Virtual Address Hashing for Uniform Page-Table Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In virtual memory environments, as physical memory increases, page-table entries become unevenly distributed, leading to inefficient mapping and translation, which prolongs execution times due to long overflow chains.
Innovation Solution
An extended virtual address hashing method is employed to distribute page-table entries uniformly throughout a large hash table, using a two-stage hashing algorithm where the primary hashing function is combined with an overflow hashing function that utilizes the virtual address to select chunks within the hash table, ensuring even distribution and minimizing traversal time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If physical memory size increases, then memory capacity improves, but page-table entry distribution becomes uneven leading to longer translation time
Solution Approach 1:
The hash table is divided into multiple chunks, where each chunk contains a portion of the page table entries. This segmentation allows the system to distribute entries across multiple smaller units rather than having them all in a single large table, improving cache utilization and reducing traversal time when physical memory size increases.
Solution Approach 2:
The patent implements a two-stage hashing algorithm where the first stage determines which chunk to access and the second stage finds the specific entry within that chunk. This local quality approach optimizes the search process by first narrowing down to a specific chunk (local region) and then searching within that smaller scope, rather than searching the entire hash table uniformly.
2Ease of manufacture
If standard hashing function is used, then implementation simplicity is maintained, but page-table entries are not uniformly distributed throughout the hash table
Solution Approach 1:
The hashing process is segmented into two distinct stages: first computing a preliminary hash value to determine the chunk index, then computing a secondary hash within that chunk. This segmentation transforms a single complex distribution problem into two simpler stages, achieving uniform distribution while maintaining implementation feasibility.
Solution Approach 2:
The patent introduces a dimensional approach to hashing by using the two-stage process: the first stage operates in one dimension (chunk selection) and the second stage operates in another dimension (entry selection within chunk). This dimensional decomposition allows uniform distribution across the entire hash table while keeping each individual stage relatively simple to implement.
Data Source
AI summary
Embodiments include methods, apparatus, and systems for virtual address hashing. One embodiment evenly distributes page-table entries throughout a hash table so applications do not generate a same hash index for mapping virtual addresses to physical addresses.