Communications of the ACM
Adaptive Optimal Load Balancing in a Nonhomogeneous Multiserver System with a Central Job Scheduler
IEEE Transactions on Computers
Parallel machines scheduling with nonsimultaneous machine available time
Discrete Applied Mathematics
The PVM concurrent computing system: evolution, experiences, and trends
Parallel Computing - Special issue: message passing interfaces
A taxonomy of scheduling in general-purpose distributed computing systems
IEEE Transactions on Software Engineering
Process Mapping Given by Processor and Network Dynamic Load Prediction
Euro-Par '99 Proceedings of the 5th International Euro-Par Conference on Parallel Processing
Communication Kernel for High Speed Networks in the Parallel Environment LANDA-HSN
IWCC '99 Proceedings of the 1st IEEE Computer Society International Workshop on Cluster Computing
Models for Dynamic Load Balancing in a Heterogeneous Multiple Processor System
IEEE Transactions on Computers
PAPER: influence of deterministic customers in time sharing scheduler
ACM SIGOPS Operating Systems Review
| Hi-index | 0.00 |
This paper describes a tool called Network-Analyser that increases the efficiency of clusters for parallel computing. Two important features are provided by this tool. First, it is organized in three hierarchical levels to collect information on workstations state. Secondly it provides several taskallo cation policies. In this paper a new policy that uses both instantaneous and average workstations loads to estimate tasks expected execution times, is developed. The multiprocessors workstation model used is based on round robin queuing system. A differential equation is established to describe transient state of the queuing system. Then an integral equation allows to estimate the expected execution time of regular parallel applications by mixing the deterministic processes of the parallel application with the stochastic load of the workstation.