Dr. John Thomson


I have moved to the University of Innsbruck, Austria! My new website is here.


This page is no longer kept up to date.


Research Associate, Institute of Computer Systems Architecture, School of Informatics, University of Edinburgh


I lead the compiler work package of the Milepost project at the University of Edinburgh - an E.U. funded project which develops compiler technology that can automatically learn how to best optimise programs for re-configurable heterogeneous embedded architectures. I work together with the other partners on the project - IBM, INRIA, ARC International and CAPS Enterprises.


My research interests include optimising compilers, embedded systems, applied machine learning, parallelisation techniques, software power-saving techniques and microarchitecture.


I completed my PhD in 2008 on ‘Using Machine Learning to Automate Compiler Optimisation’ under the supervision of Professor Mike O’Boyle. I obtained a BSc with first class honours from the University of Exeter in 2004.


Publications


  1. •Reducing Training Time and Confidence Calculation using Clustering in a Machine Learning Compiler
    John Thomson, Michael O’Boyle, Bjoern Franke and Grigori Fursin
    In Proceedings of International Workshop on Languages and Compilers for Parallel Computing LCPC 09) -     PDF

  2. •Workload Characterization of Embedded Applications Using Decision Trees and Clustering (under submission)
    Damon Fenacci, Bjoern Franke and John Thomson  (under prep for ACM SIGPLAN/SIGBED Conference on Languages, Compilers and Tools for Embedded Systems, LCTES ’10)

  3. •MILEPOST GCC: machine learning based research compiler.
    Grigori Fursin, Cupertino Miranda, Olivier Temam, Mircea Namolaru, Elad Yom-Tov, Ayal Zaks, Bilha Mendelson, Phil Barnard, Elton Ashton, Eric Courtois, Francois Bodin, Edwin Bonilla, John Thomson, Chris Williams, Michael O'Boyle.
    In Proceedings of the GCC Developers' Summit, Ottawa, Canada, June 2008. - PDF

  4. •Using Machine Learning to Focus Iterative Optimization
    Felix Agakov, Edwin Bonilla, John Cavazos, Björn Franke, Michael O’Boyle, John Thomson, and Chris Williams.
    Proceedings of the 4th Annual International Symposium on Code Generation and Optimization (CGO), New York, NY, March 2006 (acceptance rate 29%) - PDF
    Best presentation award. Primary author.

  5. •Predictive Search Distributions
    Edwin Bonilla, Christopher K.I. Williams, Felix Agakov, John Cavazos, John Thomson and Michael F.P. O'Boyle
    In Proceedings of the 23rd International Conference on Machine Learning: ICML'06, Pittsburgh, PA, June 2006. (acceptance rate 19%) - PDF

  6. •Probabilistic Source-Level Optimisation of Embedded Programs
    Björn Franke, Michael O’Boyle and John Thomson
    Proceedings of the 2005 Conference on Languages, Compilers and Tools for Embedded Systems (LCTES'05), Chicago, IL, June 2005, (acceptance rate 25%) - PDF

  7. •Using Machine Learning to Automate Compiler Optimisation
    PhD Thesis. University of Edinburgh, 2008. -     PDF

  8. •Using Models for to Speed Up Smart Compilation (under submission)
    John Thomson, John Cavazos, Edwin Bonilla and Michael O'Boyle
    (submitted to ACM Transactions on Architecture and Code Optimization)

Contact:

Room 1.02 Informatics Forum,

University of Edinburgh,

10 Crichton Street,

Edinburgh,

EH8 9AB,

UK.

Email: