Core Scheduling Using Energy Crossover Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional core scheduling in heterogeneous architectures fails to account for changes in efficiency of cores at different frequencies, leading to inefficient operation and limited throughput, particularly in systems with varying core types and manufacturing tolerances.
Innovation Solution
A scheduler that adjusts core scheduling based on energy crossover points, providing feedback to the operating system to migrate work or reduce workload on inefficient cores, thereby optimizing power consumption and performance by leveraging the most efficient cores for task execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional core scheduling is used in heterogeneous architectures, then core type-based task assignment is simple to implement, but system efficiency deteriorates due to ignoring frequency-dependent efficiency changes
Solution Approach 1:
The patent implements a feedback mechanism where the controller monitors core efficiency at different frequencies and provides this information to the operating system. The OS then adjusts scheduling decisions based on this feedback, dynamically optimizing task assignment to cores operating at their most efficient frequencies while maintaining implementation simplicity through structured information flow.
Solution Approach 2:
The patent transitions from static core type-based scheduling to dynamic frequency-aware scheduling. The system continuously monitors operating frequencies and efficiency metrics, adapting scheduling decisions in real-time to match current operational conditions, thereby resolving the contradiction between simple implementation and high efficiency.
2Speed
If cores operate at higher frequencies to increase throughput, then processing speed improves, but power consumption increases and efficiency deteriorates beyond crossover points
Solution Approach 1:
The patent changes the scheduling parameter from static core type assignment to dynamic frequency-based assignment. By monitoring operating frequencies and identifying crossover points where efficiency transitions occur, the system adjusts task allocation to keep cores operating in their most efficient frequency ranges, thereby maintaining high throughput without excessive power consumption.
Solution Approach 2:
The controller preliminarily determines crossover points and efficiency characteristics of cores before scheduling tasks. This advance preparation allows the operating system to make informed scheduling decisions that prevent cores from operating in inefficient high-frequency states, thus avoiding excessive power consumption before it occurs.
3Device complexity
If frequency-agnostic scheduling is used, then scheduling complexity remains low, but manufacturing tolerance variations cause inefficient core operation
Solution Approach 1:
The patent introduces a feedback loop that monitors actual core efficiency and provides this information to the scheduler. This feedback mechanism enables the system to compensate for manufacturing tolerance variations by adapting to actual core performance characteristics rather than relying on theoretical specifications, thereby improving reliability without excessive complexity.
Solution Approach 2:
The patent applies local quality optimization by tailoring scheduling decisions to individual core characteristics. Each core's efficiency profile and crossover points are monitored separately, allowing the scheduler to optimize task assignment for each core's specific operational characteristics, accommodating manufacturing variations while maintaining overall system efficiency.
Data Source
AI summary
Core scheduling based on energy crossover is described. In one or more implementations, a system includes a plurality of cores, a controller configured to communicate feedback associated with efficiency of the plurality of cores, and an operating system. The operating system is configured to receive the feedback and adjust core scheduling responsive to at least one of the plurality of cores operating in an inefficient state based on the feedback. The controller may monitor operation of the cores, determine crossover points indicating transitions between efficient and inefficient frequency ranges for the cores, and detect when operating frequencies are proximate to the crossover points. Core scheduling adjustments may include migrating work between cores or reducing workload while maintaining operating frequencies to optimize efficiency.