Processor Power Limits for Isolated Virtual Partitions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional power management systems in computer processors fail to isolate power limits effectively across virtual partitions, leading to uneven performance degradation and the 'noisy neighbor' problem, where one virtual partition's power spike affects others undesirably.
Innovation Solution
Implementing a power management unit that generates separate power limit tables for each virtual partition based on the ratio of processing engines within each partition, using a hardware-software interface to manage power limits and isolate power use between different virtual partitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional power management systems manage power limits collectively for all virtual partitions, then overall power control is simplified, but power spikes from one virtual partition affect other partitions undesirably (noisy neighbor problem)
Solution Approach 1:
The patent divides the collective power management system into separate power limit tables for each virtual partition. Each partition receives a dedicated power limit allocation based on its processing engine ratio, creating isolated power control domains that prevent noisy neighbor effects while maintaining individual reliability.
2Reliability
If separate power limit tables are generated for each virtual partition based on processing engine ratio, then power isolation and performance optimization for active partitions are improved, but device complexity and management overhead increase
Solution Approach 1:
The patent applies local quality by allocating power limits according to the specific characteristics of each virtual partition's processing engine ratio. Each partition receives customized power allocation tailored to its workload requirements, optimizing local performance while the centralized power management unit coordinates overall power distribution.
Solution Approach 2:
The system dynamically adjusts power limit parameters for each virtual partition based on processing engine ratios and workload conditions. The power management unit modifies power allocation parameters in real-time to balance isolation requirements with system-wide power efficiency, managing complexity through adaptive parameter control.
3Ease of operation
If power limits are shared collectively across all virtual partitions, then management overhead is reduced, but performance variance increases due to uneven power distribution during power spikes
Solution Approach 1:
By segmenting power limits into separate tables for each virtual partition, the system maintains ease of operation through centralized management while ensuring each partition receives appropriate power allocation. This segmentation prevents performance degradation during power spikes by isolating power-consuming workloads to their designated partitions.
Solution Approach 2:
The power management unit monitors power consumption and workload conditions across virtual partitions, using feedback to dynamically adjust power limit allocations. This feedback mechanism ensures performance consistency by reallocating power limits based on actual usage patterns and partition-specific requirements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In an embodiment, a processor includes multiple processing engines and a power control unit. The power control unit is to receive a mapping of multiple virtual partitions to sets of the processing engines, and in response to a receipt of the mapping of multiple of virtual partitions: access a power limit table for the processor, and generate multiple virtual partition power limit tables based on the power limit table for the processor, where each virtual partition power limit table is associated with a different virtual partition. Other embodiments are described and claimed.