Showing papers on "Benchmark (computing) published in 1973"
TL;DR: Measurements made on one multi‐terminal system are described and an attempt made to justify the resulting benchmark, which of necessity is changed in detail in order to run on the new system.
Abstract: A basic premise of all benchmarks is that the pattern of system use does not change when moving work from one system to another. Measurements made on one multi-terminal system are described and an attempt made to justify the resulting benchmark, which of necessity is changed in detail in order to run on the new system. © 1973 Wiley Periodicals, Inc.
8 citations
IBM1
TL;DR: APET is an approach for manufacturing synthetic benchmark or yardstick programs that exercise the computer system that measure selected hardware or software functions, as they interact within an application under real operating conditions.
Abstract: This article describes a technique for controlling computer system activity in order to measure selected hardware or software functions, as they interact within an application under real operating conditions. APET is an approach for manufacturing synthetic benchmark or yardstick programs that exercise the computer system. © 1973 Wiley Periodicals, Inc.
5 citations
TL;DR: Policy suggestions are considered for the potential user of a commercial time-shared system to improve performance in terms of both faster turnaround time and lower processing costs.
Abstract: This paper presents the results of an analysis of a commercial time-shared system. System performance in terms of response times and processing costs were studied for a mix of benchmark programs. The number of users on-line during various periods throughout the day were studied with the result that a significantly difTerent pattern of user behaviour hour to hour throughout the day was found. The benchmark performance was found to vary considerably in conjunction with this varying utilization pattern. Policy suggestions are considered for the potential user of such a system to improve performance in terms of both faster turnaround time and lower processing costs. The methodology of the research is extended so that a potential purchaser could implement relevant portions of this research in his own analysis of the commercial time-shared environments.
2 citations