Dynamic Weight Calculation for Digital Power Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional thermal measurement mechanisms lack granularity and repeatability, making it difficult to modulate power and thermal activity in individual components of modern processors, especially as application programs execute and processor loads vary.
Innovation Solution
A power management system that dynamically calculates and adapts weight factor values for digital power estimation, using signal activity and voltage regulator measurements to accurately estimate and control power consumption across processor cores, allowing for adaptive performance adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional thermal measurement mechanisms are used, then power management can be implemented at the chip level, but measurement precision and granularity are insufficient for individual component control
Solution Approach 1:
The patent divides the chip into multiple power domains, each with its own power management capabilities. Instead of measuring total chip power, the system segments power measurement into individual domain measurements, enabling precise control of specific components while maintaining manageable system complexity through modular architecture.
Solution Approach 2:
The patent introduces digital power estimation mechanisms as intermediary systems that correlate easily measurable signals (voltage, frequency, activity counters) with actual power consumption. These intermediaries translate simple measurements into accurate power estimates without requiring direct complex thermal or power measurements at each component level.
2Measurement precision
If fixed weight factor values are used in digital power estimation, then the system is simple to operate, but accuracy degrades as processor loads and application programs vary
Solution Approach 1:
The patent transitions from static fixed weight factors to dynamic adaptive weight factors that automatically adjust based on observed power consumption patterns. The system continuously monitors actual power usage and recalibrates weight factors to match current workload characteristics, maintaining accuracy across varying processor loads and application programs.
Solution Approach 2:
The patent implements feedback mechanisms where measured power consumption data is fed back into the power estimation model to continuously refine and update weight factor values. This closed-loop system automatically adapts to changing conditions without manual intervention, improving estimation accuracy while the automated nature keeps complexity manageable.
3Productivity
If power management controls the entire chip, then thermal limits are protected, but individual component performance cannot be optimized
Solution Approach 1:
The patent segments the chip into multiple independently controllable power domains, each with its own power management and thermal monitoring. This allows performance optimization in domains with available thermal headroom while maintaining safety in domains approaching thermal limits, enabling simultaneous performance improvement and thermal protection.
Solution Approach 2:
The patent applies different power management strategies to different regions of the chip based on local thermal conditions and workload characteristics. Each power domain can operate with customized voltage, frequency, and power limits tailored to its specific thermal environment, allowing high-performance operation in cool regions while maintaining safety in hot regions.
Data Source
AI summary
A system includes a power management unit that may be configured to estimate the power consumed by at least a portion of each of one or more processor cores during operation of each processor core. The power management unit may be configured to generate a sum of activity values and normal weight factor values for a predetermined set of signals within each processor core to estimate the power consumed. The power management unit may also be configured to adaptively generate and selectively use new weight factor values to estimate the power consumed based upon a total measured dynamic power consumed by each processor core during operation.