Most commercial multicomputers use space-slicing schemes in which each scheduling decision has an unknown impact on the future: should a job be scheduled, risking that it will block other larger jobs later, or should the processors be left idle for now in anticipation of future arrivals? This dilemma is solved by using gang scheduling, because then the impact of each decision is limited to its time slice, and future arrivals can be accommodated in other time slices. This added flexibility is shown to improve overall system utilization and responsiveness. Empirical evidence from using gang scheduling on a Cray T3D installed at Lawrence Livermore National Lab corroborates these results, and shows conclusively that gang scheduling can be very effective with current technology.
|Original language||American English|
|Title of host publication||Job Scheduling Strategies for Parallel Processing - IPPS 1997 Workshop, Proceedings|
|Editors||Dror G. Feitelson, Larry Rudolph|
|Number of pages||24|
|ISBN (Print)||3540635742, 9783540635741|
|State||Published - 1997|
|Event||3rd Workshop on Job Scheduling Strategies for Parallel Processing, IPPS 1997 - Geneva, Switzerland|
Duration: 5 Apr 1997 → 5 Apr 1997
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||3rd Workshop on Job Scheduling Strategies for Parallel Processing, IPPS 1997|
|Period||5/04/97 → 5/04/97|
Bibliographical notePublisher Copyright:
© Springer-Verlag Berlin Heidelberg 1997.