Asymmetric Processing Elements Task Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern microprocessors face challenges in achieving high processing throughput while minimizing power consumption, particularly in devices like mobile internet devices and laptops, where transitioning between power states is time-consuming and not all tasks require full processing throughput.
Innovation Solution
Incorporating asymmetric processing elements with different power consumption, performance, and voltage characteristics, allowing tasks to be dynamically migrated between higher and lower power cores based on performance needs, using instruction translation logic to adapt architectures like CISC to RISC, and employing power management logic to optimize power and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If the processor is placed in a low-power state to reduce power consumption, then power consumption is reduced, but the time required to transition to and from the low-power state increases
Solution Approach 1:
The processor is divided into multiple processing elements with different power consumption characteristics and performance capabilities. This segmentation allows the system to select appropriate processing elements based on task requirements, avoiding the need to transition a single processor to low-power state and thereby reducing transition time overhead while maintaining power savings.
Solution Approach 2:
The system dynamically assigns tasks to different processing elements based on real-time power and performance requirements. This dynamic assignment allows the system to optimize power consumption without incurring transition penalties, as tasks can be moved between processing elements that are already in appropriate power states.
2Productivity
If the processor operates at high processing throughput to meet performance requirements, then processing speed is improved, but power consumption increases
Solution Approach 1:
Different processing elements are designed with different local qualities - specifically, different power consumption characteristics and performance capabilities. High-performance processing elements handle computationally intensive tasks requiring high throughput, while low-power processing elements handle simpler tasks, thereby optimizing the overall power-performance balance of the system.
Solution Approach 2:
The system changes operational parameters by selecting different processing elements based on task requirements. When high processing throughput is needed, high-performance processing elements are activated; when lower performance suffices, low-power processing elements are used. This parameter change approach allows the system to adapt power consumption to actual performance needs.
3Device complexity
If a single processor is used to handle all tasks, then device complexity is reduced, but the ability to optimize power consumption for different task requirements is limited
Solution Approach 1:
The processing system is designed with multi-functionality, where multiple processing elements can handle various types of tasks. Each processing element is capable of executing a broad range of instructions, but they differ in performance and power characteristics. This universality allows the system to adapt to different task requirements while maintaining a relatively simple overall architecture.
Solution Approach 2:
The system employs asymmetric processing elements with different power consumption characteristics and performance capabilities rather than identical symmetric processors. This asymmetry enables fine-grained power optimization by matching task requirements with appropriate processing elements, thereby achieving better power efficiency without significantly increasing system complexity.
Data Source
AI summary
Techniques to control power and processing among a plurality of asymmetric cores. In one embodiment, one or more asymmetric cores are power managed to migrate processes or threads among a plurality of cores according to the performance and power needs of the system.