Browse Prior Art Database

Reference Associative Cache Mapping

IP.com Disclosure Number: IPCOM000051496D
Original Publication Date: 1981-Feb-01
Included in the Prior Art Database: 2005-Feb-10
Document File: 3 page(s) / 49K

Publishing Venue

IBM

Related People

Mathis, JR: AUTHOR [+3]

Abstract

Disclosed is a Reference Associative Cache Mapping algorithm which uses tag lines to supply information from the Requestor (processor, channel, etc.) which describe the nature of the request. This information is used to make a single access into the directory to determine if the needed data is in the CACHE.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 3

Reference Associative Cache Mapping

Disclosed is a Reference Associative Cache Mapping algorithm which uses tag lines to supply information from the Requestor (processor, channel, etc.) which describe the nature of the request. This information is used to make a single access into the directory to determine if the needed data is in the CACHE.

The number of different reasons for a main storage reference determines the number of CACHE pages required to support the system. Since the number of reasons for storage access is limited, the number of CACHE pages is likewise limited. Thus, the implementation cost of the Reference Associative algorithm is reduced due to the number of CACHE pages. The reasons for an access would include: - instruction fetch, - operand (1 and/or 2) data fetch, and - channel fetch.

For each access, the Requestor will supply the main storage address and a "reference ID" (tag). The reference ID is used to determine which one of the several CACHE pages is to be examined to determine if the data is in CACHE. Thus, a SINGLE directory access and a SINGLE address comparison are required to determine if the data is in CACHE.

The Reference Associative algorithm is designed to provide maximum benefit for read operations of sequential data and/or instructions. For nonsequential read accesses the possibility of a page fault occurring and the system performance being degraded is increased for the Reference Associative algorithm over the more conventional algorithms. The frequency of occurrence of page faults for the Reference Associative algorithm can be reduced by increasing the size of the CACHE pages, as is done for other algorithms. For a write access, the appropriate CACHE page is updated immediately. The choice of immediate or deferred update of main storage is dependent on the requirements of a given system implementation.

The Reference A...