Improving scheduling techniques in heterogeneous systems with dynamic, on-line optimisations

Computational performance increasingly depends on parallelism, and many systems rely on heterogeneous resources such as GPUs and FPGAs to accelerate computationally intensive applications. However, implementations for such heterogeneous systems are often hand-crafted and optimised to one computation scenario, and it can be challenging to maintain high performance when application parameters change. In this paper, we demonstrate that machine learning can help to dynamically choose parameters for task scheduling and load-balancing based on changing characteristics of the incoming workload. We use a financial option pricing application as a case study. We propose a simulation of processing financial tasks on a heterogeneous system with GPUs and FPGAs, and show how dynamic, on-line optimisations could improve such a system. We compare on-line and batch processing algorithms, and we also consider cases with no dynamic optimisations.

Application of automatic load transfer on feeder utilisation balance

This paper looks at how automatic load transfer may be used as a possible planning tool to help deliver faster connections for customers. A trial on an area of overhead line Network is presented to show how improvements in % feeder utilisation may be realised by changing the location of the open point. The reported Network data is compared to calculated data under two different configurations over a two week trial period. The results show that ALT open point determination in the presence of generation is different from a load only circuit and that the open points may not be fixed with time. Looking at improvements in Network headroom may not be conducive to other improvements in the network such as loss reduction or improving voltage profiles.