EPCC Project Manager
2404 James Clerk Maxwell Building,
The King's Buildings,
+44 (0) 131 650 5201 / 6555
+44 (0) 7598317015
Since 2008, Iain has worked at EPCC, specialising in performance analysis and optimization of HPC applications. He collaborates with a range of researchers around the UK and Europe in fields as diverse as Computational Chemistry, Mathematics and Biomechanical Engineering. He develops scalable software to allow researchers to harness HPC resources such as HECToR and ARCHER - the UK National HPC Services - and the European PRACE infrastructure. Most of his work has been in the area of Materials Science, and he is a developer on the CP2K project - a powerful program enabling Density Functional Theory and other atomistic modelling techniques to be applied a diverse range of systems using tens of thousands of CPU cores at a time.
Iain has a wide-ranging interest in high performance computing architectures and programming models including MPI, OpenMP, CUDA and OpenCL. He teaches several classes on EPCC's MSc in HPC on Numerical Algorithms, Performance Optimisation, and organises the 'HPC Ecosystem' course, which looks at how applications developers, HPC vendors and users are collaborating to tackle the upcoming challenge of Exascale computing.
Believing that outreach to the public about scientific research is an important part of science itself, Iain is a developer for the PrimeGrid volunteer computing project, which promotes mass participation in the solution of mathematical problems, as well as finding very large - million digit - prime numbers.
Prior to joining EPCC, Iain earned a First Class BSc. (Hons) in Computer Science & Physics from Edinburgh University in 2005 before spending two and a half years working for IBM on SAN storage virtualization.
In his limited spare time, Iain enjoys white water kayaking, hill walking and is an avid space exploration and research enthusiast.
- CP2K-UK - an EPSRC-funded project to improve the usability of CP2K, support the growth of the CP2K user and developer community in the UK via networking activities, improved software engineering, and new algorithm development. EPCC leads the work, in collaboration with KCL and UCL, in addition to many more group involved in the wider network.
- CP2K-NSF - supporting UK-US project led by Prof. Ilja Siepmann (University of Minnesota) and Dr. Ben Slater (UCL) to develop new Monte Carlo methods in CP2K to allow the study of reactive phase and sorption equilibria.
- ARCHER - the new UK National HPC Service based on the Cray XC30 platform. Supporting CP2K usage and acting as liaison to the UK Materials Chemistry HPC Consortium.
- ExTASY - we are developing an "Extensible Toolkit for Advanced Samping and analYsis" to couple novel algorithms for large time-step molecular dynamics, advanced analysis tools, and large scale ensemble calculations in order to overcome the Grand Challenge of effective sampling of complex macro-molecules. Funded jointly by EPSRC and NSF, the project team comprises Rice U., Rutgers U., IBM, Duke U., U. of Edinburgh, U. of Nottingham and Imperial College London.
- NAIS - as part of the Centre for Numerical Algorithms and Intelligent Software, Iain is developing improved programs for primality testing using auto-tuning and accelerator hardware.
- PhD - Iain is studying part-time towards a PhD in materials physics, supervised by Prof. Graeme Ackland at the University of Edinburgh, to investigate the properties and structure of silicate melts found at high temperature and pressure in the Earth's mantle
- MSc in HPC - organising the 'HPC Ecosystem' course and teaching Fast Fourier Transforms as part of 'Parallel Numerical Algorithms'
- TPLS/Microchannels - collaboration with Dr. Prash Valluri (UoE Engineering) and Dr. Lennon O Naraigh (UCD) to impelement more scalable solvers in their TPLS (Two-Phase Level Set) fluid dynamics code, which models the interface between two phases of a fluid using ultra-high resolution DNS.
- EPCC Publicity & Outreach - taking EPCC to other departments, universities and science fairs (e.g. British Science Festival, Bang Goes The Borders). Developing publicity videos for EPCC projects.
