Dynamic Memory Reconfiguration to Delay Page Swapping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing virtual memory management systems face performance overhead due to the invocation of the least recently used (LRU) algorithm for page swapping, which consumes processing cycles and introduces latency from I/O operations when determining which pages to swap out, especially during temporary memory bursts.

Innovation Solution

Implement a dynamic main memory reconfiguration method where the virtual memory manager augments or reduces real memory allocations instead of triggering page swapping when free space falls below a threshold, delaying or avoiding page swapping by adding or removing logical memory blocks to satisfy memory access directives.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the LRU algorithm is invoked to determine which memory pages to swap out, then memory virtualization can ensure data availability, but processing overhead and performance latency increase due to cycling through the page frame table and I/O operations

Engineering Contradiction:
Improvedata availabilityVSAvoidprocessing overhead
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The system pre-identifies stealable pages and maintains a free list of candidate pages before swapping is needed. This preliminary action allows the system to quickly select pages for swapping without cycling through the entire page frame table during memory pressure events, thus reducing processing overhead while ensuring data availability.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary data structure (free list) that mediates between the page frame table and the swapping decision-making process. This free list acts as a pre-filtered set of candidate pages, reducing the computational burden on the virtual memory manager during swapping operations.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If page swapping is performed frequently to maintain memory allocation, then memory access requirements can be satisfied, but system performance deteriorates due to I/O latency to fixed storage

Engineering Contradiction:
Improvememory allocation flexibilityVSAvoidI/O latency
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system performs preliminary identification and categorization of pages into stealable and non-stealable groups, maintaining a ready list of swap candidates. This preparation reduces the need for frequent, urgent swapping operations by having pages pre-selected for potential swapping, thereby reducing I/O latency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The free list structure enables the system to self-manage swap candidates without requiring intensive intervention from the virtual memory manager during memory pressure events. The pre-organized free list allows the system to service memory allocation requests more efficiently by having swap candidates already identified and ready.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS8751724B2Dynamic memory reconfiguration to delay performance overhead
Publication Date: 2014.06.10 LENOVO GLOBAL TECHNOLOGIES SWITZERLAND INTERNATIONAL GMBH
  • US8751724B2 patent drawing
  • US8751724B2 patent drawing

AI summary

Embodiments of the present invention provide a method, system and computer program product for dynamic main memory reconfiguration in virtual memory management. In an embodiment of the invention, a method for dynamic main memory reconfiguration in virtual memory management can include receiving a memory access directive in a host computer, determining a low free space condition in a memory allocation to satisfy the memory access directive, augmenting the memory allocation with a mapping to additional memory in the host computer in lieu of page swapping in response to the low free space condition, and satisfying the memory access directive. Additionally, the method can include determining an excess free space condition in the memory allocation and removing from the memory allocation a selection of allocated memory in the host computer.