Comparing performance heatmaps

David Krakov, Dror G. Feitelson*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

9 Scopus citations

Abstract

The performance of parallel job schedulers is often expressed as an average metric value (e.g. response time) for a given average load. An alternative is to acknowledge the wide variability that exists in real systems, and use a heatmap that portrays the distribution of jobs across the performance load space. Such heatmaps expose a wealth of details regarding the conditions that occurred in production use or during a simulation. However, heatmaps are a visual tool, lending itself to high-resolution analysis of a single system but not conducive for a direct comparison between different schedulers or environments. We propose a number of techniques that allow to compare heatmaps. The first two treat the heatmaps as images, and focus on the differences between them. Two other techniques are based on tracking how specific jobs fare under the compared scenarios, and drawing underlying trends. This enables a detailed analysis of how different schedulers affect the workload, and what leads to the observed average results.

Original languageAmerican English
Title of host publicationJob Scheduling Strategies for Parallel Processing - 17th International Workshop, JSSPP 2013, Revised Selected Papers
PublisherSpringer Verlag
Pages42-61
Number of pages20
ISBN (Print)9783662437780
DOIs
StatePublished - 2014
Event17th International Workshop on Job Scheduling Strategies for Parallel Processing, JSSPP 2013 - Boston, MA, United States
Duration: 24 May 201424 May 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8429 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th International Workshop on Job Scheduling Strategies for Parallel Processing, JSSPP 2013
Country/TerritoryUnited States
CityBoston, MA
Period24/05/1424/05/14

Fingerprint

Dive into the research topics of 'Comparing performance heatmaps'. Together they form a unique fingerprint.

Cite this