CPU Hot-Plug Management via Thread Load Rearrangement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CPU hot-plug mechanisms in portable devices fail to optimally balance power consumption and system performance, as they do not effectively assess the impact of CPU plugging/unplugging on thread workload distribution and CPU frequency, leading to inefficient power usage and performance trade-offs.
Innovation Solution
A performance management method that involves thread loading rearrangement and CPU frequency evaluation to determine the optimal number of active CPUs by using a lookup table to assess power consumption values, allowing for dynamic adjustments based on evaluated performance values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a CPU is plugged in to handle additional thread workload, then system performance is improved, but power consumption increases and battery life decreases
Solution Approach 1:
The patent performs preliminary evaluation of thread loading rearrangement and CPU frequency adjustment before actually plugging in a CPU. By calculating predicted performance values for different CPU configurations in advance, the system determines whether adding a CPU is truly necessary, thereby avoiding unnecessary power consumption while maintaining performance requirements.
Solution Approach 2:
The patent changes operational parameters (CPU frequency, thread distribution) before changing the number of active CPUs. By adjusting existing parameters to maximize utilization of current CPUs, the system may achieve performance improvements without the power penalty of activating additional CPUs.
2Ease of operation
If existing evaluation factors (CPU frequency, usage, thread count) are used to determine hot-plug decisions, then decision-making is simplified, but the accuracy of predicting actual performance impact is insufficient
Solution Approach 1:
The patent implements a feedback mechanism where the system predicts performance outcomes of potential CPU hot-plug actions, compares these predictions with actual performance requirements, and uses this information to make informed decisions. This feedback loop improves prediction accuracy by continuously learning from the relationship between thread distribution, CPU frequency, and actual performance impact.
3Use of energy by moving object
If thread loading rearrangement and CPU frequency evaluation are performed before CPU hot-plug decisions, then power-performance optimization is improved, but computational overhead and decision complexity increase
Solution Approach 1:
The patent segments the complex evaluation process into distinct modules: thread loading analysis, CPU frequency evaluation, performance prediction, and decision-making. By organizing the complexity into separate, manageable segments with defined interfaces, the system achieves thorough power-performance optimization while keeping the overall decision process structured and controllable.
Data Source
AI summary
Performance management methods for an electronic device with multiple central processing units (CPUs) are provided. First, thread loading rearrangement and CPU frequency evaluation are performed to obtain a plurality of evaluated performance values for different amounts of CPUs, wherein the plurality of evaluated performance values are relevant to power consumption values of the multiple CPUs. It is then determined whether to adjust an amount of used CPUs based on the plurality of evaluated performance values corresponding to the different amounts of CPUs.