The AS4DR (Adaptive Scheduling for Distributed Resources) scheduling method presented in this paper aims at maximizing throughput, when processing several data streams by divisible load applications on star-shaped distributed memory platforms, with available speeds for communicating and computing which may be poorly estimated, or varying over time. The total workload is supposed to be unknown. According to the computation cost model, AS4DR can either maximize throughput, or CPU utilization by avoiding data-starvation of the computing units. An experimental assessment of the adaptation of the workload distribution to the variation of the communicating and computing speeds has been performed that shows that the use of AS4DR can significantly improve the throughput. This paper also experimentally assesses a resource selection method to set up star-shaped clusters of distributed resources, so as to process efficiently a set of data streams with AS4DR.

The AS4DR (Adaptive Scheduling for Distributed Resources) scheduling method experimented in this paper aims at maximizing the CPU use efficiency when executing divisible load applications on heterogeneous distributed memory platforms. AS4DR adapts the scheduling to: the unawareness of the total workload, both the unspecification and the variation over time of the execution parameters (available communication speed, available computing speed, etc.). This paper presents the first experimental assessments of the adaptivity of the scheduling with this method.

In this paper, we present an adaptive method for scheduling parallel applications on heterogeneous multi-processor platforms, in a dynamic context. When executing divisible load applications according to a master-worker model, this method delivers the workload through multiple rounds and can avoid contentions in the use of the network. Such contentions avoidance is a means to avoid idle time, thus to maximize the throughput. Before presenting the proposed scheduling method, the paper revisits a runtime method which motivated its development. The method presented in this paper can be used to schedule parallel applications whose total workload is large but unknown a priori.