Using Covariance Analysis as an aid to interpret the results of a performance measurement
SIGMETRICS '76 Proceedings of the 1976 ACM SIGMETRICS conference on Computer performance modeling measurement and evaluation
Studies in Markov models of computer systems
ACM '75 Proceedings of the 1975 annual conference
Computer performance evaluation: an empirical approach
ACM SIGMETRICS Performance Evaluation Review
Empirical investigation of the effectiveness of several computer performance evaluation tools
ACM SIGMETRICS Performance Evaluation Review
Experimenting with a trace-driven simulator of a batch computing system
SICOSIM4 Proceedings of the 4th annual symposium on SIGCOSIM: management and evalution of computer technology
Evaluating system changes under uncontrolled workloads: a case study
IBM Systems Journal
Approximate techniques for modeling the performance of complex systems
Computer Languages
Hi-index | 0.00 |
This paper describes how a regression model of the Research Center's APL system was constructed and used for evaluating system changes. The regression methodology that is generally used for performing such an analysis is believed to have been expanded by the introduction of multiple regression equations for estimating the cumulative distribution function of a performance variable. A model is developed which shows that the size of the workspace, but not the number of workspaces in core, significantly affects the performance of the Research Center's system. In particular, the model indicates that increasing the workspace size from 5 to 7 tracks causes 11 percent of all requests to be degraded by at least 0.4 seconds under mean load conditions.