- PRACE - as part of the Applications work package of PRACE, we are porting and optimisng CP2K for the Intel Xeon Phi co-processor architecture. Three previous projects under PRACE have also extended mixed-mode OpenMP parallelism in CP2K, and developed an auto-tuning framework for key computational kernels, to support efficient massively parallel MP2 calculations on the PRACE Tier-0 HPC resources.
- Novel Asynchronous Algorithms and Software for Large Sparse Systems - an EPSRC-funded collaboration with Universities of Manchester, Leeds, Hull and Strathclyde to develop the next generation of methods for solving linear algebra problems with billions of unknowns using tens of thousands of CPU cores.
- HPC Simulations of complex solids and clusters using static lattice techniques - EPSRC Software Development project with UCL and Curtin University (Australia) to integrate GULP, the 'General Utility Lattice Program' with a new controller code in order to support a wide range of materials science applications.
- ChipSHOP - a collaboration with biological scientists and engineers from Heriot-Watt and Edinburgh Universties formed as a result of winning the Heriot-Watt Crucible in May 2012 to develop software for in-silico design and optimisation of micro-fluidic devices for applications in separation of biological fluids such as DNA/blood plasma, and cells/buffer solution.
- TEXT "Towards EXaflop applicaTions" - an EU FP7 research project exploring the use of the MPI and SMPSs as a hybrid programming model to allow high productivity and efficient use of current and future HPC architectures.
- Commercial Projects - optimised use of FFT in a key code for a Scottish Oil & Gas services SME.
- HECToR dCSE: VOX-FE - implemented parallel I/O for improved scalability and performance.
- HECToR dCSE: Quantum Espresso - modified FFT schemes for SMP cluster systems.
- HECToR dCSE: CP2K - three projects improving MPI and OpenMP parallelisation in all key areas of the code including grid operations, FFT, sparse matrix multiplication
- ACE "Architecture Comparison Exercise" - benchmarking and performance analysis on 5 representative UK codes on a range of HPC architectures, to understand the links between hardware characteristics and application performance.
- HPCx - supporting the 'Complementary Capability Computing' initiative: parallel visualisation, interactive MD, job submission portlets.
- "Linear instability, nonlinear instability, and ligament dynamics in three-dimensional laminar two-layer liquid/liquid flows", L. O'Naraigh, P. Valluri, D. Scott, I. Bethune, P. Spelt, Journal of Fluid Mechanics, Jul 2014, 750: 464-506
- "Performance analysis of asynchronous Jacobi's methods implemented in MPI, SHMEM and OpenMP", I. Bethune, M. Bull, N. Dingle, N. Higham, International Journal of High Performance Computing Applications, Feb 2014, 28: 97-111
- "Mapping application performance to HPC architecture", A. Gray, I. Bethune, R. Kenway, L. Smith, M. Guest, C. Kitchen, P. Calleja, A. Korzynski, S. Rankin, M. Ashworth, A. Porter, I. Todorov, M. Plummer, E. Jones, L. Steenman-Clark, B. Ralston, C. Laughton, Comp. Phys. Comm. Volume 183, Issue 3, March 2012, Pages 520-529
- "CP2K Performance from Cray XT3 to XC30", I. Bethune, F. Reid, A. Lazzaro, Cray User Group 2014
- "Automated regression testing and code coverage analysis of the CP2K application", M. Misic, I. Bethune, M. Tomasevic, International Conference on Software Testing 2014
