Page List Management for Accurate Hot-Cold Page Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing LRU mechanisms struggle with inaccurate classification of pages as hot or cold, leading to high CPU overhead and inefficient resource management due to frequent reverse mapping and high CPU usage, especially as memory capacity increases.
Innovation Solution
Implement a page list management method and device using a Compute Express Link (CXL) device to collect detailed statistical information on page access through an access tracking module, classify pages as hot or cold based on threshold values, and adjust the LRU list accordingly, reducing CPU overhead and improving classification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequent reverse mapping is performed to accurately classify pages as hot or cold, then classification accuracy is improved, but CPU overhead increases
Solution Approach 1:
The patent introduces a page list as an intermediary data structure that stores page access statistics. Instead of performing frequent reverse mapping operations to classify pages, the system uses the pre-computed page list to determine hot and cold pages, thereby reducing CPU overhead while maintaining classification accuracy.
Solution Approach 2:
The patent performs preliminary computation by generating and maintaining a page list that records page access patterns. This pre-computed information is then used to classify pages as hot or cold without requiring frequent reverse mapping operations, thus reducing real-time CPU overhead.
2Productivity
If LRU list is frequently adjusted based on page access patterns, then memory management efficiency is improved, but system complexity increases
Solution Approach 1:
The patent segments the memory management task into two independent parts: (1) collecting page access statistics and generating a page list, and (2) using the page list to adjust the LRU list. This segmentation allows each part to be optimized independently, improving overall efficiency while managing complexity.
Solution Approach 2:
The patent uses a simplified representation (page list) that copies essential information about page access patterns without requiring complex reverse mapping operations. This simplified data structure enables efficient LRU list adjustment while reducing system complexity.
3Measurement precision
If detailed statistical information on page access is collected, then classification accuracy is improved, but CPU usage increases
Solution Approach 1:
The patent implements a mechanism where the system collects page access statistics and maintains a page list autonomously. The page list is updated based on natural page access patterns without requiring intensive CPU processing, allowing the system to serve itself in collecting and utilizing access information.
Data Source
AI summary
A method includes obtaining statistical information about page access by scanning a page table entry related to a page table, classifying a page in the page table entry as a hot page or a cold page based on the statistical information, generating a page list comprising the classified hot page or the classified cold page, and transmitting the page list to a host device, wherein the page list is used to adjust a least recently used (LRU) list.