Computing Publications

Publications Home » A new tool for the performance an...

A new tool for the performance analysis of massively parallel computer systems

Anton Stefanek, Richard Hayden, Jeremy T. Bradley

Conference or Workshop Paper
Eighth Workshop on Quantitative Aspects of Programming Languages (QAPL 2010), March 27-28, 2010, Paphos, Cyprus
March, 2010
Electronic Proceedings in Theoretical Computer Science
Volume 28
pp.159–181
ISSN 2075-2180
DOI 10.4204/EPTCS.28.11
Abstract

We present a new tool, GPA, that can generate key performance measures for very large systems. Based on solving systems of ordinary differential equations (ODEs), this method of performance analysis is far more scalable than stochastic simulation. The GPA tool is the first to produce higher moment analysis from differential equation approximation, which is essential, in many cases, to obtain an accurate performance prediction. We identify so-called switch points as the source of error in the ODE approximation. We investigate the switch point behaviour in several large models and observe that as the scale of the model is increased, in general the ODE performance prediction improves in accuracy. In the case of the variance measure, we are able to justify theoretically that in the limit of model scale, the ODE approximation can be expected to tend to the actual variance of the model.

Keywords
AESOP
PDF of full publication (1.3 megabytes)
(need help viewing PDF files?)
PDF of presentation slides (486 kilobytes)
BibTEX file for the publication
N.B.
Conditions for downloading publications from this site.
 

pubs.doc.ic.ac.uk: built & maintained by Ashok Argent-Katwala.