Simple Analytic Performance Models for Streaming Data Applications Deployed on Diverse Architectures

Authors

Jonathan C. Beard, Washington University in St Louis
Roger D. Chamberlain, Washington University in St LouisFollow
Mark A. Franklin, Washington University in St LouisFollow

Document Type

Technical Report

Department

Computer Science and Engineering

Publication Date

2013

Filename

WUCSE-2013-2.pdf

Technical Report Number

WUCSE-2013-2

Abstract

Modern hardware is inherently heterogeneous. With heterogeneity comes multiple abstraction layers that hide underlying complex systems. While hidden, this complexity makes quantitative performance modeling a difficult task. Designers of high-performance streaming applications for heterogeneous systems must contend with unpredictable and often non-generalizable models to predict performance of a particular application and hardware mapping. This paper outlines a computationally simple approach that can be used to model the overall throughput and buffering needs of a streaming application on heterogeneous hardware. The model presented is based upon a hybrid maximum flow and decomposed discrete queueing model. The utility of the model is assessed using a set of real and synthetic benchmarks with model predictions compared to measured application performance.

Comments

Permanent URL: http://dx.doi.org/10.7936/K7TX3CK9

Recommended Citation

Beard, Jonathan C.; Chamberlain, Roger D.; and Franklin, Mark A., "Simple Analytic Performance Models for Streaming Data Applications Deployed on Diverse Architectures" Report Number: WUCSE-2013-2 (2013). All Computer Science and Engineering Research.
https://openscholarship.wustl.edu/cse_research/100