An Experimental Approach for Job Scheduling Optimization to Improve the System Usage Efficiency

R. Ito (Japan)

Keywords

Queuing system, job scheduling, resource management, optimization, PC clusters and parameter settings

Abstract

We propose a scientific approach to optimize the system usage with parameter configuration, mainly by developing simulator software for NQS with which the system usage pattern could be virtually simulated with various parameter configurations, consequently makes us possible to see how optimally the system is used with the current configuration. Furthermore this scientific approach is useful to investigate a property of each parameter.

Important Links:



Go Back