Browse Prior Art Database

Main Storage Call Reduction

IP.com Disclosure Number: IPCOM000046575D
Original Publication Date: 1983-Aug-01
Included in the Prior Art Database: 2005-Feb-07
Document File: 1 page(s) / 11K

Publishing Venue

IBM

Related People

Crabtree, RP: AUTHOR [+2]

Abstract

A considerable overhead reduction is realized by calling a main storage supervisor for a block of storage elements at one time rather than calling for individual elements on an as-needed basis. In multiprogramming and multiprocessing systems it is frequently necessary to dynamically obtain a small element of storage from a main storage pool in order to queue a job from one task to another. After the job has been dequeued by a receiving task, the element of storage is available for return to the main storage pool. This obtaining and returning of elements presents an overhead burden in terms of calls to the main storage supervisor. One solution to reducing this burden is not to return the element where a pointer in common addressable memory is zero.

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

Page 1 of 1

Main Storage Call Reduction

A considerable overhead reduction is realized by calling a main storage supervisor for a block of storage elements at one time rather than calling for individual elements on an as-needed basis. In multiprogramming and multiprocessing systems it is frequently necessary to dynamically obtain a small element of storage from a main storage pool in order to queue a job from one task to another. After the job has been dequeued by a receiving task, the element of storage is available for return to the main storage pool. This obtaining and returning of elements presents an overhead burden in terms of calls to the main storage supervisor. One solution to reducing this burden is not to return the element where a pointer in common addressable memory is zero. In this case, the element is available for use the next time a service needs an element. When no element is available, a normal call to the main storage supervisor is made. Implementation of this technique requires either hardware disablement, or an instruction that does a fetch/store reference while blocking other central processing units during the sequence. Another solution is to call for a block of elements at one time. When all are being used (pointer is zero), and another element is needed, an additional block is called. Pointer management within the block is by the calling service, and the supervisor is freed for other chores.

1