the device monitors one or more thermal-performance parameters and determines a plurality of operation scenarios for a plurality of corresponding low-power policies. Based on corresponding operation scenarios, the device selects corresponding low-power policy. The device applies different low-power strategy for temperature control based on low-power policies. Different low-power policy is applied to different low-power techniques, such as the DVFS, the CPU hot-plug, and the task migration. In another novel aspect, the device obtains one or more user-defined policy for each corresponding low-power technique. The selection of each low-power policy is further based on its corresponding user-defined policy. In one embodiment, the user-defined DVFS policy includes power policy, performance policy, and DVFS-balanced policy. The user-defined CPU hot-plug policy includes conservative policy, aggressive policy, and hot-plug-balanced policy. The user-defined task-migration policy includes performance policy, and task-migration-balanced policy.