- "Linear and nonlinear instability and ligament dynamics in 3D laminar two-layer liquid/liquid flows", L. O'Naraigh, P. Valluri, D. Scott, I. Bethune, P. Spelt, Bulletin of the American Physical Society, Vol 58, 18, Nov 2013
- "A hybrid approach for extreme scalability when solving linear systems", N. Brown, M. Bull, I. Bethune, Exascale Applications and Software Conference (EASC), 2013
- "Extending the generalized Fermat prime search beyond one million digits using GPUs", I. Bethune, M. Goetz, Proceedings of the 10th International Conference on Parallel Processing and Applied Mathematics, PPAM 2013, Lecture Notes in Computer Science 8384, 106-113, 2014 (Slides)
- "PRACE DECI (Distributed European Computing Initiative) Minisymposium, C. Johnson, A. Carter, I. Bethune, K. Stratford, M. Alava, V. Cardoso, M. Asif, B. S. A. Schuberth, T. Weinzierl, Applied Parallel and Scientific Computing Lecture Notes in Computer Science Volume 7782, 2013, pp 43-60
- "Improving the performance of CP2K on the Cray XT", I. Bethune, Proceedings of the Cray User Group (2010), May 2010 (Slides)
Technical Reports & White Papers
- "Optimising CP2K for the Intel Xeon Phi", F. Reid, I. Bethune, PRACE White Paper, 2013
- "Evaluating CP2K on Exascale Hardware: Intel Xeon Phi", F. Reid, I. Bethune, PRACE White Paper, 2013
- "Solving Large Sparse Linear Systems using Asynchronous Multisplitting", N. Brown, J. M. Bull, I. Bethune, PRACE White Paper, 2013
- "Performance Enhancement and Optimization of the TPLS and DIM Two-Phase Flow Solvers", D. Scott, L O Naraigh, I. Bethune, P. Valluri, P. Spelt, HECToR dCSE Report, Jun 2013
- "High Performance MP2 for Condensed Phase Simulations", R. Reyes, I. Bethune, PRACE White Paper, 2013
- "Adding Parallel I/O to PARA-BMU", N. Johnson, I. Bethune, HECToR dCSE Report, Oct 2012
- "Application Porting from External Developers", J. Labarta, I. Bethune, M. Garcia, V. Lopez. G. Rilley, TEXT Project Deliverable, Aug 2012
- "CP2K - Scalable Atomistic Simulations for the PRACE Community", I. Bethune, A. Carter, K. Stratford, P. Korosoglou, PRACE White Paper, 2012
- "CP2K - Sparse Linear Algebra on 1000s of Cores", I. Bethune, HECToR dCSE Report, Jan 2012
- "Million Atom KS-DFT with CP2K", I. Bethune, A. Carter, X. Guo, P. Korosoglou, PRACE White Paper, 2011
- "Improving the scalability of CP2K on multi-core systems", I. Bethune, HECToR dCSE Report, Sep 2010
- "Improving the performance of GWW", I. Bethune, HECToR dCSE Report, Aug 2009
- "Improving the performance of CP2K on HECToR", I. Bethune, HECToR dCSE Report, Jul 2009
- "Parallel Visualisation on HPCx", I. Bethune, HPCx Technical Report, Apr 2008
- Acceleration of Diagrammatic Determinantal Quantum Monte Carlo Calculations using GPUs, Apr 2014 - PDF
- Developing a User-Centric CP2K Benchmark Suite, Feb 2014 - PDF
- CP2K: Capability and Scalability, Apr 2013 - PDF
- High Performance Computing with CP2K, Jul 2012 - PDF
- Heriot-Watt Crucible Bio Poster, Feb 2012 - PDF
- "High-Performance Computing", Omega Tau Podcast Episode, Aug 2014
- "Porting CP2K to the Intel Xeon Phi", ARCHER Technical Forum Webinar, July 2014, (Slides)
- "CP2K: Scalable atomistic simulation on HECToR", HECToR dCSE Technical Meeting, Mar 2014, (Slides)
- "PrimeGrid: Searching for a New World Record Prime Number", Mathematics Today, Vol: 49 No. 6 Dec 2013, 250-251
- "PrimeGrid: Join in the search for a world record prime number", British Science Festival, Sep 2013 (Slides, Audio)
- "Density Functional Theory for Dummies", EPCC Seminar, Jun 2013, (Slides)
