Computational Storage Power Allocation for Sudden Power-Off Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computing devices with centralized architectures face limitations in meeting high-performance and high-efficiency demands due to excessive current consumption by computational modules, leading to potential sudden power-offs and affecting internal and peripheral devices.
Innovation Solution
A computational storage device with a power management integrated circuit that distributes power, a computing module for specific operations, and a performance monitoring module to monitor workloads and power states, requesting additional power when needed, and a method for power management across multiple devices to stabilize power supply.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If computational modules perform complex calculations including AI or machine learning operations, then computation performance is improved, but current consumption increases excessively causing sudden power-offs
Solution Approach 1:
The performance monitoring module continuously monitors workload and power state in advance, and the power management integrated circuit proactively adjusts power distribution before power-out occurs. This preliminary action prevents sudden power-offs by anticipating power demands and adjusting supply accordingly, while allowing computational modules to maintain high-performance operations.
Solution Approach 2:
The system implements a feedback loop where the performance monitoring module detects workload and power state, communicates with the power management integrated circuit, which then adjusts power distribution dynamically. This closed-loop feedback mechanism ensures computation performance is maintained while preventing excessive current consumption that would cause power-outs.
2Device complexity
If a centralized architecture is used to control overall operations, then system control is simplified, but performance is limited by the bottleneck of specific circuits or components
Solution Approach 1:
The system segments the centralized control function by distributing computational tasks across multiple computational modules within the storage device. Each module operates semi-independently, reducing the bottleneck effect of any single component while maintaining overall system coordination through the power management integrated circuit and performance monitoring module.
Solution Approach 2:
The power management integrated circuit dynamically adjusts power distribution to computational modules based on real-time workload and power state monitoring. This dynamic allocation allows the system to optimize performance by directing power to active modules while reducing power to idle modules, thereby increasing overall processing throughput without excessive complexity.
3Device complexity
If power is distributed statically to computational modules, then power management is simplified, but computational performance is limited by fixed power allocation
Solution Approach 1:
The power management integrated circuit implements dynamic power distribution by continuously adjusting power allocation to computational modules based on real-time workload and power state monitoring. This dynamic approach increases computational throughput by providing more power to active modules and reducing power to idle modules, while the integrated circuit design keeps the overall management complexity manageable.
Solution Approach 2:
The performance monitoring module and power management integrated circuit work together to enable the system to self-regulate power distribution based on actual computational needs. This self-service mechanism allows the system to optimize its own power allocation without external intervention, improving computational throughput while maintaining manageable complexity through automated monitoring and adjustment.
Data Source
AI summary
A computational storage device that is configured to be electrically connected to a host, comprising: a power management integrated circuit that is configured to distribute power; a storage that is configured to store data using a first voltage provided from the power management integrated circuit; a computing module that is configured to perform an operation that is allocated by the host using a second voltage provided from the power management integrated circuit; and a performance monitoring module that is configured to monitor a first workload, a first amount of computation, and/or a first power state of the storage and/or is configured to monitor a second workload, a second amount of computation, and/or a second power state of the computing module and is configured to send a request for additional power to the host based on a monitoring result that is generated by the performance monitoring module.