Browse Prior Art Database

Overlap Buffering

IP.com Disclosure Number: IPCOM000083007D
Original Publication Date: 1975-Mar-01
Included in the Prior Art Database: 2005-Feb-28
Document File: 2 page(s) / 20K

Publishing Venue

IBM

Related People

Melkun, BF: AUTHOR

Abstract

Processing programs often use multiple I/O buffers to allow overlap between I/O activity and processing data with a computer. Since buffers require storage space, a more efficient use may be made of storage space using the buffering arrangement shown in the diagram to provide the de sired overlap.

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 100% of the total text.

Page 1 of 2

Overlap Buffering

Processing programs often use multiple I/O buffers to allow overlap between I/O activity and processing data with a computer. Since buffers require storage space, a more efficient use may be made of storage space using the buffering arrangement shown in the diagram to provide the de sired overlap.

Each buffer A and B has two segments, 1 and 2 and 2 and 3, one of them (2) being a common segment. This buffering technique assumes that the reading and writing programs are aware of the technique being used. The programs treat the segments in a conventional manner by using blocking within segments. An example of a user of this technique is a compiler with a work data set. The Write Sequence for the technique is as follows: Buffer/Segment 1 2 A 3 A 2 B 1 B Operation F F W F C F W F C . . . n F means fill segment n, x W means write buffer x, x C The Forward Read Sequence is: B B 2 A 3 A 2 B 1 R C P R P C P R P . . . n P means process segment n, x R means read into buffer x, x C means check read for buffer x.

The Forward Read Sequence requires that segment sizes be identical (since a write of A is read into B). A backward read can be done (requiring a rule like `last buffer out is always A'), and provides an additional advantage of allowing segment 2 to be a different size than 1 and 3. This allows tuning its relative size to the ratio of processing to I/O wait an inverse relationship.

1

Page 2 of 2

2

[This page contains 2 pictures or other non-text obje...