Stacked Die Thermal Telemetry for Virtual Temperature Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing thermal management schemes in multichip modules fail to accurately predict temperature changes due to thermal impacts from neighboring dies, leading to inaccurate performance management and potential reliability issues.
Innovation Solution
A thermal management scheme that utilizes telemetry information from both local and remote thermal sensors to estimate temperature virtually, allowing for precise performance management by adjusting voltage, frequency, and refresh rates based on the estimated temperature.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If thermal sensors are placed only on local dies without considering neighboring dies, then device complexity is reduced, but measurement precision deteriorates due to inaccurate temperature prediction
Solution Approach 1:
The patent introduces an intermediary computational model that processes thermal data from multiple sources (local and remote sensors) to predict temperatures on dies without direct sensors. This mediator translates remote thermal measurements into accurate local temperature predictions, resolving the contradiction between reduced sensor quantity and maintained measurement precision.
Solution Approach 2:
The patent creates virtual copies of thermal sensor data by using measurements from physically distant sensors to represent temperatures on dies without direct sensing. The computational model generates virtual temperature readings that accurately reflect the thermal state of target dies, enabling precise measurement without placing sensors on every die.
2Measurement precision
If thermal sensors are placed on every die to ensure accurate temperature monitoring, then measurement precision is improved, but device complexity and area increase
Solution Approach 1:
The computational model acts as an intermediary that infers temperatures on sensor-less dies by processing data from remote sensors. This intermediary system provides accurate temperature monitoring for all dies without requiring physical sensors on each die, thus maintaining measurement precision while reducing device complexity.
Solution Approach 2:
The patent makes remote thermal sensors serve multiple functions: they not only monitor their local die temperatures but also provide data for predicting temperatures on neighboring dies. This multi-functionality allows a single sensor to contribute to the thermal monitoring of multiple dies, reducing the total sensor count while maintaining comprehensive monitoring accuracy.
3Ease of operation
If thermal impacts from neighboring dies are ignored, then ease of operation is improved through simpler performance management, but reliability deteriorates due to inaccurate performance settings
Solution Approach 1:
The patent implements a feedback mechanism where thermal data from all dies (including neighboring dies) is continuously collected and fed into a computational model. This model dynamically adjusts performance settings based on the predicted thermal state of each die, ensuring reliable performance management that accounts for thermal impacts while maintaining systematic simplicity through automated feedback control.
4Area of stationary object
If the number of thermal sensors is reduced, then area efficiency is improved, but measurement precision deteriorates without considering thermal impacts from aggressor dies
Solution Approach 1:
The computational model serves as an intermediary that processes thermal data from reduced sensor sets and generates accurate temperature predictions for all dies. This intermediary system compensates for the reduced sensor count by using thermal physics-based modeling to infer temperatures on dies without direct sensors, maintaining measurement precision while improving area efficiency.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces the number of thermal sensors needed, decouples dielet designs, and enhances thermal management accuracy, thereby improving power and area efficiency while maintaining reliability and performance.
Implementation Method 1
hot spots in a first set of dies (aggressor dies) can impact thermal constraints of a second set of dies
Data Source
AI summary
A thermal management scheme, for a multichip module, that is aware of various dies in a stack (horizontal and/or vertical) and heat generated from them, local hot spots in a victim die, and hot spots in aggressor die(s). Each victim die receives telemetry information from thermal sensors located in aggressor dies as well as local thermal sensors in the victim die. The telemetry information is used to enable a virtual sensing scheme where temperature for a target die (e.g., a victim die) and/or its intellectual property (IP) domain is estimated or calculated. The estimated or calculated temperature is then used for performance management of the victim and/or aggressor dies in the stack.