Storage Controller ML Clock Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage devices face challenges in maintaining performance while significantly reducing power usage, as the determination of clock values greatly affects performance and power consumption, leading to a trade-off between the two.
Innovation Solution
The implementation of a storage device with a controller that includes a parameter storage for power parameters derived using machine learning, which adjusts the frequency of internal components by clock division, gating, or gearing to optimize power efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If clock values are increased to improve performance, then performance is improved, but power consumption increases
Solution Approach 1:
The patent applies dynamic clock adjustment by determining optimal clock values for different power states (active, background, idle, sleep) based on machine learning predictions. The clock value is dynamically changed according to the predicted power state, allowing the system to optimize between performance and power consumption in real-time based on actual operational conditions.
Solution Approach 2:
The patent changes the clock value parameter based on machine learning predictions of power states. By predicting which power state the system will enter and adjusting the clock value accordingly, the system can prepare optimal performance levels for upcoming operations while reducing clock values during low-power states to minimize energy consumption.
2Use of energy by moving object
If machine learning models are used to predict power states and determine clock values, then power efficiency is improved, but device complexity increases
Solution Approach 1:
The storage device performs self-optimization by incorporating machine learning capabilities within the controller to automatically predict power states and determine optimal clock values. The device monitors its own operational patterns, predicts future states, and adjusts parameters autonomously without external intervention, thereby improving power efficiency while containing complexity within the controller.
Solution Approach 2:
The patent replaces manual or heuristic clock tuning mechanisms with machine learning-based automated determination. Instead of using fixed thresholds or simple timers to decide when to adjust clock values, the system uses ML models that process operational data, predict power states, and optimize clock values accordingly, substituting complex mechanical tuning processes with intelligent algorithms.
3Device complexity
If manual parameter optimization is used, then device complexity is reduced, but productivity and development time increase
Solution Approach 1:
The system automates the parameter optimization process by embedding machine learning models that automatically determine optimal clock values for different power states. This self-service capability eliminates the need for manual tuning and testing by developers, significantly reducing development time and productivity costs while maintaining optimal performance and power efficiency.
Solution Approach 2:
The patent performs preliminary training of machine learning models using historical operational data before the actual deployment. The models learn optimal clock value mappings during training, and once deployed, they automatically apply these learned patterns without requiring manual optimization during operation, thereby accelerating development and reducing time-to-market.
Data Source
AI summary
A storage device includes at least one nonvolatile memory device, and a controller controlling the at least one nonvolatile memory device. The controller includes a parameter storage storing a power parameter indicating a clock value of each of internal configurations for each power state. The power parameter is a value derived by performing machine learning considering performance, peak power, and average power.