- "Research in a Nutshell", 1 minute Video, Feb 2012
- "Prime Numbers, Large Integer Arithmetic, and a 2PF Task Farm", EPCC Seminar, Sep 2011, (Slides, Audio)
- "Mixed-mode Parallelism in CP2K: A Case Study", DEISA/PRACE Spring School, Mar 2011, (Slides)
- Project highlights videos, 2011 - HPC Europa, OGSA-DAI
- "Improving the scalability of Domain Decomposition within CP2K", HECToR dCSE Technical Meeting, Sep 2009 (Slides)
Supervision of Research Projects
- "Multi-platform parallel code coverage and regression testing with CP2K", M. Misic, PRACE Summer of HPC, 2013 (Video)
- "Diagrammatic Determinantal Quantum Monte Carlo Calculations on GPUs", M. Schmitt, MSc in HPC, 2013
- "Accelerated Primality Testing Using GPUs", D. McBain, MSc in HPC, 2013
- "Improving the Computing Efficiency of the Regional Genomic Relationship Mapping Approach", J. C. Ortuno, MSc in HPC, 2012
- "Optimising the DBCSR GPU Implementation", J. Chetty, MSc in HPC, 2011
- "Getting to know the LLVM compiler", G. Ye, MSc in HPC, 2011
- "Investigation into hardware acceleration of HPC kernels, within a cross platform OpenCL environment", S. Fraser, MSc in HPC, 2010
- "GPU Acceleration of HPC Applications", A. Richardson, MSc in HPC, 2009
- "Optimising EM", J. Read, MSc in HPC, 2009
A list of programs which Iain has contributed to the development of:
- CP2K - Highly scalable, modular, open source molecular dynamics program, featuring wide range of force evaluation methods from potentials, linear scaling DFT, Hartree-Fock and MP2. Large, distributed developer community led by Prof. Joost VandeVondele (ETH Zurich) and Prof. Jurg Hutter (University of Zurich).
- TPLS - CFD software for ultra-high resolution DNS of turbulent two-phase flows, developed by Dr. Prash Valluri and Dr. Lennon O Naraigh.
- VOX-FE - Finite Element Analysis software for voxel data sets. Targetted at mechanical properties of bones. Developed in collaboration with Prof. Michael Fagan's group at University of Hull.
- Quantum Espresso - Popular suite of open source codes for materials science. Mainly developed in Italy (DEMOCRITOS, SISSA, CINECA).
- GULP - Lattice dynamics code development by Julian Gale, Curtin University
- LSDMap - Analysis software for automatic determination of collective variables for MD trajectory data. Developed by Prof. Cecilia Clementi's group, Rice University
- Genefer - Programs for finding large Generalized Fermat Primes. Developed in collaboration with various people from the PrimeGrid project.
- PRPNet - A client/server application for prime search projects, developed by Mark Rodenkirch and used by PrimeGrid
A list of grants which Iain has held:
|Large scale voxel based modelling||Co-proposer||ARCHER eCSE||May-14||-||£50,700|
|TPLS: Optimised Parallel I/O and Visualisation||Co-proposer||ARCHER eCSE||Apr-14||-||£57,900|
|SI2-CHE: Collaborative Research: Developing First Principles Monte Carlo Methods for Reactive Phase and Sorption Equilibria in the CP2K Software Suite||Co-I||EPSRC||Sep-13||-||£15,000|
|SI2-CHE: ExTASY: Extensible Tools for Advanced Sampling and analYsis||Co-I||EPSRC||Jul-13||-||£277,000|
|Modelling Complex Oxides using CP2K on Intel Xeon Phi||Proposer||PRACE-1IP Ext.||Jul-13||Dec-13||6 person-months|
|Boiling in microchannels: integrated design of closed-loop cooling system for devices operating at high heat fluxes||Project Partner||EPSRC||May-13||Oct-13||£23,000|
|Evaluating CP2K on Exascale Hardware: Intel Xeon Phi||Proposer||PRACE-3IP||Jan-13||Jun-13||6 person-months|
