Large Memory Page Activity Identification for VM Performance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In computer systems with virtual machines, the indiscriminate demotion of large pages to small pages can lead to performance degradation due to increased mapping time for active pages, especially when memory is scarce, as existing methods lack the ability to identify inactive large pages effectively.
Innovation Solution
A method to identify activity levels of large pages by selecting a large page, updating mappings to small pages, tracking accesses, and determining the activity level based on access patterns, allowing for intelligent selection of pages for decomposition and memory reclamation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If large pages are used to decrease TLB misses, then performance is improved, but memory waste increases when physical memory is scarce
Solution Approach 1:
The patent dynamically adjusts page size based on activity levels. Large pages are used for inactive memory regions to reduce TLB misses, while active regions maintain small page granularity. This dynamic adaptation allows the system to optimize between TLB performance and memory utilization based on actual access patterns.
Solution Approach 2:
The patent applies different page sizes to different regions of memory based on their activity characteristics. Inactive regions use large pages to minimize TLB misses, while active regions use small pages to enable fine-grained memory management and prevent waste. This local differentiation resolves the contradiction by applying the appropriate page size where needed.
2Loss of substance
If large pages are randomly selected for demotion to small pages, then memory pressure is reduced, but performance degradation occurs due to increased mapping time for active pages
Solution Approach 1:
The patent uses activity level monitoring as feedback to guide page demotion decisions. By tracking access patterns and identifying inactive large pages, the system receives feedback about which pages are safe to demote without impacting performance. This feedback mechanism prevents random demotion of active pages while still reducing memory pressure through targeted demotion of inactive pages.
Solution Approach 2:
The patent performs preliminary identification of inactive pages before demotion occurs. By monitoring activity levels and identifying candidate pages in advance, the system ensures that only truly inactive pages are demoted to small pages. This preliminary action prevents performance degradation by avoiding demotion of active pages while still achieving memory pressure reduction.
Data Source
AI summary
Large pages that may impede memory performance in computer systems are identified. In operation, mappings to selected large pages are temporarily demoted to mappings to small pages and accesses to these small pages are then tracked. For each selected large page, an activity level is determined based on the tracked accesses to the small pages included in the large page. By strategically selecting relatively low activity large pages for decomposition into small pages and subsequent memory reclamation while restoring the mappings to relatively high activity large pages, memory consumption is improved, while limiting performance impact attributable to using small pages.