Graham Riley

I am a Lecuter in Computer Science. Following a BSc in Physics from Manchester University, I worked in real-time simulation, first as a software engineer with Rediffusion Flight Simulation and then with Ferranti Simulation and Training as a software team leader, software system designer and project manager. I obtained an MSc (by research) in October 1996 and have was a member of the Centre for Novel Computing (CNC ) in the School of Computer Science from its inception in 1990 until March 2013 when I joined the Advanced Processing Technologies Group (APT) in the School.

Current and Recent Projects
The main research projects that I am currently involved with as an investigator are: IS-ENES-2, an EC project developing new infrastructures for Earth System Modelling and the NERC/Met Office-funded 'GungHo' project which is seeking to develop a new dynamics code for Met Office`s Unified Model (UM), which is portably scalable, for use on future many-core supercomputers. GungHo is funded under the Next Generation Weather and Climate Prediction (NGWCP) programme. (The original meaning of Gung Ho comes from a Chinese expression for 'working together harmoniously'). In addition, I am the project manager of the EPSRC programme grant PAMELA (a Panoramic Approach to the Many-corE LAndscape), an APT project seeking to develop new low-power SoC solutions for advanced device-based applications such as 3D scence understanding. I have had funding in 2014 and 2015 from STFC/Daresbury to support Computational Science Activities (CSA) related to the U.K. Met Office.

Past projects include: ERMITAGE, an EC project support Integrated Assessment Modelling for Climate Impact studies, GSUM, an EPSRC-funded HPC Software Development project looking to develop generic parallel strategies for the UM from the Met Office, and an EU project, METAFOR, focussed on metadata developments for Earth System Modelling. Recent projects include: IntBioSim, a BBSRC-funded e-Science project and GENIEfy, a NERC-funded e-Science project, RealityGrid, and EPSRC-funded e-Science project and APART project completed a few years ago, and was a working group consisting of European and US partners which focussed on techniques and tools to support automatic performance analysis. APART was funded jointly by the EC and NSF.

I have also worked, with Rupert Ford (ex-CNC, now at STFC Daresbury), as a consultant for the Met Office, on the FLUME project investigating aspects of the software architecture for their next generation Unified Model.

Research Interests
(i) Techniques, tools and methods for developing high performance applications for parallel machines, including performance analysis and performance improvement, now-a-days in the context of energy efficient computing for extreme-scale systems.

(ii) Software architectures for the flexible composition and deployment of scientific coupled modelling.

(iii) Performance control techniques for component-based applications executing on distributed computing resources.

I have previously worked on a joint project between computer science and Bioinformatics called PrePRINTS in which CS techniques were applied to create web-based tools which interact with a distributed, heterogeneous, high performance server environment supporting the creation and curation of protein sequence databases. I am always on the lookout for new collaborators with applications requiring high performance.

I also teach on modules in the Parallel Computing in the Multi-core era theme of the Advanced Computing MSc course and on the 3rd year Chip Multiprocessors course. Applications for PhD studies are always welcome.

E-mail

graham.riley at manchester.ac.uk
Telephone
 UK 0161 275 5724
 +44 161 275 5724
Snail-mail
 School of Computer Science,
 University of Manchester,
 Oxford Road,
 Manchester M13 9PL,
 United Kingdom