Dr. Francis Russell
email address graphic


I have left Imperial College and moved to Hadean.

I was previously a research associate under the supervision of Wayne Luk in the Custom Computing research group. Before that, I was a member of the Software Performance Optimisation research group under the supervision of Paul Kelly.

My most recent research involved the simulation of chaotic systems on FPGAs in the context of atmospheric modelling.

Previous research involved the investigation of the applicability of Domain Specific Languages to the quantum chemistry package ONETEP. Further details can be found in slides I've presented on this work here.

Prior to that, I looked at the use of active libraries in the context of optimising the finite element method. The outcome of this work was a library, Excafé, that generated optimised C code for performing finite element local assembly. This work was presented at the FEniCS'11 workshop.


These are the slides I presented at the FCCM 2015 conference. They cover work exploring the implementation of chaotic systems on FPGAs. Specifically, I look at the Lorenz'96 chaotic system and evaluate the power and performance benefits to an FPGA system when the precision is reduced. Due to the chaotic nature of the system, we can reduce precision further than we might otherwise be able, since we do not have to match the behaviour of a reference implementation to produce valid results.
These are the slides I presented at the June 2013 PSL (Particle Science Language) project meeting. They discuss my work on improving the performance of the Fourier interpolation of 3D datasets. This is a performance-critical kernel in the linear-scaling quantum chemistry application ONETEP.
These are the slides I presented at the FEniCS'11 workshop on using symbolic algebra to reduce the operation count of local assembly in the finite element method. More information on the library developed to explore these techniques, Excafé, can be found here.
These are a slightly updated version of the slides that were presented at my 4th year MEng individual project presentation.


Bartosz D. Wozniak, Freddie D. Witherden, Francis P. Russell, Peter E. Vincent and Paul H. J. Kelly. Computer Physics Communications, Volume 202, May 2016, Pages 12-22.
Peter D. Düben, Francis P. Russell, Xinyu Niu, Wayne Luk and T. N. Palmer. Journal of Advances in Modelling Earth Systems, Volume 7, Pages 1393-1408, August 2015.
Francis P. Russell, Peter D. Düben, Xinyu Niu, Wayne Luk and T. N. Palmer. 2015 IEEE 23rd Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM), Pages 171-178, May 2015.
Doru Thom Popovici, Francis P. Russell, Karl Wilkinson, Chris-Kriton Skylaris, Paul H. J. Kelly and Franz Franchetti. Parallel and Distributed Processing Symposium (IPDPS), 2015 IEEE International, May 2015, Pages 743-752.
Francis P. Russell, Karl Wilkinson, Paul H. J. Kelly and Chris-Kriton Skylaris. Computer Physics Communications. Volume 187, February 2015, Pages 8-19. Available online 7 October 2014.
Francis P. Russell and Paul H. J. Kelly. ACM Transactions on Mathematical Software. Volume 39, Issue 4, July 2013, Article 26. Pre-print.
Francis P. Russell, Michael R. Mellor, Paul H. J. Kelly and Olav Beckmann. Science of Computer Programming, Volume 76, Issue 4, April 2011, pp. 227-242. Accepted Author Manuscript.
Francis P. Russell, Michael R. Mellor, Paul H. J. Kelly and Olav Beckmann. In Proceedings of the Second International Workshop on Library-Centric Software Design (LCSD'06), pp. 5-13.


In this thesis, I explore an approach called "active libraries". These are libraries that take part in their own optimisation, enabling both high-performance code and the presentation of intuitive abstractions.

I investigate the use of active libraries in two domains. Firstly, dense and sparse linear algebra, particularly, the solution of linear systems of equations. Secondly, the specification and solution of finite element problems.

Extending my earlier (MEng) thesis work, I describe the modifications to my linear algebra library "DESOLA" required to perform sparse-matrix code generation. I show that optimisations easily applied in the dense case using code-transformation must be applied at a higher level of abstraction in the sparse case. I present performance results for sparse linear system solvers generated using DESOLA and compare against an implementation using the Intel Math Kernel Library. I also present improved dense linear-algebra performance results.

Next, I explore the active-library approach by developing a finite element library that captures runtime representations of basis functions, variational forms and sequences of operations between discretised operators and fields. Using captured representations of variational forms and basis functions, I demonstrate optimisations to cell-local integral assembly that this approach enables, and compare against the state of the art.

As part of my work on optimising local assembly, I extend the work of Hosangadi et al. on common sub-expression elimination and factorisation of polynomials. I improve the weight function presented by Hosangadi et al., increasing the number of factorisations found. I present an implementation of an optimised branch-and-bound algorithm inspired by reformulating the original matrix-covering problem as a maximal graph biclique search problem. I evaluate the algorithm's effectiveness on the expressions generated by our finite element solver.

This is the project report for my 4th year MEng individual project. It describes a C++ numerical library that delays evaluation of the calls made to it. When the results of the computation are required, optimised code is generated at runtime to perform the delayed operations.
This is the project report from my 3rd year group project. The project was an experiment with a filesystem where files are identified by metadata rather than filenames. This makes the filesystem usage similar to that of a database.
Views can be defined on the filesystem filtering on selection criteria. These views are notified when changes, addition or removal of files alter the results. Many tasks become easier under such a system. Music players no longer need to search and index media folders. They can just create a view containing all music files, retrieve any metadata through the view and be notified of all changes.


I lectured the first-year undergraduate C course in 2012 and 2013.
I have previously co-lectured the first-year optional Advanced Java Programming course with Dr. Tristan Allwood.