TLB Search Optimization via Page Size Presence Vector
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current translation lookaside buffer (TLB) search operations are inefficient due to the need for separate lookups for each page size, leading to slowed throughput and increased power consumption.
Innovation Solution
Implementing a TLB that uses a page size presence vector and prioritizer, such as an age matrix, to determine which page sizes to search and in what order, reducing the number of search operations required to find a matching entry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the TLB performs separate lookups for each page size to ensure accurate translation, then translation accuracy is maintained, but TLB throughput decreases and power consumption increases
Solution Approach 1:
The patent applies preliminary action by pre-establishing an age matrix that stores historical page size usage frequencies before translation requests arrive. When a translation request comes in, the system queries the pre-computed age matrix to immediately obtain the most likely page size without performing multiple sequential lookups. This pre-computation approach maintains translation accuracy while significantly improving throughput by eliminating redundant lookup operations.
Solution Approach 2:
The patent introduces an age matrix as an intermediary data structure between the page table and the translation request. This intermediary stores pre-analyzed page size probability information, allowing the system to quickly determine the most likely page size without directly querying multiple page table entries. The age matrix acts as a mediator that reduces the complexity of multi-page-size lookups while preserving translation accuracy.
2Reliability
If the TLB performs multiple lookups for different page sizes to find the correct translation, then complete search coverage is achieved, but power consumption increases
Solution Approach 1:
The patent applies partial action by performing only the necessary number of lookups to achieve translation. Instead of exhaustively searching all possible page sizes, the system uses the age matrix to identify the most probable page size first and performs lookups in decreasing order of probability. Once a match is found, the search terminates, avoiding unnecessary additional lookups. This approach maintains search completeness for valid translations while reducing power consumption by eliminating redundant search operations.
3Adaptability or versatility
If the TLB supports multiple page sizes with separate lookups to maintain versatility, then adaptability to different page sizes is achieved, but device complexity increases
Solution Approach 1:
The patent applies universality by designing a single age matrix data structure that handles multiple page sizes through a unified approach. Rather than implementing separate lookup mechanisms for each page size, the system uses one universal age matrix that stores probability information for all supported page sizes. This universal structure maintains adaptability to different page sizes while reducing device complexity by eliminating the need for multiple specialized lookup paths.
Data Source
AI summary
Systems, apparatuses, and methods for limiting translation lookaside buffer (TLB) searches using active page size are described. A TLB stores virtual-to-physical address translations for a plurality of different page sizes. When the TLB receives a command to invalidate a TLB entry corresponding to a specified virtual address, the TLB performs, for the plurality of different pages sizes, multiple different lookups of the indices corresponding to the specified virtual address. In order to reduce the number of lookups that are performed, the TLB relies on a page size presence vector and an age matrix to determine which page sizes to search for and in which order. The page size presence vector indicates which page sizes may be stored for the specified virtual address. The age matrix stores a preferred search order with the most probable page size first and the least probable page size last.