Multi-Processor Model Weight Allocation for Stable Task Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for task processing using multiple processing units in electronic devices face performance degradation due to insufficient available computing resources, leading to reduced efficiency.
Innovation Solution
A method and device that configure first model weights for each processing unit based on computing resource information, enabling synchronous execution of tasks by multiple processing units, thereby optimizing resource allocation and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple processing units allocate fixed computing resources to perform model inference, then task processing can be performed using parallel computing, but performance degrades when a processing unit receives a new task due to insufficient available computing resources
Solution Approach 1:
The patent implements dynamic resource allocation by allowing processing units to flexibly adjust their computing resource allocation based on real-time task requirements and available resources. Instead of fixed allocation, the system dynamically configures computing resources for each task, enabling processing units to adapt to varying workloads and maintain stable performance across different scenarios.
2Power
If multiple processing units are used to perform model inference, then computing power is increased, but resource allocation is not optimized leading to reduced task processing efficiency
Solution Approach 1:
The patent applies local quality by configuring different computing resource allocations for different processing units based on their specific capabilities and the requirements of individual tasks. Each processing unit receives a customized resource configuration rather than a uniform allocation, optimizing the match between computing power and task demands to improve overall processing efficiency.
Solution Approach 2:
The system changes the parameter of computing resource allocation from fixed to variable, allowing dynamic adjustment of resource distribution based on task characteristics and processing unit performance. This parameter change enables the system to optimize computing power utilization and improve task processing efficiency across multiple processing units.
3Device complexity
If fixed computing resources are allocated to each processing unit, then resource management is simplified, but task processing accuracy and speed deteriorate due to insufficient available computing resources
Solution Approach 1:
The patent implements dynamic resource allocation that automatically adjusts computing resource distribution based on real-time system state and task requirements. This dynamic approach maintains task processing accuracy and speed by ensuring adequate resources are available, while the system manages the complexity through automated configuration rather than manual intervention.
Data Source
AI summary
A task processing method includes: in response to obtaining a target processing task, obtaining computing resource information of multiple processing units of an electronic device; and configuring first model weights of each of the multiple processing units based on the computing resource information, so that the multiple processing units synchronously execute the target processing task based on their respective first model weights.