Dynamic Thermal Time Constant Determination for Processing Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Processing systems, such as system-on-a-chip (SoC), face challenges in predicting and managing thermal time constants due to dynamic changes in workload, power consumption, and heat dissipation, which can lead to temperature exceeding the thermal design power (TDP), potentially damaging the system.
Innovation Solution
The system dynamically determines thermal time constants at runtime using feedback controllers like PI or PID, comparing measured temperatures to predicted values based on power consumption and heat dissipation rates, and adjusts performance states through closed-loop feedback control to prevent overheating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the processing system operates at high power consumption to improve performance, then productivity increases, but temperature rises above the thermal envelope causing system damage
Solution Approach 1:
The patent implements dynamic thermal management by continuously monitoring temperature and adjusting performance states in real-time. The system transitions between different performance states (e.g., boosting, sprinting, throttling) based on current thermal conditions, allowing the processing system to operate at high performance when thermally safe and reduce performance when approaching thermal limits, thus resolving the contradiction between productivity and temperature control
Solution Approach 2:
The system employs periodic performance states including boosting periods (high performance), sprinting periods (temporary high performance with thermal awareness), and cooling periods (reduced performance for thermal dissipation). These periodic transitions allow the system to achieve high average productivity while ensuring temperature remains within the thermal envelope through structured cycles of high and low performance operation
2Temperature
If thermal management control is increased to prevent overheating, then temperature control improves, but system complexity increases due to multiple performance states and transitions
Solution Approach 1:
The patent implements feedback-based thermal management where temperature sensors continuously monitor the thermal state and feed this information back to the performance state manager. The system adjusts performance states based on feedback from temperature readings, power consumption measurements, and thermal model predictions, creating a closed-loop control system that manages temperature effectively while using standardized feedback mechanisms to avoid excessive complexity
Solution Approach 2:
The system manages thermal complexity by changing key parameters such as performance state duration, transition thresholds, and thermal model parameters based on operating conditions. Rather than implementing complex hardware controls, the patent adjusts software-controlled parameters like boost duration, sprint timing, and throttle levels to achieve effective thermal management with minimal additional system complexity
3Measurement precision
If performance states are dynamically adjusted to manage thermal time constants, then thermal management accuracy improves, but processing time is lost during state transitions
Solution Approach 1:
The system performs preliminary thermal analysis and pre-calculates optimal performance state transitions based on the thermal model and predicted workload. By anticipating thermal conditions and preparing transition timing in advance, the system minimizes unnecessary transitions and reduces the total time lost to state changes while maintaining accurate thermal time constant management through pre-planned performance adjustments
Data Source
AI summary
A processing system includes one or more processing units to perform operations and one or more sensors to measure a temperature concurrently with the one or more processing units performing the operations. The processing system also includes a controller to receive feedback indicating the temperature and to determine a peak temperature and a thermal time constant for heating of the processing system based on a comparison of the measured temperature to a first temperature that is predicted based on the peak temperature and a previously determined thermal time constant for heating. Some embodiments of the controller can control a performance state of the processing system based on the peak temperature and the thermal time constant for heating of the processing system.