Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Automatic Partial Deletion of High Speed Buffer Entries

IP.com Disclosure Number: IPCOM000082500D
Original Publication Date: 1974-Dec-01
Included in the Prior Art Database: 2005-Feb-28
Document File: 2 page(s) / 40K

Publishing Venue

IBM

Related People

Ahearn, TP: AUTHOR [+4]

Abstract

A mechanism is described which automatically deletes portions of a high-speed buffer on the occurrence of a buffer data error or an address compare error. Although only two types of errors are described, the choice was arbitrary and other types of errors may be included in the disabling process.

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 89% of the total text.

Page 1 of 2

Automatic Partial Deletion of High Speed Buffer Entries

A mechanism is described which automatically deletes portions of a high- speed buffer on the occurrence of a buffer data error or an address compare error. Although only two types of errors are described, the choice was arbitrary and other types of errors may be included in the disabling process.

The logic shown permits a partial disabling of a high-speed buffer, or cache, and associated address index array and compare circuitry. The arrangement comprises eight rows of 64 entries. Each entry in the index array being a block address identifier, and each entry in the cache being a block of data.

When one of the specified errors occur, the row identifier of the buffer address is decoded and an associated row delete latch is enabled, thereby disabling a portion of the array and its corresponding compare logic. The state of the row delete latch is fed into the replacement algorithm, thereby inhibiting any of the row's member blocks from being referenced. Thus, the address array, the row compare logic and the data array card, any of which could have caused one of the errors, are disabled and eliminate the reoccurrence of the same (or related) buffer check and at the same time allows the system to continue processing.

In order to maintain the system integrity, a degradation latch is set and stored in the machine check interruption code. The delete latches are also displayed for visual inspection and extended logout...