Spatial Workload Scheduling for Thermal-Aware Multi-Core Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-core processing systems face performance limitations due to thermal considerations, necessitating power throttling to maintain thermal limits, which reduces processing performance.
Innovation Solution
A control circuitry system that manages workload distribution across processing elements based on their spatial layout, deactivating active elements and activating inactive elements based on timing parameters and spatial distance to optimize heat dissipation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If processing power is increased by increasing clock speed of processing elements, then processing capacity is improved, but system temperature increases and thermal limits are exceeded
Solution Approach 1:
The patent implements dynamic workload scheduling that continuously monitors thermal conditions and adjusts processing element activation/deactivation in real-time. The scheduler dynamically migrates workloads between processing elements based on current thermal states, allowing the system to operate at high performance when thermal conditions permit while automatically reducing heat generation when thermal limits are approached, thus resolving the contradiction between processing capacity and temperature control
Solution Approach 2:
The patent divides the processing system into multiple independent processing elements with distinct thermal zones. By segmenting the workload across these spatially separated elements and selectively deactivating elements in hot zones while activating elements in cooler zones, the system can maintain high overall processing capacity while managing local temperature peaks, effectively resolving the contradiction between total processing capacity and localized temperature control
2Temperature
If clock speed of processing elements is throttled to reduce power consumption and system temperature, then thermal requirements are satisfied, but processing performance is reduced
Solution Approach 1:
Rather than uniformly throttling all processing elements, the patent implements dynamic selective throttling where only specific processing elements in thermal hot zones are deactivated or throttled, while elements in cooler zones continue operating at full performance. This dynamic, location-based approach maintains processing performance in cooler regions while controlling temperature in hot regions, resolving the contradiction between temperature control and performance maintenance
Solution Approach 2:
The patent applies different operational states to different processing elements based on their local thermal conditions. Elements in cool zones maintain high clock speeds and full performance, while elements in hot zones are throttled or deactivated. This local differentiation allows the system to satisfy thermal requirements in hot zones without globally reducing processing performance, resolving the contradiction between localized temperature control and overall performance
Data Source
AI summary
Aspects of the present disclosure relate to an apparatus comprising a plurality of processing elements having a spatial layout, and control circuitry to assign workloads to said plurality of processing elements. The control circuitry is configured to, based on a timing parameter, determine one or more active processing elements to deactivate; determine, based on the spatial layout, one or more inactive processing elements to activate; and deactivate said one or more active processing elements and activate said one or more inactive processing elements.