A Load Balancing Model for 3D HPC Image Reconstruction Applications To Reduce Unnecessary Overheads
Jose Antonio Alvarez-Bermejo (Universidad de Almeria)Javier Roca-Piera (Universidad de Almeria)
Abstract:
Scientific applications are high-resource demanding software. HPC constitute a suitable framework to deploy such applications. Porting them to parallel platforms is a tedious task that implies adding a complexity layer to the software. There is still a gap between parallel architectures and want-to-be-parallel software. Abstractions are an useful tool to bridge the gap. This paradigm shift can be used to exploit parallelism and add adaptivity layers easily by means of load balancing strategies. In this paper we present results in porting scientific software to parallel platforms.
Keywords:
Parallel and Distributed Computing
