Computing Device Thermal Control Instability Detection via Frequency Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Enterprise-level computing servers experience feedback-control instabilities in fan speeds and CPU chip frequencies, leading to thermal oscillations, energy wastage, and reduced compute performance, due to challenges such as quantization of sensor signals, time lags in telemetry transmission, time-varying flow impedance, competition among local controllers, and non-synchronicity of telemetry metrics.
Innovation Solution
A thermal stability assurance system that autonomously detects feedback control instability by executing workloads, recording thermal telemetry, converting it into a frequency domain, and generating alerts based on dissimilarity in the frequency domain, without modifying existing hardware, addressing issues like quantization, telemetry lags, and airflow oscillations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Temperature
If feedback control is used to regulate server temperature, then thermal control is achieved, but feedback control instability occurs causing oscillations in fan speed and CPU frequency
Solution Approach 1:
The patent implements a feedback mechanism where the service processor monitors thermal telemetry data and compares it against expected values. When deviations are detected indicating oscillation, the system adjusts fan speeds and CPU frequencies to counteract the instability, thereby maintaining temperature control while eliminating oscillations.
Solution Approach 2:
The service processor acts as an intermediary between the thermal control system and the hardware components. It receives thermal telemetry, analyzes it for oscillation patterns, and mediates control adjustments to fan speeds and CPU frequencies, preventing direct feedback instability from causing harmful oscillations.
2Reliability
If thermal telemetry is monitored continuously, then feedback control stability can be detected, but energy consumption increases
Solution Approach 1:
The system performs thermal telemetry analysis at selective intervals rather than continuously. The service processor monitors thermal data and only initiates detailed oscillation detection and mitigation when deviations are detected, reducing unnecessary energy consumption while maintaining reliable stability detection.
3Temperature
If fan speed oscillations occur, then thermal control feedback loop is active, but energy wastage increases significantly
Solution Approach 1:
The system uses feedback monitoring to detect fan speed oscillations by comparing thermal telemetry against expected values. When oscillations are detected, the service processor adjusts fan speeds to eliminate the oscillatory behavior, thereby reducing the cubic energy wastage associated with oscillating fan operation while maintaining effective thermal regulation.
4Temperature
If P-State oscillations are triggered by temperature oscillations, then thermal feedback control is responsive, but compute performance decreases
Solution Approach 1:
The service processor serves as an intermediary that detects temperature oscillations before they trigger harmful P-State switching. By identifying oscillation patterns early through thermal telemetry analysis, the system can adjust fan speeds and workload distribution to prevent P-State oscillations, thereby maintaining both temperature responsiveness and compute performance.
Data Source
AI summary
Systems, methods, and other embodiments associated with detecting feedback control instability in computer thermal controls are described herein. In one embodiment, a method includes executing a workload on the computing system, wherein the workload varies between a minimum and a maximum at a workload frequency. The method includes recording thermal telemetry from the computing system during execution of the workload. The method includes converting the recorded thermal telemetry into a frequency domain. The method includes detecting whether thermal control of the computing system exhibits feedback control instability based on dissimilarity in the frequency domain between the transformed thermal telemetry and the workload frequency. And, the method includes generating an electronic alert that indicates whether the thermal control of the computing device exhibits the feedback control instability.