Processor Throttling via Power Management Agents
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current power management systems for multicore processors often force processors to operate at the lowest frequency and voltage mode in response to platform events, leading to unnecessary performance penalties and inefficient energy usage, as they fail to optimize throttling levels based on user activity and power capabilities.
Innovation Solution
A proactive power management system that includes a platform target agent to set an optimal throttling power target for the processor and power management agents to determine optimal throttling targets for each processing element, allowing for dynamic control of operating points to maintain performance while conserving energy, even in worst-case conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If the processor operates at the lowest frequency and voltage mode in response to platform events, then power consumption is reduced, but performance deteriorates unnecessarily
Solution Approach 1:
The patent applies local quality by differentiating throttling levels across different processing elements based on their specific workload characteristics and user activity patterns. Instead of uniformly throttling all cores to the lowest frequency and voltage mode, the system selectively applies optimized throttling targets to individual processing elements, allowing some to maintain higher performance levels while others are throttled, thereby resolving the contradiction between power consumption reduction and performance maintenance.
Solution Approach 2:
The system implements dynamics by continuously monitoring platform events, user activity, and power capabilities to dynamically adjust throttling targets in real-time. The proactive power management agents adapt the operating points of processing elements based on changing conditions, enabling the processor to transition smoothly between performance states rather than being stuck at fixed lowest modes, thus maintaining performance when needed while reducing power when appropriate.
2Reliability
If the processor is forced to minimum operating point, then protection against electrical failure is ensured, but energy efficiency is compromised
Solution Approach 1:
The patent applies preliminary action through proactive power management agents that anticipate platform events and user activity patterns before they fully manifest. By predicting when throttling will be needed and pre-positioning the processor at optimized intermediate operating points rather than immediately forcing minimum operation, the system maintains protection against electrical failure while avoiding unnecessary energy loss that would occur with aggressive throttling to minimum points.
Solution Approach 2:
The system implements feedback mechanisms where power management agents continuously monitor platform events, power capabilities, and processing element performance, then use this feedback to adjust throttling targets dynamically. This closed-loop control ensures that the processor maintains adequate protection levels while optimizing energy efficiency based on actual system conditions, preventing both over-throttling and under-protection scenarios.
3Device complexity
If uniform throttling is applied to all processing elements, then simplicity is maintained, but user activity and priority information are not optimized
Solution Approach 1:
The patent applies segmentation by dividing the processor into separate processing elements with individual throttling control. Instead of applying uniform throttling to all cores, the system segments the throttling management into per-element control based on user activity and priority information. Each processing element receives customized throttling targets from dedicated power management agents, enabling fine-grained optimization of user activity while maintaining manageable complexity through modular agent architecture.
Data Source
AI summary
In one embodiment, a processor includes: a plurality of processing elements to perform operations; a power management agent (PMA) coupled to the plurality of processing elements to control power consumption of the plurality of processing elements; and a throttling circuit coupled to the PMA. The throttling circuit is to determine a throttling power level for the plurality of processing elements based at least in part on translation information communicated from the PMA. Other embodiments are described and claimed.