|Performance enhancement and optimization of the TPLS and DIM two-phase flow solvers||Co-proposer||HECToR dCSE||Sep-12||Apr-13||£75,000|
|High Performance MP2 for Condensed Phase Simulations||Co-proposer||PRACE-2IP||Apr-12||Oct-12||6 person-months|
|Adding Parallel I/O to PARA-BMU||Co-proposer||HECToR dCSE||Feb-12||Sep-12||£39,000|
|HPC simulations of complex solids and clusters using static lattice techniques||Co-I||EPSRC||Oct-11||Sep-13||£75,000|
|CP2K: Scalable Atomistic Simulation for the PRACE community||Co-proposer||PRACE-1IP||Jun-11||Jun-12||10 person-months|
|Million Atom KS-DFT with CP2K||Co-proposer||PRACE-1IP||Dec-10||May-11||6 person-months|
|CP2K - Sparse Linear Algebra on 1000s of cores||Co-proposer||HECToR dCSE||Oct-10||Dec-11||£56,000|
|Improving the performance of GWW||Co-proposer||HECToR dCSE||Feb-10||May-10||£38,000|
|Improving the scalability of CP2K on multi-core systems||Co-proposer||HECToR dCSE||Sep-09||Sep-10||£56,000|
A non-exhaustive list of research highlights from Iain's collaborators:
- Dr. Ben Slater and others at UCL use CP2K to study the properties of ice crystasl surfaces e.g. Large variation of vacancy formation energies in the surface of crystalline ice", Watkins, Pan, Wang, Michaelides, VandeVondele and Slater, Nature Materials 10, 794-798 (2011), "Point defects at the ice (0001) surface", Watkins, VandeVondele and Slater, Proceedings of the National Academy of Sciences of the USA 107(28), 12429-12434 (2010)
- Prof. Joost VandeVondele and colleagues in Zurich found a mechaism for regeneration of coatings of dye-sensitised solar cells using atomistic simulation with CP2K - "An atomistic picture of the regeneration process in dye sensitized solar cells" - Shiffmann, VandeVondele, Hutter, Urakawa, Wirz and Baiker, Proceedings of the National Academy of Sciences of the USA 107(11), 4830-4833 (2010)
- Dr. Carole Morrison from Edinburgh University's School of Chemisty studies proton transport through pores in biological membranes using CP2K - Simulating Proton Transport through a Simplified Model for Trans-Membrane Proteins, Shepherd and Morrison, J. Phys. Chem. B 114(20), 7047-7055 (2010)
- Prof. Michael Fagan's group use VOX-FE to study the biomechanical properties of bones - "Why do humans have chins? Testing the mechanical significance of modern human symphyseal morphology with finite element analysis.", Groning, Liu, Fagan and O'Higgins, Americal Journal of Physics Anthropology 144, 593-606 (2011)
- Prof. Merlyne De Souza uses Quantum Espresso to calculate the electronic structure and properties of carbon nanotubes, with applications in device electronics - "Role of hybridization on the Schottky barrier height of carbon nanotube field effect transistors", Casterman, De Souza, Tahraoui, Durkan and Milne, Phys. Rev. B 79 (2009)
- Prof. Cecilia Clementi's research group are applying the Locally-scaled Diffusion Map method using LSDMap to determine the fundamental dynamics of a wide variety of macromolecular systems e.g. "Polymer reversal rate calculated via locally scaled diffusion map", Zheng, Rohrdanz, Maggioni and Clementi, J. Chem Phys. 134 (2011)
Employment & Education
|2012 - Present
||Project Manager, EPCC, The University of Edinburgh
|2008 - 2012
||HPC Applications Consultant, EPCC, The University of Edinburgh
|2005 - 2008
||Software Engineer, IBM United Kingdom Ltd.
|2001 - 2005
||BSc. (Hons) 1st Class, Computer Science and Physics, The University of Edinburgh