Graph Theory and their Application to Interconnection Networks for Very-Large-Scale Computing Systems
As illustrated by SciDac's "Paving the Roadmap to Exascale", interconnection networks are known to be one of the main limiting factors to computer systems scalability because of the communication and synchronisation penalties suffered by applications and their significant overheads in terms of cost, power and performance. Datacentres, compute clusters and supercomputers will equally benefit from better performing, more efficient interconnection networks.
This project will be placed within the context of the EPSRC-granted INPUT project whose main objective is to confront practical negative effects in very large scale networks by means of a careful analysis of graph theoretic properties and how to best exploit these properties in practical scenarios. This is an excellent opportunity for developing a multidisciplinary career on computer architecture and graph theory with close contact with two internationally recognized research groups, The APT group at the University of Manchester and the Acid group at the University of Durham.