DMR API: Improving cluster productivity by turning applications into malleable (2005.05910v2)

Abstract: Adaptive workloads can change on--the--fly the configuration of their jobs, in terms of number of processes. In order to carry out these job reconfigurations, we have designed a methodology which enables a job to communicate with the resource manager and, through the runtime, to change its number of MPI ranks. The collaboration between both the workload manager---aware of the queue of jobs and the resource allocation---and the parallel runtime---able to transparently handle the processes and the program data---is crucial for our throughput-aware malleability methodology. Hence, when a job triggers a reconfiguration, the resource manager will check the cluster status and return an action: an expansion, if there are spare resources; a shrink, if queued jobs can be initiated; or none, if no change can improve the global productivity. In this paper, we describe the internals of our framework and how it is capable of reducing the global workload completion time along with providing a smarter usage of the underlying resources. For this purpose, we present a thorough study of the adaptive workloads processing by showing the detailed behavior of our framework in representative experiments and the low overhead that our reconfiguration involves.