The interaction of parallel and sequential workloads on a network of workstations
Proceedings of the 1995 ACM SIGMETRICS joint international conference on Measurement and modeling of computer systems
Effective distributed scheduling of parallel workloads
Proceedings of the 1996 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Coordinating parallel processes on networks of workstations
Journal of Parallel and Distributed Computing
The Hector Distributed Run-Time Environment
IEEE Transactions on Parallel and Distributed Systems
A Case for NOW (Networks of Workstations)
IEEE Micro
An Empirical Investigation of Load Indices for Load Balancing Applications
Performance '87 Proceedings of the 12th IFIP WG 7.3 International Symposium on Computer Performance Modelling, Measurement and Evaluation
Modeling and characterizing parallel computing performance on heterogeneous networks of workstations
SPDP '95 Proceedings of the 7th IEEE Symposium on Parallel and Distributeed Processing
Hi-index | 0.00 |
Networks of workstations (NOWs) have become important and cost-effective parallel platforms for scientific computations. In practice, a NOW system is heterogeneous and non-dedicated. These two unique factors make scheduling policies on multiprocessor/multicomputer systems unsuitable for NOWs, but the coscheduling principle is still an important basis for parallel process scheduling in these environments. The main idea of this technique is to schedule the set of tasks composing a parallel application at the same time, to increase their communication performance. In this article we present an explicit coscheduling algorithm implemented in a Linux NOW, of PVM distributed tasks, based on Real Time priority assignment. The main goal of the algorithm is to execute efficiently distributed applications without excessively damaging the response time of local tasks. Extensive performance analysis as well as studies of the parameters and overheads involved in the implementation demonstrated the applicability of the proposed algorithm.