Browse Prior Art Database

Cache Prefetching With Chaining

IP.com Disclosure Number: IPCOM000043722D
Original Publication Date: 1984-Sep-01
Included in the Prior Art Database: 2005-Feb-05
Document File: 1 page(s) / 11K

Publishing Venue

IBM

Related People

Giraud, FA: AUTHOR [+4]

Abstract

Any cache prefetching algorithm can be improved using the following scheme which creates a chaining effect. The scheme prompts the prefetching algorithm and indicates how the prefetched line should be placed in the cache as follows: 1. The algorithm is activated each time a cache reference is made to a non-MRU (most recently used) line of the cache. 2. The prefetch line is inserted into the cache with the status of non-MRU. MRU, which signifies most recently used, refers to a set of age-use status indicators maintained by the cache. These age-use indicators are set and reset within predefined subsets of the cache and indicate within that subset which line was most recently used, second most recently used, etc.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Cache Prefetching With Chaining

Any cache prefetching algorithm can be improved using the following scheme which creates a chaining effect. The scheme prompts the prefetching algorithm and indicates how the prefetched line should be placed in the cache as follows: 1. The algorithm is activated each time a cache reference is made to a non-MRU (most recently

used) line of the cache. 2. The prefetch line is inserted into the cache with the status of non-MRU. MRU, which signifies most recently used, refers to a set of age-use status indicators maintained by the cache. These age-use indicators are set and reset within predefined subsets of the cache and indicate within that subset which line was most recently used, second most recently used, etc. An access to the subset will reset these indicators if the line is not the MRU line of the subject. Since the prompting of the prefetch mechanism is based on MRU change only, the requirement for any additional bit-per-line to distinguish prefetched but unused lines is eliminated.

1