Combining performance and priority for scheduling resizable parallel applications

• We describe a job scheduler that exploits malleable parallel applications.
• Processor allocations are based on job performance and on cluster workload.
• We present scheduling scenarios and algorithms that exploit malleability.
• Potential benefits when only some jobs are malleable and when job priority varies are studied.

We illustrate and evaluate the potential impact of dynamic resizability on parallel job scheduling. Our ReSHAPE framework includes a job scheduler that supports dynamic resizing of malleable parallel applications. We propose and evaluate new scheduling policies and strategies enabled by the ReSHAPE framework. These strategies use both application performance and user-assigned priorities to inform decisions about expanding or contracting the set of processors assigned to a particular job. Experimental results show that the scheduling policies significantly improve individual job turn around time as well as overall cluster utilization.