Workload Action Space Reduction for Edge Device Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional techniques for optimizing workload execution in cloud and edge environments often rely on fixed hardware configurations, failing to account for the potential benefits of different hardware settings during various phases of a workload, leading to suboptimal performance and efficiency.
Innovation Solution
The approach involves identifying a group of hardware configurations or settings that can be dynamically applied by edge devices in real-time, reducing the hardware setting space to a smaller number of dominating settings or combinations, which are then used to improve workload performance and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional fixed hardware configurations are used for workload execution, then device complexity is reduced and ease of operation is improved, but workload performance and efficiency deteriorate
Solution Approach 1:
The patent implements dynamic hardware configuration by allowing edge devices to switch between different hardware settings during workload execution. The system transitions from static fixed configurations to dynamic adjustable configurations, enabling the hardware to adapt its state based on workload requirements while maintaining manageable complexity through automated control
Solution Approach 2:
The system changes hardware parameters (settings) dynamically during workload execution. Different hardware configurations are applied at different phases of workload execution to optimize performance. The system explores and identifies optimal parameter combinations without requiring manual intervention, thus improving productivity while managing complexity through automated parameter tuning
2Productivity
If comprehensive hardware setting exploration is performed to find optimal configurations, then workload performance is improved, but time and computational resources increase
Solution Approach 1:
The system performs preliminary exploration of hardware settings during idle periods or before workload execution. By pre-identifying and caching optimal or near-optimal hardware configurations, the system reduces the time required during actual workload execution. The action determination circuitry prepares potential actions in advance, so when a workload arrives, optimal settings are already determined or can be quickly selected from pre-evaluated options
Solution Approach 2:
Instead of exhaustively exploring all possible hardware setting combinations, the system performs partial exploration focusing on the most promising configurations. The action determination circuitry identifies and evaluates a subset of critical hardware settings rather than all possible combinations, achieving good enough performance with significantly reduced time and resource investment
3Adaptability or versatility
If dynamic hardware configuration is implemented, then adaptability to different workload phases is improved, but device complexity increases
Solution Approach 1:
The edge device performs self-configuration by automatically determining optimal hardware settings without external intervention. The action determination circuitry autonomously evaluates workload requirements and selects appropriate hardware configurations. This self-service approach enables high adaptability while managing complexity internally without requiring complex external configuration management infrastructure
Solution Approach 2:
The system implements feedback mechanisms where the performance of different hardware configurations is monitored and used to guide future configuration selections. The action determination circuitry learns from past performance data to make better configuration decisions, enabling adaptability through data-driven automation rather than complex rule-based systems
Data Source
AI summary
An example apparatus includes at least one programmable circuit to analyze workload runs for a plurality of combinations of enabled setting to determine a subset of the plurality of combinations that satisfy a target performance metric; run a workload for a second combination of enabled settings to generate a result, the second combination combining enabled settings from two or more of the subset of the plurality of combinations; analyze the result to determine the second combination satisfies the target performance metric; and deploy the second combination and the subset of the plurality of combinations to a device to process a second workload using at least one of the second combination of the subset of the plurality of combinations.