Virtual Machine Graphics Power Management via Activity History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtualization systems underutilize graphics processing core performance by allocating it on a coarse basis to individual virtual machines, failing to manage power consumption effectively when multiple virtual machines with different rendering requirements share the core.
Innovation Solution
A power management scheme that monitors and controls the graphics processing core's power settings based on the activity history context of each virtual machine, optimizing performance by adjusting frequency and voltage settings on a finer grain level, considering temperature, power consumption, and estimated active time periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the graphics processing core is allocated to individual virtual machines on a coarse basis, then each virtual machine can independently control the graphics processing core, but the graphics processing core performance is underutilized and power consumption cannot be effectively managed
Solution Approach 1:
The patent segments the graphics processing core allocation into fine-grained time slices, dividing the core into multiple allocation units that can be dynamically assigned to different virtual machines. This allows the system to switch between VMs frequently while maintaining the appearance of independent control, thereby resolving the contradiction between ease of operation and productivity.
Solution Approach 2:
The patent implements dynamic allocation of the graphics processing core to virtual machines based on actual workload demands and priority levels. The allocation is not static but continuously adjusted through a hypervisor that monitors usage patterns and reassigns time slices accordingly, enabling both independent control and high utilization simultaneously.
2Ease of operation
If the graphics processing core is allocated to individual virtual machines on a coarse basis, then each virtual machine can independently control the graphics processing core, but power consumption is not optimized
Solution Approach 1:
The patent dynamically adjusts the power state of the graphics processing core based on which virtual machine is currently active and its power management requirements. The hypervisor coordinates with graphics drivers of active VMs to set appropriate frequency and voltage levels, ensuring power optimization while maintaining independent control capabilities.
Solution Approach 2:
The patent changes operational parameters (frequency, voltage, power state) of the graphics processing core based on the active virtual machine's requirements. Different VMs may require different power management profiles, and the system adjusts these parameters dynamically to optimize energy usage while preserving each VM's ability to control the core independently when allocated.
3Productivity
If fine-grain multiplexing of the graphics processing core between virtual machines is implemented, then performance is optimized and power consumption is reduced, but system complexity increases
Solution Approach 1:
The patent enables graphics drivers within each virtual machine to self-report their power management requirements and performance needs to the hypervisor. This self-service mechanism reduces the complexity of hypervisor oversight by allowing VMs to autonomously provide information about their needs, simplifying the coordination required for fine-grain multiplexing.
Solution Approach 2:
The patent implements feedback loops where graphics drivers of active virtual machines communicate their current workload and power requirements back to the hypervisor, which then adjusts the graphics processing core allocation and power settings accordingly. This automated feedback mechanism reduces manual oversight complexity while enabling optimized fine-grain multiplexing.
Data Source
AI summary
A method and apparatus determines an activity history context for each of a plurality of virtual machines sharing use of a graphics processing core. Each activity history context provides information related to a power setting of at least one engine of the graphics processing core during at least one prior use of the graphics processing core by the corresponding virtual machine. The method and apparatus controls a power setting of the at least one engine of the graphics processing core based on the activity history context corresponding to an active virtual machine using the graphics processing core.