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

Internal Performance Measurement for Data Processing System

IP.com Disclosure Number: IPCOM000080478D
Original Publication Date: 1973-Dec-01
Included in the Prior Art Database: 2005-Feb-27
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Hedeman, WR: AUTHOR [+2]

Abstract

The internal performance of a data processing system is customarily measured, by running test job streams on the system and monitoring the performance by hardware devices. By a modification of this method, the internal performance of a proposed system can be determined by running a test job stream for the proposed system on an existing system. The results produced by the hardware monitors on the existing system are adjusted for differences in the performance of components of the two systems, to give the internal performance of the proposed system.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 62% of the total text.

Page 1 of 1

Internal Performance Measurement for Data Processing System

The internal performance of a data processing system is customarily measured, by running test job streams on the system and monitoring the performance by hardware devices. By a modification of this method, the internal performance of a proposed system can be determined by running a test job stream for the proposed system on an existing system. The results produced by the hardware monitors on the existing system are adjusted for differences in the performance of components of the two systems, to give the internal performance of the proposed system.

Improved storage technology is one example of these differences in the components of two data processing systems. Operations on main storage produce delays relating to access time (the time to supply data in response to a fetch request), the cycle time (the time between beginning successive storage operations), and delays produced when the store is busy and thus cannot immediately respond to an access request. For a nondestructive read store, the fetch cycle time is close to the access time whereas in a destructive read memory, the fetch cycle time is significantly longer than the access time. Thus, the performance of an individual fetch operation in the proposed system is adjusted for an improvement in access time, and the performance of successive fetch operations is further adjusted for the improvement in cycle time.

To calculate the effect of delays that ar...