Processor Vmin Calibration Using Ring Oscillator Aging Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The initial low end operating voltage set for semiconductor devices increases over time due to stress effects like Hot Carrier Injection and Bias Temperature Instability, leading to errors and failures if not adequately compensated, while existing guardband techniques are inefficient in power usage.
Innovation Solution
Perform High Temperature Operating Life stress testing to gather degradation data, create a prediction model, and dynamically adjust the low end operating voltage based on ring oscillator oscillations to set an optimal voltage limit for the processor, repeating this process during operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed guardband is added to the initial Vmin value to compensate for Vmin increase over time, then reliability is improved, but power efficiency deteriorates
Solution Approach 1:
The patent applies dynamics by transitioning from a static fixed guardband approach to a dynamic adaptive guardband system. The guardband is continuously adjusted based on real-time monitoring of Vmin shifts and actual device degradation, allowing the system to optimize power efficiency while maintaining reliability. The guardband increases gradually as degradation is detected rather than being fixed at a pessimistic initial value.
Solution Approach 2:
The patent implements feedback mechanisms through continuous monitoring of device performance parameters and Vmin shifts. The system uses this feedback to dynamically adjust the guardband value, creating a closed-loop control system that adapts to actual device degradation patterns. This feedback-driven approach eliminates the need for overly conservative initial guardband settings.
2Reliability
If a fixed pessimistic guardband is employed from the beginning of usage to ensure good yields and quality metrics, then reliability is improved, but power efficiency deteriorates
Solution Approach 1:
The system dynamically adjusts the guardband based on actual device degradation rather than using a fixed pessimistic value. The guardband evolves over time as the device ages, starting low and increasing only when degradation is detected, thereby maintaining quality metrics while optimizing power efficiency throughout the device lifecycle.
Solution Approach 2:
The system performs self-diagnosis and self-adjustment by continuously monitoring its own degradation patterns and automatically adjusting the guardband accordingly. This self-service capability eliminates the need for external recalibration and allows the system to maintain optimal performance characteristics throughout its operational life.
3Use of energy by moving object
If the low end operating voltage is set at or near the initial Vmin, then power efficiency is improved, but reliability deteriorates over time
Solution Approach 1:
The operating voltage is dynamically adjusted based on real-time degradation monitoring. The system starts with a low end voltage near the initial Vmin for optimal power efficiency, then gradually increases the voltage threshold as degradation is detected, maintaining both power efficiency and reliability throughout the device lifecycle.
Solution Approach 2:
The system uses feedback from continuous monitoring of device performance and Vmin shifts to dynamically adjust the operating voltage. This feedback mechanism allows the system to maintain operation near the initial Vmin when the device is healthy, improving power efficiency, while automatically adjusting upward when degradation patterns indicate reliability risks.
Data Source
AI summary
The low end operating voltage of an integrated circuit is adjusted. Oscillations are counted at a ring oscillator on the integrated circuit over a designated period of clock cycles. Based on the number of oscillations, a prediction model associated with a first set of device degradation data and a second set of static random-access memory (SRAM) low end operating voltage data is used to select a low end operating voltage limit for a processor on the integrated circuit. The low end operating voltage of the processor is set based on the selected low end operating voltage limit. These steps are repeated multiple times during operation of the processor. A method of testing integrated circuits to provide the data employed to produce the prediction model is also provided.