Dynamic Power Control via Work-Loop Deadline Proximity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Portable computing systems face challenges in optimizing power management due to unpredictable workloads, making it difficult to estimate future system busy periods and adjust power consumption effectively to avoid latencies or overreactions.
Innovation Solution
A system that dynamically controls power consumption by measuring performance during work-loops, determining derived completion times and deadline proximities, and adjusting power based on statistical distributions of completion times and deadline proximities to optimize power management for multiple work-loops executing on the device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If the system reduces power consumption during idle periods by powering down modules or reducing clock frequencies, then battery life is optimized, but the system experiences latency when rapidly powering up modules and increasing frequencies to meet future workloads
Solution Approach 1:
The system performs preliminary actions by keeping critical system modules in a low-power but not fully powered-down state during idle periods, and maintains higher clock frequencies for anticipated workloads. This allows the system to rapidly transition to full power when needed without experiencing the full latency of cold startup, thus balancing energy savings with response time requirements.
Solution Approach 2:
The system dynamically adjusts power states and clock frequencies based on predicted workload characteristics. By continuously monitoring system state and using predictive algorithms, the system can smoothly transition between power states rather than making abrupt changes, reducing both energy consumption and transition latency.
2Speed
If the system increases power consumption to rapidly respond to workload increases, then response time is improved, but the system experiences overreactions and unnecessary power consumption during moderate workloads
Solution Approach 1:
The system implements feedback mechanisms by continuously monitoring actual workload characteristics and comparing them against predicted workload patterns. This allows the system to adjust power consumption levels dynamically, increasing power only when actual workload matches the predicted high-workload pattern, and avoiding overreactions when workloads are moderate, thus optimizing the balance between response time and energy consumption.
Solution Approach 2:
The system changes operational parameters such as clock frequencies and voltage levels based on predicted workload characteristics. By using predictive algorithms to estimate future workload intensity, the system can pre-adjust parameters to optimal levels, avoiding both excessive power consumption and insufficient response capability.
3Productivity
If the system uses traditional real-time operating systems to predict future workloads, then scheduling effectiveness is improved, but the system cannot accurately estimate power requirements for non-periodic workloads
Solution Approach 1:
The system transitions from static, periodic workload assumptions to dynamic workload prediction that adapts to non-periodic patterns. By continuously learning from actual workload behavior and adjusting predictions accordingly, the system can accurately estimate future power requirements for both periodic and non-periodic workloads, maintaining scheduling effectiveness while improving power estimation reliability.
Solution Approach 2:
The system employs self-service mechanisms through machine learning algorithms that automatically learn workload patterns and predict future behavior without requiring manual configuration or periodicity assumptions. This allows the system to accurately estimate power requirements for diverse workload types, including non-periodic workloads, while maintaining high throughput through effective scheduling.
Data Source
AI summary
The present embodiments provide a system that dynamically controls power consumption in a computing device. During operation, the system measures the performance of the computing device while executing a work-loop. Next, the system determines a derived completion time for the work-loop based on the measured performance. (For example, the derived completion time can be an expected completion time, a maximum completion time, or more generally a completion time distribution.) The system then determines a deadline-proximity for the work-loop based on a comparison between the derived completion time and a deadline for the work-loop. (For example, the deadline-proximity can be an expected deadline-proximity, a minimum deadline-proximity, or more generally a deadline-proximity distribution.) Finally, the system controls the power consumption of the computing device based on the determined deadline-proximity for the work-loop.