Computer Simulation and Modelling
Computer Simulation and Modelling
Using Processor-Cache Affinity Information in Shared-Memory Multiprocessor Scheduling
IEEE Transactions on Parallel and Distributed Systems
Task scheduling performance in distributed systems with time varying workload
Neural, Parallel & Scientific Computations
An Integrated Approach to Parallel Scheduling Using Gang-Scheduling, Backfilling, and Migration
IEEE Transactions on Parallel and Distributed Systems
Gang Scheduling with Memory Considerations
IPDPS '00 Proceedings of the 14th International Symposium on Parallel and Distributed Processing
Efficient scheduling of scientific workflows in a high performance computing cluster
CLADE '08 Proceedings of the 6th international workshop on Challenges of large applications in distributed environments
Performance Implications of Cache Affinity on Multicore Processors
Euro-Par '08 Proceedings of the 14th international Euro-Par conference on Parallel Processing
Journal of Systems and Software
Scheduling Gangs with Different Distributions in Gangs' Degree of Parallelism in a Multi-Site System
BCI '09 Proceedings of the 2009 Fourth Balkan Conference in Informatics
Gang scheduling in a two-cluster system with critical sporadic jobs and migrations
SPECTS'09 Proceedings of the 12th international conference on Symposium on Performance Evaluation of Computer & Telecommunication Systems
Enhancements to the decision process of the self-tuning dynp scheduler
JSSPP'04 Proceedings of the 10th international conference on Job Scheduling Strategies for Parallel Processing
The Journal of Supercomputing
Hi-index | 0.00 |
Gang scheduling is an efficient resource management scheme for distributed systems which combines elements of time sharing and space sharing. It is a suitable technique particularly in the case when parallel tasks have to be running concurrently to make progress in communication. This paper studies the impact on scheduling performance when dynamically generated sequential gangs exist in the workload. In the case of sequential gangs, a subsequent gang can be dynamically generated after the execution of the initial gang based on affinity information which resides on the caches of the previously seized processors. The performance of different gang-scheduling algorithms is examined for various cases of workload compositions which range from cases with a low demand for dynamically generated gangs to cases with a high ratio of sequential gangs to solitary gangs. A simulation model is implemented to address associated performance issues. Copyright © 2011 John Wiley & Sons, Ltd.