CPU Task Scheduling with External IP Feedback for Power Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing task scheduling methods for application processors with multiple cores do not effectively utilize feedback data from both internal cores and external intellectual property blocks, leading to suboptimal performance and power efficiency, especially for GPU-centric tasks.
Innovation Solution
A task scheduling method that receives task processing requests, obtains first feedback data for multiple cores, and second feedback data for external IP blocks, such as GPUs and display devices, to dynamically assign tasks to cores based on utilization and performance requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional task scheduling methods are used that only consider internal core feedback data, then the scheduling process is simple, but the performance and power efficiency are suboptimal
Solution Approach 1:
The patent implements a feedback mechanism that collects utilization data from both internal CPU cores and external IP blocks (such as GPUs and display devices). This feedback loop enables the scheduler to make informed decisions about task placement by considering real-time resource utilization across the entire system, not just within the CPU, thereby improving processing efficiency while maintaining manageable complexity through structured data collection and analysis.
Solution Approach 2:
The scheduling system is designed to handle multiple types of tasks and resources universally. It can schedule tasks across different core types (high-performance and low-power cores) and consider the status of various IP blocks (GPU, display device, etc.). This multi-functional approach allows a single scheduling framework to optimize diverse workloads including GPU-centric tasks, general computing tasks, and display-related operations, improving overall system productivity without requiring separate scheduling mechanisms for each task type.
2Loss of energy
If feedback data from external IP blocks is incorporated into scheduling decisions, then performance and power efficiency improve, but the data collection and processing complexity increases
Solution Approach 1:
The system establishes feedback channels from external IP blocks to the scheduler, enabling continuous monitoring of their utilization states. This feedback mechanism provides real-time information about GPU, display device, and other IP block usage, allowing the scheduler to make power-efficient decisions by directing tasks to appropriate resources based on current system state, thereby reducing overall power consumption while managing data collection complexity through standardized feedback interfaces.
Solution Approach 2:
The scheduler acts as an intermediary that aggregates utilization information from multiple sources (internal cores and external IP blocks) and translates this complex data into simplified scheduling decisions. It mediates between the diverse feedback data from various IP blocks and the task allocation process, consolidating information about resource availability and utilization into a unified decision-making framework that improves power efficiency without overwhelming complexity in data handling.
3Productivity
If tasks are dynamically assigned based on multiple feedback data sources, then system performance per watt improves, but the scheduling decision complexity increases
Solution Approach 1:
The scheduling system dynamically adapts task allocation based on real-time feedback from internal cores and external IP blocks. It can adjust scheduling decisions on-the-fly in response to changing system conditions, such as varying utilization levels of GPU, display devices, and CPU cores. This dynamic approach enables the system to optimize performance per watt by directing tasks to the most appropriate resources at any given moment, while managing decision complexity through structured evaluation criteria and automated decision-making algorithms.
Solution Approach 2:
The scheduler evaluates multiple parameters including utilization data from CPU cores, GPU, and display devices to make informed scheduling decisions. It changes scheduling parameters such as task assignment, core selection, and resource allocation based on the aggregated feedback data. By systematically analyzing and responding to changes in these parameters, the system optimizes performance per watt while managing decision complexity through parameter-based evaluation frameworks that structure the scheduling decision process.
Data Source
AI summary
A task scheduling method for a central processing unit (CPU) including a plurality of cores includes receiving a task processing request, obtaining first feedback data for the plurality of cores, obtaining second feedback data for an external intellectual property (IP) block outside the CPU, and assigning a task to at least one of the plurality of cores based on the first feedback data and the second feedback data.