Human Computer Interaction (HCI)
Systems, Design & Networks
Energy
Project: PlanSea: Smart Vessel Scheduling
Project:
Performance Enhancement Infrastructure for Skeleton-based Parallel Programming
Frameworks
We present an approach to optimise parallel programs that
are designed using algorithmic skeletons. Our Performance
enhancement Infrastructure (PEI) consists of a Virtual
Instrumentation Protocol (VIP) along with a set of Performance
Enhancement Tools (PETs). . We demonstrate PEI
enhancing performance for several parallel programs written
in the FastFlow framework. The work is part of the EU FP7
ParaPhrase project.
The research applies advanced optimisation techniques to the
problem of efficiently operating a fleet of vessels to support
offshore oil and gas operations. We present demonstration
software to manage large fleets of varying vessel types given
a set of objectives and constraints. The software can simulate
operations, can predict sailing times and operational delays and
can support operational decision making.
Prof. John McCall
Robert Gordon
University
Prof. John McCall
Robert Gordon
University
Project: Gen-O-Fix - A Framework for Self-Improving Software Systems
Software maintenance dominates development cost. The
Gen-O-Fix framework allows a system to be continually
improved (e.g. make better predictions; pass more regression
tests; reduce power consumption) while it runs. The end-user is
only required to provide some measure of system quality.
Dr Jerry Swan
University of Stirling
Project: ParaPhrase: Parallel Patterns for Adaptive Heterogeneous Multicore Systems
The ParaPhrase project aims to produce a new structured
design and implementation process for heterogeneous parallel
architectures (e.g. combining CPUs and GPUs). In the ParaPhrase
approach, developers can exploit a variety of high-level design
patterns to develop component-based applications that can be
mapped to the available parallel hardware resources.These
components may subsequently be dynamically re-mapped to meet
application needs and hardware availability.
Vladimir Janjic
University of St
Andrews