Browse Prior Art Database

Pageable Memory Element Design

IP.com Disclosure Number: IPCOM000049480D
Original Publication Date: 1982-Jun-01
Included in the Prior Art Database: 2005-Feb-09
Document File: 2 page(s) / 70K

Publishing Venue

IBM

Related People

Parks, LS: AUTHOR [+4]

Abstract

This article describes a design to improve the performance of a small processor (microprocessor) on a set of applications which are regular. By regular we mean those in which a large part of the processing required can be viewed as the manipulation of an array of data such that the result is an array, and the manipulation can be decomposed into transformations of rows of the data. For example: F(DA(1))=DA(2) F(D)=(f(i))(f(2))......(f(n))(D) where (f(i)) maps rows to rows.

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 68% of the total text.

Page 1 of 2

Pageable Memory Element Design

This article describes a design to improve the performance of a small processor (microprocessor) on a set of applications which are regular. By regular we mean those in which a large part of the processing required can be viewed as the manipulation of an array of data such that the result is an array, and the manipulation can be decomposed into transformations of rows of the data. For example: F(DA(1))=DA(2)

F(D)=(f(i))(f(2))......(f(n))(D)

where (f(i)) maps rows to rows.

Some of the applications that may be processed in this fashion are some arithmetic problems, translations, encryptions and logic simulations.

This scheme, as shown in Figs. 1 and 2, is made up of a

Processor PO, two Data Arrays C and G, a Pageable Array J, a Memory O and an Incrementer P. (In the VLSI environment some or all of these may be on the same chip). Under the control of the Processor, data is read into one of the Data Arrays C and G and a "function (f(i))" is read into the Pageable Array J. Each line of the Data Array is successively sent as an address to the Pageable Array J and the "data" read from the Pageable Array J is stored in the second Data Array. The new "data" may now be used to address the Pageable Array.

This process may be repeated as desired. The Processor may query a Data Array and conditionally initiate one or a sequence of the following processes. 1. Transfer the new "data" from the Data Array to the Pageable

Array.

2. Page in a new func...