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

Non-Continuous Memory Hardware Assist Algorithm

IP.com Disclosure Number: IPCOM000039077D
Original Publication Date: 1987-Apr-01
Included in the Prior Art Database: 2005-Feb-01
Document File: 3 page(s) / 44K

Publishing Venue

IBM

Related People

Eikill, RG: AUTHOR [+3]

Abstract

Certain CPU hardware is capable of addressing a main storage configuration that is continuous and contains up to four memory cards. With the introduction of a 512 Kb memory card to the memory card family of the hardware adding to the existing 128 Kb and the 256 Kb cards, means is needed for mixing these three memory cards in the same main storage configuration. The CPU hardware cannot perform this task alone, and this algorithm is a software tool to take care of this problem that allows the CPU to address a non-continuous main storage. This software tool restrains the CPU hardware from trying to address a segment of main storage that was not there (these segments are called missing pages).

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 52% of the total text.

Page 1 of 3

Non-Continuous Memory Hardware Assist Algorithm

Certain CPU hardware is capable of addressing a main storage configuration that is continuous and contains up to four memory cards. With the introduction of a 512 Kb memory card to the memory card family of the hardware adding to the existing 128 Kb and the 256 Kb cards, means is needed for mixing these three memory cards in the same main storage configuration. The CPU hardware cannot perform this task alone, and this algorithm is a software tool to take care of this problem that allows the CPU to address a non-continuous main storage. This software tool restrains the CPU hardware from trying to address a segment of main storage that was not there (these segments are called missing pages). This presupposes a method to locate defective segments of memory (these segments are called bad pages) and maintain an active record of these segments so that the CPU will not try to access these bad pages. This algorithm constitutes a method for maintaining the same type of an active log for missing pages. The system software will allow only a specific number of bad pages to be present in the log before it will terminate operation. This number is the size of one card plus 18 Kb; this number is selected so the system may operate with one bad memory card. Since with various combinations of memory card sizes this number would be exceeded, there had to be a way to determine between bad and missing pages. This is done through the use of one byte in a UDT (unit definition table). Each card is assigned two bits to specify the size of that card. The way in which the value of this byte is determined is different in two systems. One system uses a manually configured program in which the user defines the configuration and the program translates this information into the one byte of data for the UDT. The other system uses an automatic configuration program in which the system determines the number of memory cards and the size of each card on the system and then updates the UDT with this information. In order to address all of the main storage space that is available on the system, the software must make it look to the CPU like main storage is a uniform distribution of the largest size memory ca...