Hybrid Prioritized Resource Allocation in Computing Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for managing power in thermally- or power-constrained computing devices often reduce frequency uniformly, which can unnecessarily decrease performance in non-critical components, failing to prioritize high-speed operation for essential functions.
Innovation Solution
A method for dynamically controlling clock frequencies of compute elements based on their priority, allocating greater power to higher priority components by determining minimum and maximum frequencies, and distributing frequency budgets across elements with varying weights to ensure optimal performance while adhering to thermal design limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Temperature
If frequency is reduced uniformly across all components to manage power consumption, then thermal design limits are maintained, but performance of non-critical components is unnecessarily reduced
Solution Approach 1:
The patent implements local quality by differentiating frequency scaling across different components based on their priority levels. Critical components maintain high frequencies while non-critical components receive frequency reductions, allowing thermal management to be applied locally rather than uniformly across the entire system.
Solution Approach 2:
The patent segments the system into priority-based groups (critical vs. non-critical components) and applies different frequency scaling policies to each segment. This segmentation enables selective power management that preserves performance where needed while reducing power consumption where acceptable.
2Loss of energy
If frequency is reduced non-uniformly based on power efficiency criteria, then power consumption is optimized, but the most important functions may not run at high speed
Solution Approach 1:
The patent changes the control parameter from power efficiency criteria to priority-based criteria. By using priority levels as the governing parameter, the system can guarantee high-speed operation for critical functions while optimizing power consumption for non-critical functions, directly addressing the limitation of efficiency-based approaches.
Solution Approach 2:
The patent implements feedback mechanisms that monitor both power consumption and performance requirements, using priority information to dynamically adjust frequency allocation. This feedback loop ensures that critical functions receive adequate power resources while maintaining overall system power efficiency.
Data Source
AI summary
Various embodiments comprise prioritizing frequency allocations in thermally- or power-constrained computing devices. Computer elements may be assigned ‘weights’ based on their priorities. The computer elements with higher weights may receive higher frequency allocations to assure they receive priority in processing more quickly. The computer elements with lower weights may receive lower frequency allocations and suffer a slowdown in their processing. Elements with the same weight may be grouped together for the purpose of frequency allocation.