Thermally Aware Hardware Scheduling and Binding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As integrated circuit feature sizes shrink, power density increases, leading to thermal hotspots and heat gradients that degrade circuit reliability and performance due to thermal stress, power optimization techniques often fail to address local hotspots effectively.
Innovation Solution
Implement thermally aware scheduling and binding methods that generate multiple versions of a hardware design with diverse thermal profiles, rotating between them to balance thermal activity and reduce peak temperatures, using linear programming to determine optimal operation durations for each version.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If technology scaling to smaller integrated circuit feature sizes is implemented, then device functionality and integration density are improved, but power density increases leading to thermal hotspots and reliability degradation
Solution Approach 1:
The patent implements dynamic thermal management by creating multiple versions of the hardware design with different scheduling and binding configurations. These versions exhibit diverse thermal profiles that are rotated during operation, transforming the static thermal problem into a dynamic solution where thermal characteristics change over time to balance peak temperatures.
Solution Approach 2:
The patent changes operational parameters by generating K versions of the design with different scheduling and binding parameters. Each version has distinct thermal characteristics, and by rotating between versions with different parameter sets, the system achieves thermal balance while maintaining high integration density.
2Use of energy by moving object
If power consumption optimization techniques are applied, then average power consumption is reduced, but local thermal hotspots may remain or even increase
Solution Approach 1:
The patent segments the hardware design into K different versions, each with its own scheduling and binding configuration. This segmentation allows the system to distribute thermal load across multiple operational modes, preventing concentration of heat in single hotspots while maintaining overall power efficiency.
Solution Approach 2:
The patent implements periodic rotation between K versions of the design, where each version operates for a determined duration before switching to the next. This periodic action redistributes thermal energy over time, preventing persistent hotspots while maintaining low average power consumption through efficient resource utilization.
3Temperature
If multiple versions of hardware design with diverse thermal profiles are generated and rotated, then peak temperature is reduced, but device complexity increases
Solution Approach 1:
The patent creates K versions of the same hardware design, where each version serves the same functional purpose but with different thermal characteristics. This multi-functionality approach allows a single physical device to exhibit multiple thermal profiles through software-controlled version rotation, reducing peak temperature without requiring physically separate systems.
Solution Approach 2:
The patent generates K copies of the hardware design with varied scheduling and binding configurations. These copies are not physically separate systems but rather different operational configurations of the same hardware, allowing thermal management through version rotation while maintaining design efficiency.
Data Source
AI summary
Technologies are generally described for hardware synthesis using thermally aware scheduling and binding. Multiple versions of a hardware design may be generated, each having variations of schedule and binding results. The scheduling and binding may be performed such that thermal profiles of the multiple versions have thermal peaks that are distant between the versions. The increased physical distance between the thermal peaks of the versions can give the versions unique thermal characteristics. A schedule of rotation between the multiple versions of the design may be constructed such that the thermal profile of the integrated circuit balances out during operation. A linear programming framework may be used to analyze the multiple designs and construct a thermally aware rotation scheduling and binding. For example, the K most efficient versions may be selected and then durations for operating each version within a rotation may be determined.