Browse Prior Art Database

Scatter Read

IP.com Disclosure Number: IPCOM000051643D
Original Publication Date: 1981-Feb-01
Included in the Prior Art Database: 2005-Feb-10
Document File: 2 page(s) / 70K

Publishing Venue

IBM

Related People

Hung, CY: AUTHOR [+2]

Abstract

A computer system as set forth in [*] has a slower operating master channel which can perform functions of the main channel so as to eliminate the need for a dedicated initial microprogram load (IMPL) device. When the master channel reads records from the save/restore file 5 (Fig. 1), it must either read the records contiguously during a number of revolutions or overrun occurs if an attempt is made to read the contiguous records in a single revolution. The present arrangement reduces the number of revolutions by reading the records in a scattered manner.

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 74% of the total text.

Page 1 of 2

Scatter Read

A computer system as set forth in [*] has a slower operating master channel which can perform functions of the main channel so as to eliminate the need for a dedicated initial microprogram load (IMPL) device. When the master channel reads records from the save/restore file 5 (Fig. 1), it must either read the records contiguously during a number of revolutions or overrun occurs if an attempt is made to read the contiguous records in a single revolution. The present arrangement reduces the number of revolutions by reading the records in a scattered manner.

When reading a data block which consists of four data records, it normally requires an average of 3+4 over N revolutions for reading the records contiguously, where N is the number of data records per track. With the present scatter read arrangement, the records are still arranged contiguously on the disk but records one and three are read and stored in a data buffer (Fig. 2) during a first revolution and records two and four are read and stored during a second revolution. Record two provides the necessary latency time required for the master channel to store data from record one and set up the channel request for record three. During the second revolution, record three provides this same latency time for storing record two and setting up the request for record four. One data block can be read in 1+4 over N revolutions, where N is the number of data records per track. This yields a data rate improvem...