Memory Cooling via Predictive Error Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer systems unnecessarily reduce functionality by conservatively managing cooling resources based solely on memory temperature, disregarding actual memory errors, and prohibiting operation above recommended temperatures, even if memory can handle errors at higher temperatures.
Innovation Solution
A predictive failure analysis (PFA) module dynamically monitors correctable memory errors and temperature, managing cooling resources based on both factors to optimize performance without immediately throttling systems when temperatures exceed recommended limits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If cooling resources are managed conservatively based solely on recommended temperature limits, then memory reliability is maintained, but system functionality and performance are unnecessarily reduced
Solution Approach 1:
The system implements feedback by continuously monitoring actual memory error rates and using this information to dynamically adjust cooling resource management. The PFA module monitors correctable memory errors and feeds this information back to the operating system, which then adjusts cooling resources based on actual error conditions rather than relying solely on conservative temperature thresholds, thereby resolving the contradiction between maintaining reliability and preserving productivity.
2Productivity
If memory operates above recommended temperature limits, then system performance is improved, but memory error rate increases
Solution Approach 1:
The system applies dynamics by transitioning from static temperature-based throttling to dynamic error-rate-based cooling management. The PFA module continuously monitors correctable memory errors in real-time, and the operating system dynamically adjusts cooling resources based on the actual error rate conditions. This allows the system to operate memory at higher temperatures when error rates are acceptable while maintaining reliability when errors increase, thus resolving the contradiction between productivity and reliability.
3Temperature
If cooling resources are increased to maintain recommended temperature limits, then memory temperature is controlled, but system functionality is reduced due to throttling
Solution Approach 1:
The system changes the control parameter from temperature-based throttling to error-rate-based cooling management. Instead of increasing cooling resources when temperature exceeds conservative thresholds, the PFA module monitors actual correctable memory error rates and the operating system adjusts cooling resources based on these error conditions. This parameter change allows the system to maintain appropriate memory temperature control while avoiding unnecessary throttling and preserving system functionality.
4Reliability
If conservative temperature thresholds are used for cooling management, then soft error rate is reduced, but system performance is unnecessarily limited
Solution Approach 1:
The system implements self-service by enabling memory to operate autonomously at higher temperatures when actual error rates remain within acceptable limits. The PFA module provides the operating system with information about actual correctable memory error conditions, allowing the system to self-adjust cooling resources based on real error data rather than conservative predetermined thresholds. This resolves the contradiction by reducing soft error rates through actual monitoring while maintaining performance through intelligent, data-driven decisions.
Data Source
AI summary
Operating computer memory in a computer including dynamically monitoring, by a predictive failure analysis (‘PFA’) module, correctable memory errors and memory temperature and managing cooling resources in the computer in dependence upon the correctable memory errors and memory temperature.