Metrics and benchmarking for parallel job scheduling

Dror G. Feitelson, Larry Rudolph

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

124 Scopus citations

Abstract

The evaluation of parallel job schedulers hinges on two things: the use of appropriate metrics, and the use of appropriate workloads on which the scheduler can operate. We argue that the focus should be on on-line open systems, and propose that a standard workload should be used as a benchmark for schedulers. This benchmark will specify distributions of parallelism and runtime, as found by analyzing accounting traces, and also internal structures that create different speedup and synchronization characteristics. As for metrics, we present some problems with slowdown and bounded slowdown that have been proposed recently.

Original languageEnglish
Title of host publicationJob Scheduling Strategies for Parallel Processing - IPPS/SPDP 1998 Workshop, Proceedings
EditorsDror G. Feitelson, Larry Rudolph
PublisherSpringer Verlag
Pages1-24
Number of pages24
ISBN (Print)3540648259, 9783540648253
DOIs
StatePublished - 1998
Event4th Workshop on Job Scheduling Strategies for Parallel Processing held in conjunction with IPPS/SPDP 1998 - Orlando, United States
Duration: 30 Mar 199830 Mar 1998

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1459
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference4th Workshop on Job Scheduling Strategies for Parallel Processing held in conjunction with IPPS/SPDP 1998
Country/TerritoryUnited States
CityOrlando
Period30/03/9830/03/98

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 1998.

Fingerprint

Dive into the research topics of 'Metrics and benchmarking for parallel job scheduling'. Together they form a unique fingerprint.

Cite this