Thread Scheduling Tuning System for Power-Performance Tradeoff
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional thread scheduling configurations in personal computing devices are labor-intensive, non-systematic, and lack generalization and customization capabilities, resulting in insufficient optimization of power and performance during workload execution.
Innovation Solution
A tuning system that utilizes a tradeoff indication controller and a tuning engine, combined with machine learning models, to automatically customize OS thread scheduling policies by adjusting parameters such as dynamic core count, idle states, and operating frequency, based on performance and power scores, to achieve a balanced power consumption and performance tradeoff.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional thread scheduling configurations are used, then system operation is maintained, but optimization of power and performance is insufficient
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting thread scheduling parameters (such as priority, timing, and resource allocation) based on real-time performance and power measurements. The tuning engine modifies these parameters to optimize the balance between productivity and energy consumption, transforming static scheduling configurations into dynamic, adaptive ones.
Solution Approach 2:
The patent implements feedback mechanisms where performance scores and power scores are continuously measured during workload execution. These scores feed back to the tuning engine, which uses them to iteratively adjust thread scheduling policies. This closed-loop feedback enables systematic optimization of the power-performance tradeoff.
2Productivity
If manual thread scheduling configuration is performed, then specific optimization can be achieved, but the process is labor-intensive and non-systematic
Solution Approach 1:
The patent applies self-service by enabling the system to automatically tune its own thread scheduling parameters without human intervention. The tuning engine autonomously measures performance and power scores, analyzes workload characteristics, and adjusts scheduling policies accordingly. This eliminates the need for manual configuration while maintaining systematic optimization.
Solution Approach 2:
The patent implements preliminary action through automated baseline measurements and initial tuning configurations that are performed before actual workload execution. The system pre-determines optimal scheduling parameters based on workload profiles and system characteristics, preparing the thread scheduling policy in advance to avoid time-consuming manual configuration during operation.
3Reliability
If fixed thread scheduling policies are used, then system stability is maintained, but adaptability to different workloads is insufficient
Solution Approach 1:
The patent applies dynamics by transitioning from fixed, static thread scheduling policies to dynamic, adaptive policies that automatically adjust to different workload characteristics. The tuning engine continuously monitors performance and power scores, then modifies scheduling parameters in real-time to match changing workload demands while maintaining system stability through controlled, measured adjustments.
Data Source
AI summary
An apparatus comprising: a model to generate adjusted tuning parameters of a thread scheduling policy based on a tradeoff indication value of a target system; and a workload monitor to: execute a workload based on the thread scheduling policy; obtain a performance score and a power score from the target system based on execution of the workload, the performance score and the power score corresponding to a tradeoff indication value; compare the tradeoff indication value to a criterion; and based on the comparison, initiate the model to re-adjust the adjusted tuning parameters.