GPU Resource Management for Power Savings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multiple processor environments, unused processors like GPUs cannot be shut down without losing resources stored in their memories, leading to increased power consumption and reduced battery life in mobile systems.
Innovation Solution
A method and system that manage resources by copying them from a slave GPU to a master GPU before transitioning the slave GPU to a sub-operational power state, ensuring resources are preserved and up-to-date when the slave GPU returns to a powered state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If unused GPUs are shut down to save power, then power consumption is reduced, but resources stored in their memories are lost or become out-of-date
Solution Approach 1:
The system performs a preliminary action by copying resources from the slave GPU's memory to the master GPU's memory before shutting down the slave GPU. This ensures that when the slave GPU is later restarted, its resources are already preserved and up-to-date in the master GPU's memory, eliminating the need to keep the slave GPU powered on continuously.
Solution Approach 2:
The invention uses the copying principle by creating a copy of the resources from the slave GPU's memory and transferring them to the master GPU's memory. This copy allows the master GPU to maintain access to the resources even when the slave GPU is shut down, resolving the contradiction between power savings and resource accessibility.
2Loss of information
If unused GPUs remain powered to maintain resource accessibility, then resource accessibility is maintained, but power consumption increases and battery life is reduced
Solution Approach 1:
The system performs a preliminary action by copying resources from the slave GPU's memory to the master GPU's memory before shutting down the slave GPU. This ensures that when the slave GPU is later restarted, its resources are already preserved and up-to-date in the master GPU's memory, eliminating the need to keep the slave GPU powered on continuously.
Solution Approach 2:
The invention uses the copying principle by creating a copy of the resources from the slave GPU's memory and transferring them to the master GPU's memory. This copy allows the master GPU to maintain access to the resources even when the slave GPU is shut down, resolving the contradiction between power savings and resource accessibility.
3Loss of information
If resources are copied before shutting down, then resource accessibility is maintained, but additional processing time and resources are required
Solution Approach 1:
The system performs a preliminary action by copying resources from the slave GPU's memory to the master GPU's memory before shutting down the slave GPU. This ensures that when the slave GPU is later restarted, its resources are already preserved and up-to-date in the master GPU's memory, eliminating the need to keep the slave GPU powered on continuously.
Solution Approach 2:
The invention uses the copying principle by creating a copy of the resources from the slave GPU's memory and transferring them to the master GPU's memory. This copy allows the master GPU to maintain access to the resources even when the slave GPU is shut down, resolving the contradiction between power savings and resource accessibility.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of managing resources is provided. The method includes identifying a resource associated with a processor responsive to an impending transition, and copying the identified resource from a memory associated with the GPU or to the memory associated with the GPU.